[
  {
    "start": 0.0,
    "end": 6.84,
    "text": "This podcast is presented by NXAI, your partner for time series foundation models and physical AI."
  },
  {
    "start": 19.16,
    "end": 23.28,
    "text": "And welcome to a new episode of our Industrial AI Podcast!"
  },
  {
    "start": 23.78,
    "end": 25.74,
    "text": "My name's Robert Weber and it's pleasure to talk too."
  },
  {
    "start": 26.8,
    "end": 28.92,
    "text": "Peter Sieber good morning Robert."
  },
  {
    "start": 29.1,
    "end": 33.14,
    "text": "did you arrive home well from our AI in the Alps?"
  },
  {
    "start": 33.5,
    "end": 38.78,
    "text": "Yes I arrived i think at midnight on Friday."
  },
  {
    "start": 39.46,
    "end": 48.06,
    "text": "And yeah, we spent three days at our event AI in the Alps and a wonderful Lech or Zug et lech near Lech."
  },
  {
    "start": 48.3,
    "end": 50.34,
    "text": "So it was amazing I think."
  },
  {
    "start": 51.34,
    "end": 52.3,
    "text": "what is your opinion on this?"
  },
  {
    "start": 53.14,
    "end": 56.32,
    "text": "Yeah normally you want to know who's best of us."
  },
  {
    "start": 58.58,
    "end": 62.84,
    "text": "Exactly that was my second question now!"
  },
  {
    "start": 63.8,
    "end": 75.8,
    "text": "pick out one which was I was impressed by the work that Bastjan, Bastjan Aoudi is a CEO of Simconright and he's working with simulation for injection molding."
  },
  {
    "start": 76.6,
    "end": 81.54,
    "text": "And that came up during... That was my use case session with Marco right?"
  },
  {
    "start": 81.94,
    "end": 86.2,
    "text": "That is Professor Marco Huber from the Fraunhofer Institute."
  },
  {
    "start": 87.54,
    "end": 89.04,
    "text": "as we discussed physical."
  },
  {
    "start": 89.16,
    "end": 96.06,
    "text": "We were uncertain if that was physical AI or more physics-informed AI... So it's for the buzzword bingo at The Evening Man?"
  },
  {
    "start": 96.2,
    "end": 97.8,
    "text": "Yeah, interestingly!"
  },
  {
    "start": 99.26,
    "end": 102.82,
    "text": "That is a way to written down on one of the flip charts right here."
  },
  {
    "start": 103.16,
    "end": 106.12,
    "text": "yeah independent off the buzz word Bingo."
  },
  {
    "start": 106.4,
    "end": 121.38,
    "text": "I was really impressed with what he doing and For that reason i'm gonna do a podcast recording with Bastjan over this summer so all you can kind of learn directly after our break, what it is exactly that he's doing."
  },
  {
    "start": 121.44,
    "end": 127.52,
    "text": "But basically you know... It's about simulation for injection molding."
  },
  {
    "start": 127.74,
    "end": 140.58,
    "text": "so just really-really He was using a very high number improving the quality and quantity of the molds That you produce before you then can start producing."
  },
  {
    "start": 141.04,
    "end": 143.34,
    "text": "You know whatever your producing with the injection mold."
  },
  {
    "start": 143.68,
    "end": 144.16,
    "text": "Very impressive."
  },
  {
    "start": 144.76,
    "end": 147.16,
    "text": "I think he does that together with Johannes Prandtstetter."
  },
  {
    "start": 147.4,
    "end": 148.0,
    "text": "Or am i wrong?"
  },
  {
    "start": 148.52,
    "end": 158.96,
    "text": "He has been, they have been working very closely together and there were some details that he was sharing also related to right what they had been doing together."
  },
  {
    "start": 159.08,
    "end": 171.78,
    "text": "And then as Johannes company got sold Got bought They had two kind of dismantled pieces but yeah specific stuff together."
  },
  {
    "start": 172.0,
    "end": 175.4,
    "text": "And they also met at AI in YAPS, I think three years ago or something?"
  },
  {
    "start": 176.14,
    "end": 176.52,
    "text": "Two years."
  },
  {
    "start": 176.68,
    "end": 178.26,
    "text": "Okay yeah two years."
  },
  {
    "start": 178.34,
    "end": 181.24,
    "text": "Yeah i do recall two years ago that Johannes there was a right."
  },
  {
    "start": 181.54,
    "end": 188.82,
    "text": "okay so That one best of the show award which we don't have but it's just a thought Just to share that."
  },
  {
    "start": 189.1,
    "end": 194.56,
    "text": "always at the end for The Three Days We Always Do This Pleasant Minus Thing and It Didn't Come Up."
  },
  {
    "start": 194.62,
    "end": 208.34,
    "text": "But I Just Thought this Morning Like Maybe Maybe we should do that, and if it were done at this time I would most certainly give to Zepp."
  },
  {
    "start": 209.9,
    "end": 212.86,
    "text": "inventor of LSDM, twenty seven, twenty eight years ago."
  },
  {
    "start": 213.28,
    "end": 222.62,
    "text": "Meanwhile co-founder chief scientist at NXAIM now Zepp who actually regularly reminds me off myself."
  },
  {
    "start": 222.72,
    "end": 232.12,
    "text": "he talks with his hands a lot and that is what my friends tell me or they joke about when I do it as well."
  },
  {
    "start": 232.92,
    "end": 240.0,
    "text": "so there you go very impressive, exciting and convincing way."
  },
  {
    "start": 241.12,
    "end": 262.52,
    "text": "He explained what he says And if not If I mean...I understand it like that i'm gonna share with you the revolution That is being brought to us by Tyrex too which Is The Multivariate Time Series Foundation Model Which You Could All Dear Listeners Hear About Last Episode hear about."
  },
  {
    "start": 262.56,
    "end": 264.1,
    "text": "like a week ago, we go two weeks ago."
  },
  {
    "start": 264.66,
    "end": 266.86,
    "text": "And it was actually announced on the very first day."
  },
  {
    "start": 266.96,
    "end": 267.42,
    "text": "is that right?"
  },
  {
    "start": 267.48,
    "end": 274.64,
    "text": "That we were there and then Thursday exactly exactly what went live I think on Thursday Thursday morning."
  },
  {
    "start": 274.68,
    "end": 279.72,
    "text": "yeah Right now What is so revolutionary about it?"
  },
  {
    "start": 279.94,
    "end": 286.98,
    "text": "You do not anymore need to train your model as we, We use a little bit too strong."
  },
  {
    "start": 287.16,
    "end": 304.74,
    "text": "As some of you know and I've said before many years ago... ...I was involved but the program manager never like as the scientist I understood enough and that's, i'm gonna share with you one example there just to tell."
  },
  {
    "start": 305.42,
    "end": 315.12,
    "text": "This was about a testing machine for of the structural components of the ASML chip production machines."
  },
  {
    "start": 315.58,
    "end": 327.04,
    "text": "So I'm not going to go into details as I signed an NDA a long time ago but think about something really, really structural without which machine would not function."
  },
  {
    "start": 327.72,
    "end": 331.2,
    "text": "and at that time this testing machine was broken."
  },
  {
    "start": 331.46,
    "end": 344.76,
    "text": "it had been broken for whatever reason And we were brought in, and that was at the very first time coming out of softening the industrial data solution or something like."
  },
  {
    "start": 346.4,
    "end": 352.78,
    "text": "We received a year worse off-data over about hundred fifty different sensors right?"
  },
  {
    "start": 353.48,
    "end": 361.6,
    "text": "Our job is to tell them if with means of deadtime machine learning this could have been circumvented, you know."
  },
  {
    "start": 361.94,
    "end": 365.98,
    "text": "So is machine learning better so to say than until that time?"
  },
  {
    "start": 366.4,
    "end": 367.82,
    "text": "just looking at condition monitoring?"
  },
  {
    "start": 367.88,
    "end": 373.42,
    "text": "Condition monitoring... You put a threshold level at whatever temperature and if it passes now."
  },
  {
    "start": 373.8,
    "end": 374.48,
    "text": "That's what we did."
  },
  {
    "start": 374.84,
    "end": 385.66,
    "text": "Now as soon as then A component of the testing machine The new component arrived We found out our model Did not work anymore."
  },
  {
    "start": 386.78,
    "end": 389.7,
    "text": "There is a component, but it's kind of unique."
  },
  {
    "start": 389.94,
    "end": 398.94,
    "text": "And of course... It needs to have the same spec But there are somewhere small pieces and the model would all the time come out with With a false alarm right?"
  },
  {
    "start": 399.6,
    "end": 404.76,
    "text": "So that what we've ...and That's The Kind Of Story We've Been Hearing All The Time."
  },
  {
    "start": 405.36,
    "end": 406.82,
    "text": "I Think Today Were Going To Be Listening."
  },
  {
    "start": 407.1,
    "end": 410.3,
    "text": "Is That Correct To Me?"
  },
  {
    "start": 410.46,
    "end": 422.12,
    "text": "Talking To Stefan And He Will Be Talking along the same lines why predictive maintenance has been so difficult, so cumbersome in the first ten years."
  },
  {
    "start": 422.26,
    "end": 429.72,
    "text": "Now coming to that point and I did share this example with Zepp as I was in the room with Seb."
  },
  {
    "start": 430.0,
    "end": 436.52,
    "text": "So we always are there around eight, nine persons attendees and you and i go round a little bit."
  },
  {
    "start": 437.1,
    "end": 446.76,
    "text": "And I did it for that very reason that yes, Seb confirmed Yes today You would have just plugged-in The data into Tyrex too."
  },
  {
    "start": 447.68,
    "end": 453.98,
    "text": "After some time... ...the model will start understanding what its sees."
  },
  {
    "start": 454.08,
    "end": 458.4,
    "text": "so no need as we needed to do almost ten years ago."
  },
  {
    "start": 459.2,
    "end": 463.1,
    "text": "We need to wait for another three months or four, five months I don't recall."
  },
  {
    "start": 463.56,
    "end": 466.48,
    "text": "and we needed get our new data... ...we needed build a new model."
  },
  {
    "start": 466.76,
    "end": 467.14,
    "text": "only then."
  },
  {
    "start": 467.26,
    "end": 469.0,
    "text": "so we lost like half of year."
  },
  {
    "start": 469.42,
    "end": 473.42,
    "text": "now that is the past revolution."
  },
  {
    "start": 473.68,
    "end": 479.24,
    "text": "So i wouldn't suggest best-of-the show award."
  },
  {
    "start": 480.6,
    "end": 482.06,
    "text": "Yeah, maybe there's a second thought."
  },
  {
    "start": 483.22,
    "end": 494.92,
    "text": "Maybe that would go to Tyrex too but it is really SEP who I saw him doing similar kind of discussion presentation four times in total"
  },
  {
    "start": 495.34,
    "end": 495.6,
    "text": "and"
  },
  {
    "start": 496.42,
    "end": 497.4,
    "text": "It's just amazing."
  },
  {
    "start": 498.08,
    "end": 507.42,
    "text": "you know we can have this guy Who is one the world's top ten AI researchers right?"
  },
  {
    "start": 508.66,
    "end": 520.58,
    "text": "with us two days in the Alps and provide each of our thirty plus attendees one-and-a-half hours, a small group interacting as much time that they wanted during dinners."
  },
  {
    "start": 521.72,
    "end": 523.159,
    "text": "So thank you very much Seb!"
  },
  {
    "start": 524.08,
    "end": 524.68,
    "text": "Yeah perfect."
  },
  {
    "start": 524.72,
    "end": 531.86,
    "text": "so my highlight was little bit different because we had track on AI engineering by Christian Heining."
  },
  {
    "start": 532.32,
    "end": 535.0,
    "text": "he's newbie being host for event."
  },
  {
    "start": 537.52,
    "end": 545.54,
    "text": "It was very interesting because the focus shifted a bit to the question, what will my organization look like in the future?"
  },
  {
    "start": 545.6,
    "end": 551.54,
    "text": "So no more technical discussion about which tool how to integrate cat cam."
  },
  {
    "start": 552.02,
    "end": 553.08,
    "text": "More about the topic."
  },
  {
    "start": 553.16,
    "end": 556.14,
    "text": "that team of one How realistic is that?"
  },
  {
    "start": 556.24,
    "end": 559.12,
    "text": "and how do I build The knowledge enterprise?"
  },
  {
    "start": 559.2,
    "end": 560.14,
    "text": "And there are two."
  },
  {
    "start": 560.24,
    "end": 575.7,
    "text": "memory plays A major role and we were doing a separate episode on this topic i think Because Everybody was really impressed by this session, thinking about the organization and how AI engineering organizations look like in the future."
  },
  {
    "start": 575.74,
    "end": 576.52,
    "text": "Very impressive!"
  },
  {
    "start": 576.92,
    "end": 577.94,
    "text": "Thanks a lot to Christian."
  },
  {
    "start": 578.02,
    "end": 583.08,
    "text": "it's very cool to have you on board and be our host for this session."
  },
  {
    "start": 583.62,
    "end": 601.08,
    "text": "It is interesting that their whole topic shifted from technical discussion into an organizational discussion In that track which took place in the cellar of the building, Which was very on-the-kitchen cellar right?"
  },
  {
    "start": 601.58,
    "end": 605.98,
    "text": "With a wonderful cool climate because upstairs sometimes it Was a little bit warm."
  },
  {
    "start": 606.36,
    "end": 606.86,
    "text": "Yeah We had."
  },
  {
    "start": 607.26,
    "end": 608.2,
    "text": "we had a similar."
  },
  {
    "start": 608.46,
    "end": 610.38,
    "text": "I guess i feel like a similar discussion."
  },
  {
    "start": 611.0,
    "end": 618.2,
    "text": "I think The question was raised if If they did the product management job would continue to exist."
  },
  {
    "start": 618.26,
    "end": 629.24,
    "text": "but I Think the meaning I felt like it's what Andrew Wang, now this time again in a positive sense because... ...I believe that he was one of the first maybe already six months ago."
  },
  {
    "start": 630.0,
    "end": 631.54,
    "text": "He was suggesting that."
  },
  {
    "start": 632.14,
    "end": 643.86,
    "text": "because so what follows after you start with the designer of a product, then the engineer and maybe that goes together or the product manager will decide what it is he wants to build."
  },
  {
    "start": 643.9,
    "end": 656.22,
    "text": "Then their designer and they'll pass this on may be for the software engineer because the software engineering almost like going three four five times as fast."
  },
  {
    "start": 656.4,
    "end": 662.18,
    "text": "you know in the past You don't have enough product management or designing capabilities."
  },
  {
    "start": 662.74,
    "end": 663.88,
    "text": "I think was discussion."
  },
  {
    "start": 664.46,
    "end": 668.78,
    "text": "If you're going to be doing a podcast with Chris on the topic, I'm very much looking forward."
  },
  {
    "start": 669.72,
    "end": 671.02,
    "text": "The idea is right."
  },
  {
    "start": 671.36,
    "end": 671.98,
    "text": "so who?"
  },
  {
    "start": 672.76,
    "end": 674.5,
    "text": "Who Is the originating person?"
  },
  {
    "start": 674.62,
    "end": 687.56,
    "text": "that it's the thought i believe right and You could Be one Person show for example in A big organization maybe instead of a hundred People but your Moving Capabilities through the Front."
  },
  {
    "start": 687.78,
    "end": 689.54,
    "text": "That what i Believe It is Right."
  },
  {
    "start": 689.62,
    "end": 695.14,
    "text": "So if you know What you want To do As a one-person company or as a ten person come doesn't matter."
  },
  {
    "start": 695.9,
    "end": 707.8,
    "text": "you are going to be able To do so many more things that in the past people after You know like the engineers would be doing and use of design at which maybe has never done any coding."
  },
  {
    "start": 707.88,
    "end": 711.5,
    "text": "and that's what I'm going to be talking in a couple of days with Scott Duncan."
  },
  {
    "start": 712.06,
    "end": 714.88,
    "text": "He is the guy, I believe chemical industry."
  },
  {
    "start": 715.46,
    "end": 717.78,
    "text": "he has never written a code or an agent."
  },
  {
    "start": 717.88,
    "end": 721.14,
    "text": "so maybe this example looking forward too."
  },
  {
    "start": 721.2,
    "end": 722.28,
    "text": "yet it was great session."
  },
  {
    "start": 723.06,
    "end": 728.12,
    "text": "And then there were another discussion about KPIs and AI engineering."
  },
  {
    "start": 728.62,
    "end": 729.8,
    "text": "What are the KPIs?"
  },
  {
    "start": 730.58,
    "end": 733.9,
    "text": "token consumption, but we had no answer at the end."
  },
  {
    "start": 733.98,
    "end": 736.7,
    "text": "So we also need to record an episode on KPIs."
  },
  {
    "start": 736.86,
    "end": 742.74,
    "text": "I think yeah well i mean i would say that the tokens are on a negative side."
  },
  {
    "start": 743.1,
    "end": 743.88,
    "text": "what's what i say?"
  },
  {
    "start": 744.78,
    "end": 754.48,
    "text": "In the end you have to do something like produce something and needs high quality all these good things right in relation to cost."
  },
  {
    "start": 754.7,
    "end": 761.64,
    "text": "so the tokens our costs yes maybe a couple of weeks only, couple months."
  },
  {
    "start": 761.9,
    "end": 762.72,
    "text": "Only that."
  },
  {
    "start": 762.96,
    "end": 770.3,
    "text": "yes there is price to be paid for high quality artificial intelligence language models."
  },
  {
    "start": 770.52,
    "end": 778.98,
    "text": "agents can be may not but the discussion on it's below equation right?"
  },
  {
    "start": 779.54,
    "end": 780.88,
    "text": "Yes and positive side."
  },
  {
    "start": 782.44,
    "end": 785.3,
    "text": "In the past, we were looking at writing codes."
  },
  {
    "start": 785.4,
    "end": 796.68,
    "text": "We said that a million lines of code produced when we did it on the IBM mainframe or whatever which is in itself never means quality right?"
  },
  {
    "start": 796.74,
    "end": 807.32,
    "text": "You can write what you want and I think we are moving these days towards accomplishment which is worthwhile, right?"
  },
  {
    "start": 807.46,
    "end": 808.36,
    "text": "Which has value."
  },
  {
    "start": 808.48,
    "end": 814.88,
    "text": "So maybe it's the value I guess in the end relation to the yeah but we'll find out."
  },
  {
    "start": 815.82,
    "end": 820.22,
    "text": "you're thinking You gonna join that discussion also with Christian then or have a going on?"
  },
  {
    "start": 820.26,
    "end": 820.92,
    "text": "do a separate one."
  },
  {
    "start": 821.32,
    "end": 827.48,
    "text": "Yeah Maybe i will Find somebody else who can tell us A little bit about how To measure."
  },
  {
    "start": 827.6,
    "end": 830.32,
    "text": "what are KPIs What our realistic KPIs?"
  },
  {
    "start": 830.44,
    "end": 834.66,
    "text": "Do We Have KPIs How to develop KPIs Something like this."
  },
  {
    "start": 835.38,
    "end": 836.62,
    "text": "So we cannot name all."
  },
  {
    "start": 836.78,
    "end": 837.8,
    "text": "furthest away was."
  },
  {
    "start": 838.16,
    "end": 839.28,
    "text": "But I have one more topic."
  },
  {
    "start": 839.34,
    "end": 854.74,
    "text": "when it comes to AI and engineering in tools, because I shared a picture of all our flip charts on LinkedIn And then a colleague of mine who is not able to join AI in the ups this year he wrote me what's app?"
  },
  {
    "start": 855.4,
    "end": 856.86,
    "text": "He said wow that very cool."
  },
  {
    "start": 857.02,
    "end": 862.5,
    "text": "i already put it on my AI analyzing your results."
  },
  {
    "start": 863.28,
    "end": 864.08,
    "text": "Thanks alot for that."
  },
  {
    "start": 864.4,
    "end": 867.36,
    "text": "So we need to be careful at the end, Peter."
  },
  {
    "start": 868.1,
    "end": 874.26,
    "text": "Only sharing the flip charts and people are using their AI... ...to analyze the flip chart."
  },
  {
    "start": 875.32,
    "end": 876.28,
    "text": "Greetings to Hans-Michel!"
  },
  {
    "start": 876.86,
    "end": 878.3,
    "text": "Great approach great idea."
  },
  {
    "start": 878.68,
    "end": 889.0,
    "text": "see you next time on AI On The Volcano And of course, we're very much looking forward to these results that he has received from whatever large language model."
  },
  {
    "start": 889.14,
    "end": 889.52,
    "text": "Exactly!"
  },
  {
    "start": 889.62,
    "end": 898.68,
    "text": "Now I think the yeah and i think it's perfectly okay too still share the flip charts which is the only thing that we write down."
  },
  {
    "start": 898.74,
    "end": 908.82,
    "text": "We don't do power points just remember...I think in the end my feeling is the ultimate winners always but in general, are the attendees."
  },
  {
    "start": 909.08,
    "end": 920.98,
    "text": "I mean especially those from small and mid-sized companies In addition to those of big companies Big companies like Boris, you know Boris who received a lot of interest for his book."
  },
  {
    "start": 921.34,
    "end": 924.56,
    "text": "but the guys and girls I come back to that are two women."
  },
  {
    "start": 924.64,
    "end": 927.32,
    "text": "we need more woman in The future."
  },
  {
    "start": 927.38,
    "end": 932.46,
    "text": "so dear Women female listeners listen to what?"
  },
  {
    "start": 932.5,
    "end": 933.96,
    "text": "We're going to be doing in the future."
  },
  {
    "start": 935.34,
    "end": 938.3,
    "text": "But those of the big companies they have many colleagues."
  },
  {
    "start": 938.4,
    "end": 939.84,
    "text": "You know from around the world."
  },
  {
    "start": 941.34,
    "end": 945.82,
    "text": "they can much more easily within their big company, you know."
  },
  {
    "start": 945.96,
    "end": 946.2,
    "text": "check."
  },
  {
    "start": 946.34,
    "end": 947.4,
    "text": "where am I going?"
  },
  {
    "start": 947.48,
    "end": 950.06,
    "text": "Where are my colleagues go and whereas the United States gone."
  },
  {
    "start": 950.1,
    "end": 951.24,
    "text": "Whereas China going etc."
  },
  {
    "start": 952.14,
    "end": 955.72,
    "text": "those of the smaller companies i feel that there."
  },
  {
    "start": 956.96,
    "end": 970.04,
    "text": "so happy with being able to meet face-to-face in this sessions but also into breaks uh during the meals and talk to you know, talked with other people."
  },
  {
    "start": 970.16,
    "end": 973.58,
    "text": "Talked to their attendees And at the end they go home?"
  },
  {
    "start": 973.92,
    "end": 976.3,
    "text": "That's what we always say that your quotes right."
  },
  {
    "start": 976.54,
    "end": 978.4,
    "text": "learn new People learn new things Right!"
  },
  {
    "start": 981.14,
    "end": 984.42,
    "text": "Okay, and they feel like oh yeah I'm spot-on."
  },
  {
    "start": 984.62,
    "end": 985.92,
    "text": "I am doing everything right."
  },
  {
    "start": 986.22,
    "end": 991.36,
    "text": "or They say okay ABC is good But i've learned you know XYZ."
  },
  {
    "start": 991.7,
    "end": 994.08,
    "text": "in addition And I need to do something else."
  },
  {
    "start": 994.14,
    "end": 1000.76,
    "text": "when I wanna and think that all go home with thoughts and ideas of You Know what it is today?"
  },
  {
    "start": 1000.96,
    "end": 1003.64,
    "text": "That are going to be doing from now on Perfect."
  },
  {
    "start": 1003.7,
    "end": 1017.42,
    "text": "Because you mentioned China and the US, we have had a very interesting session on sovereignty by Robert Josich from Schwarz but this sovereignty is exhausting so it's really difficult right?"
  },
  {
    "start": 1017.54,
    "end": 1020.3,
    "text": "So their whole discussion what to do?"
  },
  {
    "start": 1021.08,
    "end": 1026.42,
    "text": "I think we raised some questions or raise ideas in these sessions But its'a long way to go!"
  },
  {
    "start": 1027.48,
    "end": 1036.48,
    "text": "Everybody was very keen for going like that And it was the same when I was in Robert's session."
  },
  {
    "start": 1036.88,
    "end": 1044.44,
    "text": "The main thing that took away, and i believe everybody agreed is you try to maximize sovereignty."
  },
  {
    "start": 1045.16,
    "end": 1050.72,
    "text": "So because we were looking for You know, what is it like?"
  },
  {
    "start": 1051.2,
    "end": 1053.8,
    "text": "Maybe I'm looking not for rules."
  },
  {
    "start": 1053.9,
    "end": 1054.56,
    "text": "What does it like?"
  },
  {
    "start": 1055.26,
    "end": 1055.98,
    "text": "suggestions?"
  },
  {
    "start": 1056.06,
    "end": 1056.38,
    "text": "you Know what?"
  },
  {
    "start": 1056.42,
    "end": 1057.04,
    "text": "a suggestion?"
  },
  {
    "start": 1057.2,
    "end": 1059.68,
    "text": "one two three four five and we could come up with some."
  },
  {
    "start": 1060.2,
    "end": 1061.68,
    "text": "We talked about open source."
  },
  {
    "start": 1061.76,
    "end": 1075.16,
    "text": "We we talked about fallback solutions all those wonderful things But but was very clear that none of us And that's why i think we agreed Is that we cannot have hundred percent sovereignty."
  },
  {
    "start": 1075.76,
    "end": 1080.14,
    "text": "so you can at best Strive to maximize sovereignty."
  },
  {
    "start": 1080.64,
    "end": 1086.4,
    "text": "So you try to go from whatever is seventy two eighty percent or maybe eighty to ninety percent and that feels very much."
  },
  {
    "start": 1086.54,
    "end": 1088.24,
    "text": "Maybe, uh That's the thought I have now."
  },
  {
    "start": 1088.36,
    "end": 1089.6,
    "text": "i didn't share it at a time."
  },
  {
    "start": 1090.04,
    "end": 1093.32,
    "text": "It's Very Much The Same Like The OE Overall Equipment Efficiency."
  },
  {
    "start": 1093.4,
    "end": 1098.0,
    "text": "You Know There Is A Theoretical One Hundred Percent What Comes Out Of Your Factory?"
  },
  {
    "start": 1098.46,
    "end": 1101.56,
    "text": "You know one Hundred Cars At A hundred Percent Quality."
  },
  {
    "start": 1102.24,
    "end": 1103.58,
    "text": "You don't get that, typically."
  },
  {
    "start": 1104.54,
    "end": 1118.9,
    "text": "In what we do producing goods maybe you got ninety percent whatever in process typically last sixty-seventy percent and maybe it's similar... ...you strive to get a hundred percent but today you can not get one hundred percent."
  },
  {
    "start": 1118.98,
    "end": 1121.86,
    "text": "yet I agree its rather complex topic."
  },
  {
    "start": 1122.68,
    "end": 1130.3,
    "text": "It is little bit different from the things we have been doing before And i guess were going come back for this topic."
  },
  {
    "start": 1130.88,
    "end": 1134.84,
    "text": "Yeah, but we had a discussion because my opinion is sure."
  },
  {
    "start": 1135.86,
    "end": 1137.06,
    "text": "We lack balance right?"
  },
  {
    "start": 1137.26,
    "end": 1142.58,
    "text": "We lack what China and the US absolutely want except ASML for example."
  },
  {
    "start": 1143.08,
    "end": 1150.7,
    "text": "And I think What what should we invest and what kind of technology?"
  },
  {
    "start": 1150.84,
    "end": 1155.14,
    "text": "Should we invent that the Chinese in the US absolutely want?"
  },
  {
    "start": 1155.24,
    "end": 1157.48,
    "text": "so then We get back as a balance, right."
  },
  {
    "start": 1157.58,
    "end": 1162.7,
    "text": "So The Chinese won't X we have acts we want Y they have."
  },
  {
    "start": 1162.86,
    "end": 1165.08,
    "text": "why do you guys have B?"
  },
  {
    "start": 1165.56,
    "end": 1166.3,
    "text": "And we won't be."
  },
  {
    "start": 1166.6,
    "end": 1169.7,
    "text": "but they also want A from U from us In Europe."
  },
  {
    "start": 1169.76,
    "end": 1177.68,
    "text": "so they get a. so in the moment there's this there is no balance between big parts in the world."
  },
  {
    "start": 1178.14,
    "end": 1182.28,
    "text": "And I think maybe it's a topic for regulation, right?"
  },
  {
    "start": 1182.36,
    "end": 1192.46,
    "text": "So we also talked about regulation but or Maybe It is also a topic to invent something To invent In Europe What The Others Are Missing!"
  },
  {
    "start": 1192.88,
    "end": 1198.46,
    "text": "so i got some support from this thesis by Werner Foulhaber From ABWUG Let´s see."
  },
  {
    "start": 1198.64,
    "end": 1208.74,
    "text": "But as you mentioned A very difficult discussion But there's two, at least to it as many more probably but there is another perspective."
  },
  {
    "start": 1208.8,
    "end": 1217.36,
    "text": "So the one that you just talked about what shall we and in this case you mean Europe I guess do right which is very political?"
  },
  {
    "start": 1217.44,
    "end": 1222.3,
    "text": "And all of our... You talk about three big ones."
  },
  {
    "start": 1222.4,
    "end": 1236.2,
    "text": "well actually for us AIs concerned maybe its only too big once a day But typically we don't get it to the street, right?"
  },
  {
    "start": 1237.06,
    "end": 1242.42,
    "text": "The second perspective which is what I was very much interested in giving our listeners."
  },
  {
    "start": 1242.66,
    "end": 1254.34,
    "text": "Giving people and companies who struggle with exactly same thing not about Who am I going to vote when i can vote in democratic Europe?"
  },
  {
    "start": 1254.8,
    "end": 1258.6,
    "text": "And what is my politician gonna do as far as AI's concerned."
  },
  {
    "start": 1258.64,
    "end": 1263.92,
    "text": "That was one thing, but the other thing Is What Can I Do Today and What Should I Be Doing Tomorrow?"
  },
  {
    "start": 1264.56,
    "end": 1270.86,
    "text": "So that is a approach which I Was Very Much Interested In But Maybe The Other One We Do I Believe."
  },
  {
    "start": 1270.94,
    "end": 1277.9,
    "text": "Anyway All Time We Look At You Know What Is Coming Out Of China A Completely Different Approach To Large Language Models."
  },
  {
    "start": 1278.6,
    "end": 1280.44,
    "text": "Yeah, what's happening in the United States?"
  },
  {
    "start": 1280.54,
    "end": 1282.84,
    "text": "Oh suddenly we can't use oh In the meantime."
  },
  {
    "start": 1282.88,
    "end": 1286.06,
    "text": "We can use entropic fable again right What's happening there?"
  },
  {
    "start": 1286.44,
    "end": 1293.4,
    "text": "so those things we talk about all And those are in the end, they're becoming so political that it's yeah."
  },
  {
    "start": 1293.56,
    "end": 1298.66,
    "text": "I mean we can and we shall talk about but uh... We cannot make changes today."
  },
  {
    "start": 1299.08,
    "end": 1302.5,
    "text": "you know things that were going to talk about today and then next couple of months."
  },
  {
    "start": 1302.6,
    "end": 1308.06,
    "text": "if thats what we want there may be gonna influence politicians in y'know the necks."
  },
  {
    "start": 1308.44,
    "end": 1311.8,
    "text": "one year maybe five years or we'll see any change."
  },
  {
    "start": 1312.44,
    "end": 1313.08,
    "text": "i really believe so."
  },
  {
    "start": 1313.5,
    "end": 1326.1,
    "text": "The other thing though which is very important You listeners, you within your companies more so even the smaller and the mid ones again because you don't have So many colleagues who you can talk with."
  },
  {
    "start": 1327.78,
    "end": 1335.7,
    "text": "Maybe we should do that also Robert at some point in time to tell them what can you consider doing."
  },
  {
    "start": 1336.06,
    "end": 1339.4,
    "text": "Open source is one example, but not absolute right?"
  },
  {
    "start": 1339.52,
    "end": 1347.72,
    "text": "What's the advantage maybe also disadvantage of going open-source and all other things I'm doing."
  },
  {
    "start": 1348.86,
    "end": 1356.2,
    "text": "I also want to mention Joshi, the owner of the Rote Wand which means Red Wall Hotel."
  },
  {
    "start": 1356.46,
    "end": 1367.68,
    "text": "Which you know since as a kid he showed us Friday afternoon... ...a walk that we typically do who brought it to see water-as-energy based wood saw."
  },
  {
    "start": 1367.78,
    "end": 1369.78,
    "text": "That was really great!"
  },
  {
    "start": 1370.5,
    "end": 1379.18,
    "text": "To put nature in perspective on technology for a couple of days, right?"
  },
  {
    "start": 1379.66,
    "end": 1380.68,
    "text": "And what's the sawmill."
  },
  {
    "start": 1380.78,
    "end": 1385.26,
    "text": "Right so from exactly seventeen ninety seven."
  },
  {
    "start": 1385.3,
    "end": 1386.08,
    "text": "So yeah, wood saw."
  },
  {
    "start": 1386.12,
    "end": 1388.12,
    "text": "Wood saw."
  },
  {
    "start": 1390.16,
    "end": 1390.38,
    "text": "Yeah."
  },
  {
    "start": 1391.32,
    "end": 1411.24,
    "text": "It looks like the cloister people as they maybe putting it all together you and I believe being very happy with what we've been doing since and I think you wrote on seven years, i looked at it with the first podcast within January."
  },
  {
    "start": 1411.36,
    "end": 1425.8,
    "text": "nineteen in two thousand nineteen uh... in that one we explained a format which shared what our listeners quote-unquote why AI is and will continue to be so important for industry."
  },
  {
    "start": 1426.02,
    "end": 1428.16,
    "text": "That's what was said more than eight years ago."
  },
  {
    "start": 1428.9,
    "end": 1433.9,
    "text": "So ah.. In all humbleness ,I Think We Were Quite Visionary."
  },
  {
    "start": 1435.06,
    "end": 1439.86,
    "text": "The AI and X gatherings starting with Wurzburg, they came later."
  },
  {
    "start": 1440.86,
    "end": 1446.1,
    "text": "We should also say thank you to Barbara who has been one of the three of us."
  },
  {
    "start": 1446.18,
    "end": 1461.06,
    "text": "that we started these gatherings together was on site for a couple years And I think we provided astonishing value for the attendees which have in the meantime then become an integral part this network which I just mentioned, you know?"
  },
  {
    "start": 1461.16,
    "end": 1462.92,
    "text": "Just feeling... You know."
  },
  {
    "start": 1463.08,
    "end": 1467.16,
    "text": "I'm in sync with the other people in the industrial AI world."
  },
  {
    "start": 1467.9,
    "end": 1471.94,
    "text": "now i think there's still enough more than enough industry topics."
  },
  {
    "start": 1472.22,
    "end": 1475.76,
    "text": "so as far as i'm concerned Robert let's do some more."
  },
  {
    "start": 1475.84,
    "end": 1477.38,
    "text": "maybe you want to share Let's do it!"
  },
  {
    "start": 1477.8,
    "end": 1481.86,
    "text": "I have one more tooltip because i think Christian shared this tip with others."
  },
  {
    "start": 1482.38,
    "end": 1489.5,
    "text": "It is called Sakana AI from Japan And it's a multi-agent orchestration and model connections."
  },
  {
    "start": 1490.02,
    "end": 1491.78,
    "text": "So maybe try out ZAKANA."
  },
  {
    "start": 1491.92,
    "end": 1502.56,
    "text": "AI sounds very interesting, I already did an interview with Jakub Tomschak on ZAKARA and he also said yeah... A very interesting approach was released in June this year."
  },
  {
    "start": 1503.06,
    "end": 1504.9,
    "text": "It is called Fugu the Model."
  },
  {
    "start": 1506.08,
    "end": 1506.82,
    "text": "Maybe tryout."
  },
  {
    "start": 1506.96,
    "end": 1507.9,
    "text": "an interesting approach."
  },
  {
    "start": 1508.26,
    "end": 1509.44,
    "text": "Peter thanks alot."
  },
  {
    "start": 1509.52,
    "end": 1513.36,
    "text": "but what are our event agenda in the next coming months?"
  },
  {
    "start": 1513.42,
    "end": 1514.6,
    "text": "because we planned something."
  },
  {
    "start": 1514.86,
    "end": 1519.84,
    "text": "Yeah, but you know better than I think The Next one is AI on a volcano right?"
  },
  {
    "start": 1519.94,
    "end": 1522.32,
    "text": "Okay okay A i On the Volcano."
  },
  {
    "start": 1522.52,
    "end": 1527.28,
    "text": "yeah so everybody was ask us what does it mean AI on the volcano?"
  },
  {
    "start": 1527.48,
    "end": 1531.56,
    "text": "and yet there Is a volcano In Parkstein in Vavaria."
  },
  {
    "start": 1531.82,
    "end": 1549.54,
    "text": "And uh There's a company based at Parkstein Interlogistics companies market leader when it comes to retail logistics and they invited us in October To run our event there AI on the volcano."
  },
  {
    "start": 1549.78,
    "end": 1562.44,
    "text": "And then we plan a trip to Paris visiting this station F The startup ecosystem In France, that was also discussion of our sovereignty track about France."
  },
  {
    "start": 1563.46,
    "end": 1567.52,
    "text": "the French government is doing differently to German government, but this another story."
  },
  {
    "start": 1568.08,
    "end": 1574.28,
    "text": "But yeah if you want join apply send us an email and we will see what's possible."
  },
  {
    "start": 1574.62,
    "end": 1579.74,
    "text": "Was there a potential alternative to Paris discussed or suggested war?"
  },
  {
    "start": 1579.88,
    "end": 1581.12,
    "text": "We still wanna leave that open."
  },
  {
    "start": 1581.28,
    "end": 1585.14,
    "text": "Yeah I'm not."
  },
  {
    "start": 1585.78,
    "end": 1587.36,
    "text": "Yeah, it's a surprise."
  },
  {
    "start": 1587.44,
    "end": 1587.98,
    "text": "maybe yeah."
  },
  {
    "start": 1588.26,
    "end": 1592.52,
    "text": "Maybe we take another plane different planes at the airport right?"
  },
  {
    "start": 1593.18,
    "end": 1594.5,
    "text": "Yes there are still."
  },
  {
    "start": 1595.16,
    "end": 1598.5,
    "text": "There is enough places in Europe."
  },
  {
    "start": 1598.56,
    "end": 1600.14,
    "text": "We don't have to fly to Japan."
  },
  {
    "start": 1600.2,
    "end": 1601.38,
    "text": "We can do this sometime as well."
  },
  {
    "start": 1602.46,
    "end": 1605.1,
    "text": "But we do still have enough cities."
  },
  {
    "start": 1605.24,
    "end": 1616.78,
    "text": "I believe for the next couple of years and As us were going there was gonna be I believe more areas conglomerates, maybe developing artificial intelligence."
  },
  {
    "start": 1617.0,
    "end": 1626.44,
    "text": "Yeah I got the hint by a guy and he said you should fly to Varsha not go west but east because there's also lots happening in the East."
  },
  {
    "start": 1626.76,
    "end": 1628.1,
    "text": "Yeah, yeah."
  },
  {
    "start": 1628.2,
    "end": 1641.22,
    "text": "In general I would say if you look on a European level where things are moving there is a lot of stuff happening in the shall i say more developed from the past kind cities."
  },
  {
    "start": 1641.78,
    "end": 1648.48,
    "text": "know If we looked robotics or we're saying Munich, Zurich just an example There's another city at some point that."
  },
  {
    "start": 1649.08,
    "end": 1656.58,
    "text": "But having said that, and not only I mean it more from the general development within Europe."
  },
  {
    "start": 1656.78,
    "end": 1667.32,
    "text": "The Eastern countries are building and moving forward in becoming more important in general within Europe all of time."
  },
  {
    "start": 1667.4,
    "end": 1668.76,
    "text": "so let's look at those as well."
  },
  {
    "start": 1669.6,
    "end": 1670.84,
    "text": "Thank you Robert!"
  },
  {
    "start": 1671.0,
    "end": 1673.82,
    "text": "Thank you attendees thank you dear listeners."
  },
  {
    "start": 1673.92,
    "end": 1676.12,
    "text": "maybe want to join us next time looking forward too."
  },
  {
    "start": 1676.72,
    "end": 1676.98,
    "text": "Thank you."
  },
  {
    "start": 1677.22,
    "end": 1682.78,
    "text": "And now we switch to the main part with Stefan Zuvoloch, he was also an attendee at our AI and E-Apps event."
  },
  {
    "start": 1682.86,
    "end": 1684.7,
    "text": "so yeah... Yeah!"
  },
  {
    "start": 1684.76,
    "end": 1691.36,
    "text": "The first thing that your gonna hear is that I'm asking Stefan because there's this story when in the past i would say.. You know what?"
  },
  {
    "start": 1691.4,
    "end": 1694.16,
    "text": "We're going be doing whatever software engineer."
  },
  {
    "start": 1694.2,
    "end": 1695.68,
    "text": "We are going to do agents."
  },
  {
    "start": 1696.06,
    "end": 1699.02,
    "text": "then you dear listener, the engineer will come to your customer."
  },
  {
    "start": 1699.66,
    "end": 1706.98,
    "text": "they'll say Robert or Peter said can let me have a agent Stefan."
  },
  {
    "start": 1707.9,
    "end": 1715.36,
    "text": "He would be not happy sometimes, I believe and until a year ago it was like Peter don't say things like that."
  },
  {
    "start": 1715.46,
    "end": 1723.22,
    "text": "or maybe now Stefan has turned around one eighty degrees in the time okay?"
  },
  {
    "start": 1723.26,
    "end": 1729.98,
    "text": "That is the question i'm going to start and Stefan will explain us why he has turned."
  },
  {
    "start": 1733.02,
    "end": 1734.82,
    "text": "Talk to you soon, bye-bye Robert."
  },
  {
    "start": 1736.54,
    "end": 1740.4,
    "text": "Today I'm going to be talking once more to Stefan Zulak."
  },
  {
    "start": 1740.46,
    "end": 1752.84,
    "text": "he's the co-founder at Renumix and he and i today are going to talk about who would have thought that possible agentic AI in engineering."
  },
  {
    "start": 1752.94,
    "end": 1753.58,
    "text": "Hi Stefan!"
  },
  {
    "start": 1753.9,
    "end": 1755.54,
    "text": "Hey Peter thanks for having me back."
  },
  {
    "start": 1755.66,
    "end": 1757.24,
    "text": "uh you're welcome."
  },
  {
    "start": 1757.34,
    "end": 1764.5,
    "text": "yeah the one or the other listener may actually recognize your name or your voice last time you were here."
  },
  {
    "start": 1764.6,
    "end": 1766.84,
    "text": "that's about what do you think how long ago?"
  },
  {
    "start": 1766.88,
    "end": 1767.04,
    "text": "i don't"
  },
  {
    "start": 1768.08,
    "end": 1768.56,
    "text": "know."
  },
  {
    "start": 1769.14,
    "end": 1773.8,
    "text": "maybe like something feels like a very different world."
  },
  {
    "start": 1773.92,
    "end": 1782.38,
    "text": "I think we were talking about data-centric AI and all the things, now that we have completely different topics... Very interesting!"
  },
  {
    "start": 1782.44,
    "end": 1785.96,
    "text": "A lot has happened in those two years here."
  },
  {
    "start": 1786.3,
    "end": 1789.68,
    "text": "so nevertheless please introduce yourself to our listeners."
  },
  {
    "start": 1790.68,
    "end": 1791.98,
    "text": "Yeah, I'm Stefan Zuvlak."
  },
  {
    "start": 1792.02,
    "end": 1794.02,
    "text": "I am one of the co-founders of Renumix."
  },
  {
    "start": 1794.78,
    "end": 1805.82,
    "text": "My academic background is actually in like medical engineering and then we founded Renumics in two thousand seventeen with a goal of bringing machine learning And AI to engineering."
  },
  {
    "start": 1805.88,
    "end": 1806.1,
    "text": "very"
  },
  {
    "start": 1806.16,
    "end": 1815.92,
    "text": "good Stefan you and i talking agentic who would have thought so let me give quick explanation why I say so."
  },
  {
    "start": 1816.68,
    "end": 1821.54,
    "text": "So please hang in there for a minute, but it was the time when i would look into the future."
  },
  {
    "start": 1821.72,
    "end": 1827.92,
    "text": "actually always do that believe and uh... an eye talk about how we engineers on one of those as well."
  },
  {
    "start": 1828.08,
    "end": 1832.58,
    "text": "we should eat our own dog food right by and i think this once specifically."
  },
  {
    "start": 1832.64,
    "end": 1846.06,
    "text": "what's about having copilot support us coding and how in the not so far future personal business assistants, you know they would have agents do whatever we would've asked them."
  },
  {
    "start": 1846.1,
    "end": 1857.32,
    "text": "And then I almost felt guilty for being like a smart ass i think And say what's coming while you are there."
  },
  {
    "start": 1857.46,
    "end": 1861.82,
    "text": "You dear listener and then I would typically call out your name, I believe?"
  },
  {
    "start": 1861.88,
    "end": 1871.4,
    "text": "I was saying no like guys like Steph on you walk into a customer and they will say oh Peter Seabourg said we're gonna get these agents who can cannot get one as well please."
  },
  {
    "start": 1871.96,
    "end": 1879.3,
    "text": "so always really felt very sorry afterwards for calling out things that I believed, and still believe sometimes."
  },
  {
    "start": 1879.4,
    "end": 1882.46,
    "text": "And i'm not always right so be very careful."
  },
  {
    "start": 1882.98,
    "end": 1885.12,
    "text": "but then I believe we're going to come right?"
  },
  {
    "start": 1886.74,
    "end": 1892.46,
    "text": "You guys girls out there engineers you had to develop them also."
  },
  {
    "start": 1893.02,
    "end": 1896.72,
    "text": "part of it wasn't certain talking about agents."
  },
  {
    "start": 1897.36,
    "end": 1900.02,
    "text": "You know, we're that convinced."
  },
  {
    "start": 1900.14,
    "end": 1901.8,
    "text": "maybe I don't know a year or two years ago."
  },
  {
    "start": 1901.88,
    "end": 1906.76,
    "text": "That agents were going to play your role Or at least maybe you did not always feel happy."
  },
  {
    "start": 1907.08,
    "end": 1909.24,
    "text": "So thanks for hanging in there."
  },
  {
    "start": 1909.88,
    "end": 1915.82,
    "text": "Now We are here again back great To have you and we're talking Agents."
  },
  {
    "start": 1916.08,
    "end": 1918.58,
    "text": "seems you Have come around."
  },
  {
    "start": 1918.84,
    "end": 1919.38,
    "text": "what happened?"
  },
  {
    "start": 1920.88,
    "end": 1923.84,
    "text": "Yeah, I think that's probably a super-super long story."
  },
  {
    "start": 1924.18,
    "end": 1927.64,
    "text": "I try to maybe tell it as short as possible but yeah absolutely!"
  },
  {
    "start": 1929.04,
    "end": 1935.0,
    "text": "One was about one and half years ago... ...I think i was very big skeptic and there are several reasons for this."
  },
  {
    "start": 1935.04,
    "end": 1942.4,
    "text": "The obvious reason is back then the technology just didn't work well enough for real workflows at time."
  },
  {
    "start": 1942.46,
    "end": 1949.64,
    "text": "That was true But what's even more important in the whole field of industry."
  },
  {
    "start": 1950.36,
    "end": 1952.68,
    "text": "we had, I think quite some rough years."
  },
  {
    "start": 1952.84,
    "end": 1956.28,
    "text": "And maybe I can explain it a bit and give them some examples."
  },
  {
    "start": 1956.34,
    "end": 1966.54,
    "text": "so when we started in two thousand seventeen i think this was may be like the first year with Elon Musk promised everybody in six months there will be robot taxis on the road right?"
  },
  {
    "start": 1966.98,
    "end": 1971.66,
    "text": "We distinguish between weak machine learning or strong AI."
  },
  {
    "start": 1971.94,
    "end": 1974.06,
    "text": "actually you don't really use these terms anymore."
  },
  {
    "start": 1974.3,
    "end": 1983.36,
    "text": "at back of time he said hey weak AI is something you cannot just apply to a very narrow use case, I don't know say if there are apples and oranges on an image."
  },
  {
    "start": 1983.46,
    "end": 1987.58,
    "text": "And then strong AI is you really have in AI that has broad knowledge right?"
  },
  {
    "start": 1988.26,
    "end": 1990.4,
    "text": "At the time we only had weak AI."
  },
  {
    "start": 1991.44,
    "end": 2002.12,
    "text": "this was big problem because AI kind of like automation tool for processes but very narrow Very costly implement could fail easily."
  },
  {
    "start": 2003.04,
    "end": 2011.22,
    "text": "so i've seen think alot failures bringing good demos to production, so-to speak and very few success stories."
  },
  {
    "start": 2011.68,
    "end": 2019.32,
    "text": "So I think i was very wary about like the new hype and the new demo that might not fulfill expectations."
  },
  {
    "start": 2019.84,
    "end": 2026.44,
    "text": "And...and..I think the biggest reason is ...and we can go into this in more detail now."
  },
  {
    "start": 2026.74,
    "end": 2031.1,
    "text": "We have strong AI ..And....i actually feel it's underrated!"
  },
  {
    "start": 2031.14,
    "end": 2038.22,
    "text": "I think people underestimate how strong this is, how much background knowledge it has and so on."
  },
  {
    "start": 2038.66,
    "end": 2045.64,
    "text": "And I think that was the big change that i see very clearly now but wasn't aware at the time."
  },
  {
    "start": 2047.08,
    "end": 2047.74,
    "text": "Yeah, you're right."
  },
  {
    "start": 2047.94,
    "end": 2052.86,
    "text": "This divirement between weak and strong AI that's really part of the history of AI."
  },
  {
    "start": 2053.159,
    "end": 2070.56,
    "text": "I would say when i will do maybe up to like ten years ago When it would be giving this introduction And if we get a name there is The terminology If Weak & Strong Is Related To One Specific Person you again, Schmidt Hoover."
  },
  {
    "start": 2072.78,
    "end": 2075.659,
    "text": "Many times it is many things are related to him."
  },
  {
    "start": 2076.159,
    "end": 2079.94,
    "text": "this case there was like maybe Canadian or It's not important."
  },
  {
    "start": 2080.06,
    "end": 2082.179,
    "text": "You're right we don't talk about that anymore."
  },
  {
    "start": 2082.42,
    "end": 2092.54,
    "text": "So in the meantime thanks for that even completely overhauled your engineering software Right earlier this week This year because they fully agending approach works so well."
  },
  {
    "start": 2092.719,
    "end": 2103.16,
    "text": "I understand now We're going to be talking About That In The Moment but For Now Let Me first ask you to share with us what you see as the challenges in engineering."
  },
  {
    "start": 2103.24,
    "end": 2103.32,
    "text": "and"
  },
  {
    "start": 2107.52,
    "end": 2114.9,
    "text": "Yeah, I think obviously engineering especially European industry is in a very rough spot at the moment."
  },
  {
    "start": 2115.1,
    "end": 2117.42,
    "text": "And i think there are many reasons for that."
  },
  {
    "start": 2117.48,
    "end": 2121.66,
    "text": "we don't have to go into all of them but one really big problem"
  },
  {
    "start": 2121.78,
    "end": 2122.22,
    "text": "is that"
  },
  {
    "start": 2122.36,
    "end": 2123.42,
    "text": "In the past right?"
  },
  {
    "start": 2123.5,
    "end": 2139.64,
    "text": "We built complex products like car and did it by divide and conquering divided the cars into smaller components ECUs And now we see that doesn't work any longer."
  },
  {
    "start": 2140.12,
    "end": 2141.48,
    "text": "It's much more important."
  },
  {
    "start": 2141.54,
    "end": 2148.94,
    "text": "to end experience, obviously in automotive you have this software defined car but other industries the same applies as well."
  },
  {
    "start": 2149.48,
    "end": 2154.5,
    "text": "and what do is develop into an engineering culture."
  },
  {
    "start": 2154.54,
    "end": 2162.18,
    "text": "at moment data infrastructure super like whole system is basically built to have little data silos."
  },
  {
    "start": 2162.72,
    "end": 2176.02,
    "text": "And engineers are really like, or at least one and a half years ago they were most of the time building reports and shifting information from one data silo to another but not doing extra work."
  },
  {
    "start": 2176.54,
    "end": 2184.78,
    "text": "I believe that agentic AI is technology we read now unique right now because it's amazing."
  },
  {
    "start": 2184.86,
    "end": 2190.78,
    "text": "at break and backing open these data silos giving everyone access to data, to tools."
  },
  {
    "start": 2191.78,
    "end": 2200.66,
    "text": "To have like every stakeholder gets the ability to see the full picture And I think that says exactly what engineering needs Right"
  },
  {
    "start": 2200.72,
    "end": 2200.84,
    "text": "Now."
  },
  {
    "start": 2201.26,
    "end": 2210.5,
    "text": "Okay sounds almost likely The same problem we hear with all this other jobs as well, you know like medical."
  },
  {
    "start": 2210.58,
    "end": 2212.0,
    "text": "You come from medical technology."
  },
  {
    "start": 2212.38,
    "end": 2213.54,
    "text": "I wasn't aware of that."
  },
  {
    "start": 2213.62,
    "end": 2221.7,
    "text": "we say doctors, they're not doctoring so to speak but writing reports about what it is have been doing."
  },
  {
    "start": 2222.24,
    "end": 2223.06,
    "text": "So in general."
  },
  {
    "start": 2223.14,
    "end": 2224.62,
    "text": "then how does it work?"
  },
  {
    "start": 2224.68,
    "end": 2226.28,
    "text": "Let's go one step further."
  },
  {
    "start": 2226.42,
    "end": 2232.32,
    "text": "i understand That you think In layers Of which the agentic layer would be One."
  },
  {
    "start": 2232.42,
    "end": 2234.52,
    "text": "can you share with us?"
  },
  {
    "start": 2235.78,
    "end": 2239.3,
    "text": "Yeah, I think it makes sense to have like a model and at most people."
  },
  {
    "start": 2239.38,
    "end": 2250.38,
    "text": "I think you use in some sort of others to have basically three layer cakes so-to speak where At the bottom You have your data sources right?"
  },
  {
    "start": 2250.64,
    "end": 2255.22,
    "text": "Your different databases systems but also tools used."
  },
  {
    "start": 2255.64,
    "end": 2265.68,
    "text": "i think its a known problem that at, for example the German Automotive OEM you have like hundreds or even thousands of different tools and databases."
  },
  {
    "start": 2265.82,
    "end": 2268.46,
    "text": "And most often they don't really talk to each other."
  },
  {
    "start": 2268.58,
    "end": 2271.18,
    "text": "You'll get a new tool for specialized work."
  },
  {
    "start": 2272.18,
    "end": 2273.26,
    "text": "We have to standardize this."
  },
  {
    "start": 2273.5,
    "end": 2282.46,
    "text": "This is something on its way now but I feel with technology also talk deep about that, like MCP this will be standardized."
  },
  {
    "start": 2282.52,
    "end": 2302.58,
    "text": "And on top of that you have a new layer and we'll have these agentic layers or agents in network if you'd like an ecosystem of agents who can use the bottom data tool to actually get information execute work do studies."
  },
  {
    "start": 2303.84,
    "end": 2316.12,
    "text": "This is for me the top layer with different teams the engineers, at different stakeholders in projects to get them reports and studies they need."
  },
  {
    "start": 2316.68,
    "end": 2323.02,
    "text": "This constitutes this three-layer thing where the agentic layer is like innovation."
  },
  {
    "start": 2323.12,
    "end": 2333.58,
    "text": "that's a new layer we didn't have before possible to break these data silos where previously this was just impossible."
  },
  {
    "start": 2333.72,
    "end": 2335.52,
    "text": "Very good, and that's actually."
  },
  {
    "start": 2335.56,
    "end": 2337.86,
    "text": "we're probably gonna move into death for a moment."
  },
  {
    "start": 2337.92,
    "end": 2339.56,
    "text": "but let me check one more time."
  },
  {
    "start": 2339.7,
    "end": 2341.98,
    "text": "I like his idea of three layer cake."
  },
  {
    "start": 2343.74,
    "end": 2344.92,
    "text": "you need put it the right way."
  },
  {
    "start": 2346.26,
    "end": 2351.9,
    "text": "everybody is writing about robots which were not doing specifically today."
  },
  {
    "start": 2351.96,
    "end": 2353.18,
    "text": "if do so thats okay."
  },
  {
    "start": 2354.96,
    "end": 2361.1,
    "text": "And then everybody's talking about the components of robots."
  },
  {
    "start": 2361.94,
    "end": 2369.46,
    "text": "Everybody at the same time seems to find out that actuators have about sixty percent of their bill of materials, I believe right?"
  },
  {
    "start": 2369.66,
    "end": 2379.18,
    "text": "Then there is this person writing also making a cake with ten layers but then the robot is at the bottom and I say, understand everything you do."
  },
  {
    "start": 2379.28,
    "end": 2381.42,
    "text": "But why did he put the robot to the bottom?"
  },
  {
    "start": 2382.06,
    "end": 2385.84,
    "text": "And it's that way around right."
  },
  {
    "start": 2386.34,
    "end": 2391.3,
    "text": "so um i follow what It Is That You're Saying and I Understand."
  },
  {
    "start": 2391.4,
    "end": 2398.74,
    "text": "You Say The Agenting Is The Middle One and Then The User Interface Is At Top and I'm With You Exactly In That Thinking."
  },
  {
    "start": 2398.84,
    "end": 2404.02,
    "text": "and Than in a Moment We'll Be Talking About the example that you have, this onion for engineers."
  },
  {
    "start": 2404.1,
    "end": 2409.1,
    "text": "But let's make sure somewhere today we're going to be talking MCP."
  },
  {
    "start": 2409.54,
    "end": 2413.36,
    "text": "I don't think it is at its right point yet but lets not forget that."
  },
  {
    "start": 2413.96,
    "end": 2415.58,
    "text": "Let us say a little bit with data."
  },
  {
    "start": 2415.98,
    "end": 2420.22,
    "text": "You already said before or mentioned a couple of times Data silos here."
  },
  {
    "start": 2420.28,
    "end": 2422.66,
    "text": "We've always heard about them and all the time."
  },
  {
    "start": 2422.76,
    "end": 2434.66,
    "text": "So how then in this time of agentic AI In case for engineering How do NOT manage data and how to best manage the data."
  },
  {
    "start": 2435.38,
    "end": 2452.12,
    "text": "Yeah, I think... And we are super happy to include MCP here as an example where agents can make this a lot easier because with classical APIs you have like if you change something You'll have rewrite a lot of integrations With MCP."
  },
  {
    "start": 2452.42,
    "end": 2458.28,
    "text": "If some things changed The agent or LLM will pick it up And so the integration effort is much less."
  },
  {
    "start": 2458.46,
    "end": 2469.14,
    "text": "So in this case, agents can really help with integration but there are some things agents cannot help and that's mainly related to like how fast is the integration?"
  },
  {
    "start": 2469.3,
    "end": 2469.82,
    "text": "especially?"
  },
  {
    "start": 2469.92,
    "end": 2471.4,
    "text": "obviously I mean we have customers."
  },
  {
    "start": 2471.48,
    "end": 2474.66,
    "text": "they have hundreds of terabytes or fleet data."
  },
  {
    "start": 2475.74,
    "end": 2480.26,
    "text": "it's important to access these data first In order for them to be useful."
  },
  {
    "start": 2480.96,
    "end": 2483.6,
    "text": "Also i think security a super-important point."
  },
  {
    "start": 2483.68,
    "end": 2489.92,
    "text": "We all know security has become as important as never before, so to sandbox it correctly."
  },
  {
    "start": 2490.42,
    "end": 2497.74,
    "text": "And in this case I think its super-important that your data is in a place where can be carried easily"
  },
  {
    "start": 2498.32,
    "end": 2498.76,
    "text": "fast"
  },
  {
    "start": 2499.2,
    "end": 2500.72,
    "text": "and is governed correctly."
  },
  {
    "start": 2502.12,
    "end": 2524.86,
    "text": "There are probably several ways like to achieve that but we really see a big shift in the market where people gravitating towards data lake house because it's like very open layer and also to be honest, because Databricks at Microsoft have good sales teams for their fabric products."
  },
  {
    "start": 2525.26,
    "end": 2532.76,
    "text": "But I think that makes sense as an underlying data infrastructure for a lot of data."
  },
  {
    "start": 2533.0,
    "end": 2537.62,
    "text": "And I think most engineering data, like especially sensor data is there."
  },
  {
    "start": 2537.8,
    "end": 2538.96,
    "text": "it's the right choice."
  },
  {
    "start": 2539.02,
    "end": 2549.02,
    "text": "so we see with a lot customers they're migrating away both from file-based systems but also form isolated time series databases."
  },
  {
    "start": 2550.58,
    "end": 2560.56,
    "text": "If you use the word lake house I've been used to the general term of data lake, but Lake House for me is like in you."
  },
  {
    "start": 2560.72,
    "end": 2564.44,
    "text": "Is that a typical way of organizing data?"
  },
  {
    "start": 2565.5,
    "end": 2572.72,
    "text": "In a lake or an house of lakes... What's your understanding with us at Lake House?"
  },
  {
    "start": 2573.1,
    "end": 2573.72,
    "text": "Yeah, absolutely."
  },
  {
    "start": 2574.0,
    "end": 2582.76,
    "text": "We'd really love to but I think because it's a super important term... It originated probably in marketing terms maybe even by Databricks i'm not sure."
  },
  {
    "start": 2583.92,
    "end": 2585.1,
    "text": "But what is meant by that?"
  },
  {
    "start": 2585.68,
    "end": 2596.78,
    "text": "In the classical time series or the classical database What you get are good access fast access and with SQL for example You can run complex queries."
  },
  {
    "start": 2597.32,
    "end": 2604.06,
    "text": "The problem is in typical database storage is very expensive And you have to go data schemas."
  },
  {
    "start": 2604.14,
    "end": 2606.58,
    "text": "You have to clean your data, fit into the data schema right?"
  },
  {
    "start": 2606.62,
    "end": 2608.38,
    "text": "That's one story."
  },
  {
    "start": 2608.44,
    "end": 2613.74,
    "text": "The other stories is that in Data Lake basically it just has a blob storage which is very cheap."
  },
  {
    "start": 2613.84,
    "end": 2620.06,
    "text": "It about like ten times cheaper or only ninety percent cheaper whatever as database storage."
  },
  {
    "start": 2620.94,
    "end": 2625.22,
    "text": "But if we put our files there We can't use them because they become swamp."
  },
  {
    "start": 2626.24,
    "end": 2628.26,
    "text": "So that was not a successful concept."
  },
  {
    "start": 2628.7,
    "end": 2642.88,
    "text": "And the Data Lakehouse is you store your data on cheap blob storage, but you put SQL based... Typically SQL can be other... But typically SQL-based engine on top that can query the data and their engines are built in way."
  },
  {
    "start": 2643.2,
    "end": 2648.28,
    "text": "they have intelligent caching stuff for example like ClickHouse."
  },
  {
    "start": 2648.8,
    "end": 2649.22,
    "text": "it's super"
  },
  {
    "start": 2649.28,
    "end": 2649.6,
    "text": "fast."
  },
  {
    "start": 2650.06,
    "end": 2657.52,
    "text": "And the last point I would like to mention is that in data lake house, it's really dominated by open source architectures."
  },
  {
    "start": 2657.66,
    "end": 2662.98,
    "text": "It's Delta Lake which Databricks uses and Microsoft Fabric use but then there also a Pagae iceberg."
  },
  {
    "start": 2663.24,
    "end": 2665.24,
    "text": "so i think its'a good choice."
  },
  {
    "start": 2665.68,
    "end": 2669.16,
    "text": "at least gives you little bit of vendor independence."
  },
  {
    "start": 2674.92,
    "end": 2675.3,
    "text": "Very good."
  },
  {
    "start": 2675.48,
    "end": 2677.34,
    "text": "What about vector or database?"
  },
  {
    "start": 2677.44,
    "end": 2693.06,
    "text": "There was a time that we said king minus man plus woman equals queen and We would find words in the multi-dimensional space, but you know I try to realize In my brain how to get there."
  },
  {
    "start": 2693.58,
    "end": 2698.74,
    "text": "have they Have been forgotten did not come through at all?"
  },
  {
    "start": 2698.84,
    "end": 2705.0,
    "text": "Was it part of high cycle of disillusionment?"
  },
  {
    "start": 2706.64,
    "end": 2708.5,
    "text": "Yeah, I think so actually."
  },
  {
    "start": 2708.64,
    "end": 2710.68,
    "text": "There's several factors at play here."
  },
  {
    "start": 2710.78,
    "end": 2714.44,
    "text": "So first things first we still use vector databases."
  },
  {
    "start": 2714.52,
    "end": 2723.66,
    "text": "It is a super important component for any agentic AI system to find data For example the correct channel name out of like ten thousand signals."
  },
  {
    "start": 2724.5,
    "end": 2737.52,
    "text": "But We learned that this what we all did in twenty-four maybe even Is like these retrieval augmented generation REC systems Rag alone and just vector search, a loan is typically not sufficient."
  },
  {
    "start": 2737.6,
    "end": 2746.34,
    "text": "So you have to combine it with more agentic architectures that like on demand use vector search but also other types of search defined information."
  },
  {
    "start": 2746.8,
    "end": 2749.34,
    "text": "I think That's one reason why It's Not Hypey anymore."
  },
  {
    "start": 2749.38,
    "end": 2753.94,
    "text": "And the Other Reason Is That For Most Applications Its Actually Fine."
  },
  {
    "start": 2754.06,
    "end": 2756.98,
    "text": "Just Use A Reasonably Good Database."
  },
  {
    "start": 2757.02,
    "end": 2760.0,
    "text": "for example We Used Postgres so we used the postres extension."
  },
  {
    "start": 2762.56,
    "end": 2762.8,
    "text": "fine."
  },
  {
    "start": 2763.16,
    "end": 2768.5,
    "text": "We don't need some specialized high performance thing and I think that's true for most use cases."
  },
  {
    "start": 2768.94,
    "end": 2771.96,
    "text": "good one question before we then move on."
  },
  {
    "start": 2772.04,
    "end": 2775.82,
    "text": "so you mentioned the mcp example, were not going to go into MCP?"
  },
  {
    "start": 2776.68,
    "end": 2784.04,
    "text": "For those of your listeners who do want know more there are a couple places where they have been talking about MCP in their last half year."
  },
  {
    "start": 2784.16,
    "end": 2788.68,
    "text": "i guess now MCP was kind of going the other way around."
  },
  {
    "start": 2789.22,
    "end": 2791.28,
    "text": "So it came out, I don't know a year or something."
  },
  {
    "start": 2791.42,
    "end": 2797.1,
    "text": "everybody was completely impressed and then very quickly people started talking."
  },
  {
    "start": 2797.16,
    "end": 2803.64,
    "text": "you already mentioned security but is MCP receiving updates?"
  },
  {
    "start": 2803.74,
    "end": 2804.52,
    "text": "How does that work?"
  },
  {
    "start": 2804.64,
    "end": 2808.74,
    "text": "Is the community are happy with the capabilities?"
  },
  {
    "start": 2813.72,
    "end": 2818.36,
    "text": "Yeah I think it's super interesting to see how this develops and i think there are a lot of lessons to be learned."
  },
  {
    "start": 2819.26,
    "end": 2821.8,
    "text": "And, as honest as possible."
  },
  {
    "start": 2821.88,
    "end": 2827.48,
    "text": "so right now we actually don't have use cases that really use in production."
  },
  {
    "start": 2827.86,
    "end": 2838.44,
    "text": "the reason is for most generating code is a lot more efficient than tool calling."
  },
  {
    "start": 2838.7,
    "end": 2847.66,
    "text": "So for example, we use SQL and you regenerate SQL code And then I can attach my database very easily through this SQL interface."
  },
  {
    "start": 2848.18,
    "end": 2855.54,
    "text": "so on This side i would say like it's still there are many applications where mcp has to prove it's worth."
  },
  {
    "start": 2856.12,
    "end": 2861.46,
    "text": "But on the other hand, I'm absolutely one hundred percent sure at the moment that MCP will be a key factor."
  },
  {
    "start": 2861.5,
    "end": 2862.58,
    "text": "and i would give you an example."
  },
  {
    "start": 2862.64,
    "end": 2873.78,
    "text": "just the other day was in a trade fair and I saw super cool examples by Vector or vector is one of the market leaders in Ken bus technology and automotive."
  },
  {
    "start": 2873.86,
    "end": 2874.6,
    "text": "so for everyone."
  },
  {
    "start": 2874.88,
    "end": 2877.92,
    "text": "everybody in automotive knows Vector and don't sell Vector"
  },
  {
    "start": 2877.98,
    "end": 2878.62,
    "text": "databases"
  },
  {
    "start": 2879.0,
    "end": 2879.1,
    "text": "now."
  },
  {
    "start": 2879.3,
    "end": 2885.48,
    "text": "yeah no Just very In the end of the eighties or so."
  },
  {
    "start": 2885.58,
    "end": 2887.86,
    "text": "So very forward-looking name and a very cool company."
  },
  {
    "start": 2887.96,
    "end": 2900.22,
    "text": "And what they are doing is like building MCP interfaces for all their products, it's just such an enabler For example our workflows when we have some data analysis."
  },
  {
    "start": 2900.6,
    "end": 2901.66,
    "text": "You want to jump deeper."
  },
  {
    "start": 2901.82,
    "end": 2908.02,
    "text": "then you want your vector tool that has all the bells and whistles That I've been using in my last twenty years."
  },
  {
    "start": 2908.42,
    "end": 2910.28,
    "text": "Now with MCP you can easily do it."
  },
  {
    "start": 2911.38,
    "end": 2912.5,
    "text": "Just seeing how."
  },
  {
    "start": 2913.3,
    "end": 2930.9,
    "text": "I would say that these legacy engineering companies, so to speak like the software companies they are already introducing it and i think push that MCP, I think will indeed be a very important integration layer."
  },
  {
    "start": 2930.96,
    "end": 2943.04,
    "text": "That makes sense not for everything but for a lot of use cases and you would see it in little bit of delay because obviously to be useful for our system we need the integrations by vector right?"
  },
  {
    "start": 2943.1,
    "end": 2944.34,
    "text": "So its an ecosystem play."
  },
  {
    "start": 2945.38,
    "end": 2946.46,
    "text": "this will develop slowly."
  },
  {
    "start": 2948.02,
    "end": 2948.4,
    "text": "Very good."
  },
  {
    "start": 2948.46,
    "end": 2955.44,
    "text": "Let's move to Onion, RenewMix-Onion agentic AI assistant for engineering data analysis."
  },
  {
    "start": 2955.52,
    "end": 2956.2,
    "text": "is that correct?"
  },
  {
    "start": 2956.42,
    "end": 2958.52,
    "text": "Please give us a first introduction."
  },
  {
    "start": 2959.42,
    "end": 2975.58,
    "text": "Yeah if I only have a few sentences i'd say like onion is the best AI harness for uh engineering data specifically for analyzing sensor data like test data fleet data simulation production data."
  },
  {
    "start": 2975.96,
    "end": 2992.36,
    "text": "And what we really strive for is ONION to be very intuitive, so it's not like a specialized tool with lots of belts and whistles – that you don't need a handbook to learn!"
  },
  {
    "start": 2994.8,
    "end": 2996.74,
    "text": "People often say how fast was this?"
  },
  {
    "start": 2997.8,
    "end": 3002.26,
    "text": "You can pull the data very quickly see even large amounts."
  },
  {
    "start": 3004.6,
    "end": 3011.44,
    "text": "very general purpose systems like for example Power BI or Tableau, whatever just don't have the performance."
  },
  {
    "start": 3012.4,
    "end": 3015.08,
    "text": "And yeah that's like onion and nacha"
  },
  {
    "start": 3015.9,
    "end": 3016.26,
    "text": "Very good."
  },
  {
    "start": 3016.3,
    "end": 3017.98,
    "text": "Yeah, we're going to get into the details anyway."
  },
  {
    "start": 3018.54,
    "end": 3021.92,
    "text": "now harness is a relatively new term."
  },
  {
    "start": 3022.02,
    "end": 3023.82,
    "text": "there's one yet another term."
  },
  {
    "start": 3023.86,
    "end": 3029.98,
    "text": "I mean i just talked about it Robert and I talked about her Just two weeks ago because i read About somewhere."
  },
  {
    "start": 3030.02,
    "end": 3031.06,
    "text": "There's this New job?"
  },
  {
    "start": 3031.24,
    "end": 3039.76,
    "text": "I think It's The artificial AI Harness Engineer Whatever And I've got the little story To say here as well but i'm gonna do that just in A moment."
  },
  {
    "start": 3040.28,
    "end": 3047.44,
    "text": "But an ask you to share with us what is the harness element of onion?"
  },
  {
    "start": 3047.54,
    "end": 3049.74,
    "text": "What does that do, what doesn't it not"
  },
  {
    "start": 3049.82,
    "end": 3049.88,
    "text": "do."
  },
  {
    "start": 3050.62,
    "end": 3052.46,
    "text": "Yeah so yeah That's very interesting."
  },
  {
    "start": 3052.54,
    "end": 3059.34,
    "text": "It's a new word and also hopefully now we can sometimes define little bit better about who actually do or what onion is good at."
  },
  {
    "start": 3059.76,
    "end": 3071.08,
    "text": "So maybe going back through this three layer cake right now Onion as a full stack system has connections to data sources in particular can also like upload files and have a good integrated data management."
  },
  {
    "start": 3071.22,
    "end": 3082.9,
    "text": "You have this agentic layer, which is the harness I'll talk about little bit more in detail about that where we have agents that are very good at dealing with time series data problems."
  },
  {
    "start": 3083.24,
    "end": 3095.24,
    "text": "And then on top you will have the UI layer because there was no UI layer before that handled it with sufficient speed and usability engineers need."
  },
  {
    "start": 3095.42,
    "end": 3099.0,
    "text": "so as all these three components,"
  },
  {
    "start": 3099.24,
    "end": 3099.5,
    "text": "but..."
  },
  {
    "start": 3100.56,
    "end": 3103.18,
    "text": "they are modular, and I expect in the future"
  },
  {
    "start": 3103.38,
    "end": 3103.68,
    "text": "that"
  },
  {
    "start": 3104.12,
    "end": 3107.76,
    "text": "if we think of how the agentic ecosystems will play out."
  },
  {
    "start": 3108.6,
    "end": 3115.1,
    "text": "That every company will have modular agenetic ecosystems And one... The middle layer-the agentic layer."
  },
  {
    "start": 3115.22,
    "end": 3120.64,
    "text": "In this you'll probably have different agents specialized for certain tasks."
  },
  {
    "start": 3121.64,
    "end": 3124.06,
    "text": "Why do need to specialize agents?"
  },
  {
    "start": 3124.18,
    "end": 3130.14,
    "text": "It's because LLM is only a factor the biggest factor."
  },
  {
    "start": 3130.3,
    "end": 3135.26,
    "text": "but you need to control the context obviously around the LLM, right?"
  },
  {
    "start": 3135.32,
    "end": 3137.1,
    "text": "For example just take an example."
  },
  {
    "start": 3137.44,
    "end": 3149.48,
    "text": "You have some test data and want in the charging cycle of a car, then you need to give the LLM information about what signals are and data."
  },
  {
    "start": 3150.18,
    "end": 3152.26,
    "text": "What may be even typical anomalies?"
  },
  {
    "start": 3152.6,
    "end": 3157.42,
    "text": "And maybe you can also introduce some information from previous analysis."
  },
  {
    "start": 3157.54,
    "end": 3170.58,
    "text": "so there's lots descriptions that have been put into it but end up being given by other big ingredient specialized tools through this system for example search for channels very quickly, right?"
  },
  {
    "start": 3170.92,
    "end": 3175.92,
    "text": "And providing this context also with input from user but sometimes automatically."
  },
  {
    "start": 3176.7,
    "end": 3181.86,
    "text": "Providing tools and increasingly providing sandboxes and governance."
  },
  {
    "start": 3182.62,
    "end": 3183.8,
    "text": "This is the task of"
  },
  {
    "start": 3183.88,
    "end": 3184.32,
    "text": "the harness."
  },
  {
    "start": 3184.5,
    "end": 3188.3,
    "text": "I do recall there was this piece of research a couple months back."
  },
  {
    "start": 3188.44,
    "end": 3194.5,
    "text": "they looked at whatever code i think it Was called by cursor or The other one Whatever completely."
  },
  {
    "start": 3194.56,
    "end": 3196.34,
    "text": "Yeah, you know what I'm"
  },
  {
    "start": 3196.44,
    "end": 3198.8,
    "text": "referring to?"
  },
  {
    "start": 3198.84,
    "end": 3201.98,
    "text": "At least they said ninety ninety eight or two percent."
  },
  {
    "start": 3202.14,
    "end": 3207.08,
    "text": "only if you look at i guess the number of code lines are words have no idea."
  },
  {
    "start": 3207.16,
    "end": 3214.56,
    "text": "Only two percent was like related to AI maybe large language models and ninety eight percent is that it's Like The Rest."
  },
  {
    "start": 3214.76,
    "end": 3217.64,
    "text": "Maybe the rest Is For The Control Right?"
  },
  {
    "start": 3218.0,
    "end": 3220.04,
    "text": "Is The Rest What Typically?"
  },
  {
    "start": 3220.4,
    "end": 3222.4,
    "text": "What You Could Then Call The Harness Or"
  },
  {
    "start": 3223.02,
    "end": 3223.28,
    "text": "No, no."
  },
  {
    "start": 3223.5,
    "end": 3228.52,
    "text": "I think the Claude quote is the harness and I'm not sure like what exactly these numbers are referring to."
  },
  {
    "start": 3228.72,
    "end": 3249.94,
    "text": "so obviously in The Harness you don't have code for the LLM You just have code around it, prompts how to call two worlds Like if you have sub-agents stuff like this And there's a lot of work currently that has to do with the harness To make model behave good and optimize things."
  },
  {
    "start": 3250.02,
    "end": 3253.04,
    "text": "So If we look at code for Claude Code."
  },
  {
    "start": 3253.1,
    "end": 3270.92,
    "text": "It's like basically the harness, but I think there will also be like code entropic hairs around on the LLMs and i think that one point actually is where you have to say it might be dangerous for a company... For example, we are building the best harness for engineers."
  },
  {
    "start": 3272.08,
    "end": 3280.56,
    "text": "What can happen on what we see happened is that Entropic, for example as an LLM provider and occurs us also with their own models."
  },
  {
    "start": 3280.62,
    "end": 3286.56,
    "text": "They train the models specifically for their harness And That's actually a very good strategy."
  },
  {
    "start": 3286.7,
    "end": 3293.8,
    "text": "So the model has already pre-trained to fit exactly in this context With it exactly these tools and things."
  },
  {
    "start": 3294.44,
    "end": 3298.8,
    "text": "I think It's gonna be Very interesting To See how This play out."
  },
  {
    "start": 3298.88,
    "end": 3305.04,
    "text": "How i would say like Will we all still have like special, more specialized harnesses?"
  },
  {
    "start": 3305.14,
    "end": 3307.46,
    "text": "Like stuff we built in the next ten years?"
  },
  {
    "start": 3307.88,
    "end": 3308.18,
    "text": "or"
  },
  {
    "start": 3309.76,
    "end": 3316.76,
    "text": "will everything be just the entropic layer of a predefined harness that works well."
  },
  {
    "start": 3316.96,
    "end": 3326.34,
    "text": "Because if you look at how good for example Cloud Code works it's... For many applications its always amazing."
  },
  {
    "start": 3326.94,
    "end": 3330.96,
    "text": "But then the joke I would make is like as long you're still allowed to use it, right?"
  },
  {
    "start": 3332.96,
    "end": 3333.2,
    "text": "Yeah, yeah."
  },
  {
    "start": 3333.28,
    "end": 3333.64,
    "text": "That's right."
  },
  {
    "start": 3333.78,
    "end": 3340.36,
    "text": "No no absolutely I think sovereignty is an important point but also like governance isn't a potent point."
  },
  {
    "start": 3340.52,
    "end": 3349.76,
    "text": "i think these are things and And I also believe especially in our case Right now usability as Also A super Super Important Point."
  },
  {
    "start": 3350.8,
    "end": 3365.12,
    "text": "As An Engineer for Example Cloud Code or cloud co-work might be if you would Like to use it It can answer lot of stuff just goes on the loop and comes back with reasonable Answer In ten minutes or so, but as an engineer you first of all care about that."
  },
  {
    "start": 3365.18,
    "end": 3367.0,
    "text": "it's like very robust and repeatable."
  },
  {
    "start": 3367.24,
    "end": 3370.6,
    "text": "And I think their specialization can help tremendously."
  },
  {
    "start": 3371.32,
    "end": 3375.12,
    "text": ".And another thing ,and i think its really important is also efficiency."
  },
  {
    "start": 3376.32,
    "end": 3392.62,
    "text": "actually just yesterday in the US is I think, hope i'm not misquoting this because it's sort of like two thousand dollars per developer which is more than the median salary and India or something at that."
  },
  {
    "start": 3392.76,
    "end": 3404.96,
    "text": "So we are close to the point where you spend more on tokens in America then developers earn an india And I think all customers see this software development already that efficiency is key."
  },
  {
    "start": 3405.16,
    "end": 3415.58,
    "text": "so We believe as long as we have most efficient harness typically at least fifty percent less, sometimes far more or less token than general purpose harnesses."
  },
  {
    "start": 3415.8,
    "end": 3423.48,
    "text": "Even if the General Purpose Harness can get a result that this is like an important factor in deciding for"
  },
  {
    "start": 3424.0,
    "end": 3424.56,
    "text": "more specialized"
  },
  {
    "start": 3424.62,
    "end": 3424.88,
    "text": "harness?"
  },
  {
    "start": 3425.44,
    "end": 3435.98,
    "text": "Yeah I think my answer to you know are we gonna depend on one...I think the answer as always On a single solution, right?"
  },
  {
    "start": 3436.02,
    "end": 3440.18,
    "text": "So we always should be supporting you know with several solutions."
  },
  {
    "start": 3440.22,
    "end": 3446.02,
    "text": "You know the majority can go for the best one but will always need to second option I believe now."
  },
  {
    "start": 3446.12,
    "end": 3449.2,
    "text": "I want to wanna share With you little stories."
  },
  {
    "start": 3449.38,
    "end": 3456.26,
    "text": "it's from Autonomous driving so a little bit related But not directly adorably that you mentioned Ken bus For example."
  },
  {
    "start": 3456.8,
    "end": 3470.54,
    "text": "i already told The first piece when I talked To Robert because It's about the Munich Center Brewer day just a couple of days back and that's where I saw these horses, you know These big horses who carry this?"
  },
  {
    "start": 3471.32,
    "end": 3473.86,
    "text": "Cards and the beer and stuff like that."
  },
  {
    "start": 3473.9,
    "end": 3490.12,
    "text": "I'm not gonna redo that but when I was preparing with you And then it was about to the harness and I came to the horse again Because that's what typically and then I learned what the horse typically have even if consumers ride a horse."
  },
  {
    "start": 3490.5,
    "end": 3511.46,
    "text": "I believe that's our harness as well, so for humans i believe to steer the horse right and then to the story of the origin of The Autonomous Driving with my, and I told my fellow marching band colleagues who normally you know they are not really interested in engineering autonomous driving but i've told them."
  },
  {
    "start": 3511.68,
    "end": 3529.7,
    "text": "I have to tell you so share it here...I have shared that in a past because this is a story that Robert and I learned about seven years back where who meanwhile has been recognized and reported by the German government for inventing autonomous driving, in the eighties, nineties."
  },
  {
    "start": 3530.02,
    "end": 3541.72,
    "text": "And as a young man he wanted to build a machine because that's when the tractor came on site... ...and he had been used to sitting you know on the bench with his dad behind the horse!"
  },
  {
    "start": 3541.94,
    "end": 3542.5,
    "text": "When is that?"
  },
  {
    "start": 3542.6,
    "end": 3543.48,
    "text": "what falls asleep?"
  },
  {
    "start": 3543.64,
    "end": 3546.52,
    "text": "The horse would drive back home itself!"
  },
  {
    "start": 3549.04,
    "end": 3555.78,
    "text": "And then he decided, now we've got this strong mechanical horse attractor but it cannot find its way home."
  },
  {
    "start": 3556.42,
    "end": 3560.4,
    "text": "That was the reason... He said that's what I'm going to do as an engineer."
  },
  {
    "start": 3560.56,
    "end": 3562.24,
    "text": "isn't there a wonderful story?"
  },
  {
    "start": 3562.7,
    "end": 3567.8,
    "text": "Then he did actually really come up with the basics of autonomous driving."
  },
  {
    "start": 3568.58,
    "end": 3577.96,
    "text": "I want to play this a little one step further then, because if you're going have system running autonomously is that ever gonna happen?"
  },
  {
    "start": 3579.48,
    "end": 3584.74,
    "text": "The question is more like how loosely can your deal with the hardness?"
  },
  {
    "start": 3585.06,
    "end": 3588.06,
    "text": "maybe it relates to things we were just saying."
  },
  {
    "start": 3588.18,
    "end": 3591.64,
    "text": "If you've won or two are we gonna be dependent on one?"
  },
  {
    "start": 3592.16,
    "end": 3597.06,
    "text": "and i'm not sure...I am clear in my brain where its moving."
  },
  {
    "start": 3597.84,
    "end": 3598.58,
    "text": "Yeah, I think it's super."
  },
  {
    "start": 3599.42,
    "end": 3606.16,
    "text": "In the beginning we established how long i was in the past so difficult to make predictions about the future."
  },
  {
    "start": 3607.44,
    "end": 3609.2,
    "text": "You mentioned that It is actually quite interesting."
  },
  {
    "start": 3609.759,
    "end": 3628.38,
    "text": "when you started off with this topic and a genetic analysis for engineers We had big use case of fleet data where we build system for an OEM That can answer questions like show me typical charging cycles after which are many kilometers our cars typically charged."
  },
  {
    "start": 3628.82,
    "end": 3636.7,
    "text": "And previously they had to do it with Power BI and was like a long process data analyst, have to do a dashboard to answer the question took days?"
  },
  {
    "start": 3637.86,
    "end": 3645.799,
    "text": "Now we had this system where you could just ask but the system was constrained in a way that is more like hey show me this chart should be this histogram right."
  },
  {
    "start": 3646.06,
    "end": 3657.32,
    "text": "It's not really free and agentic But we did because humans has lot of control over it Because In This Case The querying the data take like minutes or even longer."
  },
  {
    "start": 3657.759,
    "end": 3660.759,
    "text": "So if the agenda makes a mistake, it's very expensive."
  },
  {
    "start": 3661.279,
    "end": 3663.74,
    "text": "so we guard where did super hard."
  },
  {
    "start": 3664.24,
    "end": 3672.9,
    "text": "and then about half year ago I think that was right there were many things."
  },
  {
    "start": 3673.58,
    "end": 3683.64,
    "text": "people for me with Opus of point six normally don't call out specific models but this is really suddenly got good."
  },
  {
    "start": 3684.06,
    "end": 3690.779,
    "text": "and it got like so good that It could write SQL queries very easily, Could go on its own."
  },
  {
    "start": 3691.46,
    "end": 3697.279,
    "text": "Could do longer workflows And I was completely blindsided with what is going on."
  },
  {
    "start": 3697.38,
    "end": 3700.68,
    "text": "Just...I remember this because just six weeks earlier Was the customer."
  },
  {
    "start": 3700.779,
    "end": 3703.36,
    "text": "they were asking me How about text to SQL?"
  },
  {
    "start": 3703.42,
    "end": 3705.06,
    "text": "And i was telling them I know!"
  },
  {
    "start": 3705.14,
    "end": 3705.74,
    "text": "It doesn't work!"
  },
  {
    "start": 3707.4,
    "end": 3712.34,
    "text": "And then just six week later one of the stuff works so well now, unbelievable."
  },
  {
    "start": 3712.779,
    "end": 3716.16,
    "text": "And um... So yeah that's just to say I think it is difficult make predictions."
  },
  {
    "start": 3716.34,
    "end": 3720.54,
    "text": "i think there are some studies where they said hey agents can do."
  },
  {
    "start": 3721.4,
    "end": 3725.8,
    "text": "Like how long they can work uninterrupted on a difficult task?"
  },
  {
    "start": 3725.88,
    "end": 3727.68,
    "text": "It like doubles, I'm not sure if you know the figures."
  },
  {
    "start": 3727.72,
    "end": 3731.319,
    "text": "it like doubles every six months or so every year."
  },
  {
    "start": 3731.42,
    "end": 3737.84,
    "text": "So its kind of Moore's Law for Agents And i think there is some truth in that."
  },
  {
    "start": 3739.46,
    "end": 3748.9,
    "text": "right now The story at least in Onion You Can Do A Lot Of Stuff Interactively and what you typically do also maybe an interesting technical aspect."
  },
  {
    "start": 3749.0,
    "end": 3750.42,
    "text": "So we actually have two phases."
  },
  {
    "start": 3750.5,
    "end": 3764.56,
    "text": "So typically the agent goes on its own, writes code does some analysis and normally detection for example And then once the agent has a result it generates charts to show the results of a human, but the charts are not code-gen."
  },
  {
    "start": 3764.64,
    "end": 3767.86,
    "text": "They're toolcodes so they just like some filters it configures."
  },
  {
    "start": 3768.24,
    "end": 3770.9,
    "text": "So we always sure that the chart is correct."
  },
  {
    "start": 3771.1,
    "end": 3785.3,
    "text": "You have this phase where agent does stuff on its own and then you have these phases Where we consolidate findings in way That's super transparent that humans can actually understand."
  },
  {
    "start": 3785.66,
    "end": 3787.5,
    "text": "Have good look at them."
  },
  {
    "start": 3787.72,
    "end": 3789.819,
    "text": "I think thats best strategy right now."
  },
  {
    "start": 3790.34,
    "end": 3792.4,
    "text": "Like, where will it go in the future?"
  },
  {
    "start": 3792.6,
    "end": 3794.24,
    "text": "Will we still have human interaction."
  },
  {
    "start": 3794.4,
    "end": 3795.04,
    "text": "Where would we have?"
  },
  {
    "start": 3795.12,
    "end": 3797.46,
    "text": "it probably very difficult to say."
  },
  {
    "start": 3799.819,
    "end": 3806.259,
    "text": "It's always difficult looking into the future but if Peter does So, yeah."
  },
  {
    "start": 3806.88,
    "end": 3809.58,
    "text": "You already kind of moved in and I was going to ask you there."
  },
  {
    "start": 3809.72,
    "end": 3812.68,
    "text": "so who is the typical onion user?"
  },
  {
    "start": 3812.86,
    "end": 3823.24,
    "text": "Maybe want to combine it with a use case And Is that person The same engineer That he or she was A year ago?"
  },
  {
    "start": 3823.779,
    "end": 3830.22,
    "text": "Or then thinking about new young engineers coming into market Are they doing things a different way?"
  },
  {
    "start": 3830.3,
    "end": 3833.22,
    "text": "And maybe if you take the use case, may be want to look at that."
  },
  {
    "start": 3833.319,
    "end": 3841.24,
    "text": "How would have done it... The engineer has done this year ago and how will he or she do by using onion?"
  },
  {
    "start": 3841.92,
    "end": 3845.12,
    "text": "Yeah I think maybe i'll start with some example use cases."
  },
  {
    "start": 3845.319,
    "end": 3847.54,
    "text": "hopefully people can relate too."
  },
  {
    "start": 3848.2,
    "end": 3851.9,
    "text": "Interesting thing is there come from all fields of engineering."
  },
  {
    "start": 3852.259,
    "end": 3853.819,
    "text": "we already had fleet data analysis."
  },
  {
    "start": 3854.279,
    "end": 3859.259,
    "text": "where Do users use my product?"
  },
  {
    "start": 3859.36,
    "end": 3864.08,
    "text": "Like when do they charge or do I have more errors with the new software version stuff like that."
  },
  {
    "start": 3864.48,
    "end": 3868.96,
    "text": "Then we have classical testing, for example you're using testing stuff is on a test bench."
  },
  {
    "start": 3869.04,
    "end": 3875.759,
    "text": "it's tortures for month, and I want to maybe investigate if there are some error logs at some points."
  },
  {
    "start": 3875.88,
    "end": 3880.46,
    "text": "And go through some events where stuff happens to understand what iIwantto improve?"
  },
  {
    "start": 3881.46,
    "end": 3890.68,
    "text": "Maybe you want to correlate that with simulation data or tune my simulations so it better fits the test... The last big area is production!"
  },
  {
    "start": 3890.779,
    "end": 3898.259,
    "text": "For example, I wanted an energy report like a bill saying why do they build so much higher in March than"
  },
  {
    "start": 3898.46,
    "end": 3899.04,
    "text": "February?!"
  },
  {
    "start": 3899.48,
    "end": 3903.56,
    "text": "And you have to go through the data, and it takes days maybe."
  },
  {
    "start": 3905.06,
    "end": 3909.48,
    "text": "With ONION or with the Gentik AI You can do all these tasks actually in a few minutes."
  },
  {
    "start": 3910.5,
    "end": 3914.46,
    "text": "So this are typical use cases for ONION In the past."
  },
  {
    "start": 3914.56,
    "end": 3918.279,
    "text": "It's interesting that we had like two worlds in engineering."
  },
  {
    "start": 3918.4,
    "end": 3919.6,
    "text": "One was the big data world."
  },
  {
    "start": 3919.74,
    "end": 3921.1,
    "text": "We had data analysis teams."
  },
  {
    "start": 3921.38,
    "end": 3933.8,
    "text": "They were working on databases Like an influx DB Or increasingly with lake house Databricks or Tableau and would make, or Kafanas is used a lot in production environments for example."
  },
  {
    "start": 3934.44,
    "end": 3936.22,
    "text": "And they were like building these dashboards!"
  },
  {
    "start": 3936.36,
    "end": 3943.74,
    "text": "Then you had engineers who have more engineering tools... For example there's Diademis one very specialized tool."
  },
  {
    "start": 3943.819,
    "end": 3945.2,
    "text": "I was mentioning the vector tooling."
  },
  {
    "start": 3945.96,
    "end": 3959.86,
    "text": "maybe some handmade tools are often testing actually some very custom tools by the manufacturer of measurement equipment And it was really difficult to answer questions because engineers sometimes for the big data stuff, they have go through that."
  },
  {
    "start": 3960.1,
    "end": 3979.02,
    "text": "Data teams and if I really want to investigate deeper actually had to pull raw file data loaded into their desktop based engineering applications in this process typically like took days easily often weeks and this is like independent from my eye."
  },
  {
    "start": 3979.24,
    "end": 3984.06,
    "text": "the big thing we actually saw many times in our projects with testing data that."
  },
  {
    "start": 3984.46,
    "end": 4000.22,
    "text": "We have to kind of bridge this world, you have to make sure engineers can super quickly access like broad data comparing a lot of runs ,a lot of testing runs or trips which actually engineers in many companies don't really do they want."
  },
  {
    "start": 4000.5,
    "end": 4003.819,
    "text": "but in their tooling, they can maybe compare two runs."
  },
  {
    "start": 4004.2,
    "end": 4007.88,
    "text": "But not seeing the patterns and hundreds or thousands of runs."
  },
  {
    "start": 4007.94,
    "end": 4008.74,
    "text": "so you want to do that?"
  },
  {
    "start": 4009.34,
    "end": 4014.96,
    "text": "You also wanna give engineers an easy-to-use tool where you could actually see signals for example."
  },
  {
    "start": 4015.0,
    "end": 4017.18,
    "text": "it's super easy but customers always love it."
  },
  {
    "start": 4017.64,
    "end": 4024.02,
    "text": "we have a very good functionality too zoom into like line charts really fast and its very accurate."
  },
  {
    "start": 4025.44,
    "end": 4029.46,
    "text": "They just love because there are details on what is going."
  },
  {
    "start": 4030.04,
    "end": 4034.779,
    "text": "And bridging this gap, I think was an important part of the onion mission."
  },
  {
    "start": 4035.04,
    "end": 4041.08,
    "text": "It's also where a lot value currently is for people that use"
  },
  {
    "start": 4041.58,
    "end": 4041.6,
    "text": "it."
  },
  {
    "start": 4041.98,
    "end": 4055.56,
    "text": "Let's stay with the use cases and maybe one that I've been actively involved in an hour talking ten years ago when everybody, since then a production world has been doing anomaly detection."
  },
  {
    "start": 4055.7,
    "end": 4056.94,
    "text": "now how is?"
  },
  {
    "start": 4057.279,
    "end": 4060.759,
    "text": "it may be onion base but in general agentic."
  },
  {
    "start": 4061.319,
    "end": 4063.48,
    "text": "Anomaly detection different from."
  },
  {
    "start": 4063.8,
    "end": 4069.5,
    "text": "you know what we were doing Dan let say classical machine learning anomaly detection."
  },
  {
    "start": 4071.0,
    "end": 4071.7,
    "text": "Yeah, that's interesting."
  },
  {
    "start": 4071.98,
    "end": 4080.88,
    "text": "I'm not sure if you know the story it because in two years back or maybe no one and a half years back also when i had like my skeptics face so to speak?"
  },
  {
    "start": 4085.08,
    "end": 4092.12,
    "text": "really had the impression, I don't know a single like productive system that does anomaly detection with machine learning."
  },
  {
    "start": 4092.5,
    "end": 4101.08,
    "text": "And then I went to AI in The Forest with you and it was really an ad hoc two people and they were saying yeah there on since yet typically doesn't work."
  },
  {
    "start": 4101.439,
    "end": 4105.359,
    "text": "or But yeah Yeah It wasn't really an eye opener for me."
  },
  {
    "start": 4105.54,
    "end": 4107.68,
    "text": "They said most but have to be more specific right?"
  },
  {
    "start": 4107.979,
    "end": 4111.0,
    "text": "This head okay here it works But it's not worth."
  },
  {
    "start": 4111.52,
    "end": 4115.059,
    "text": "It just too costly to implement, to maintain and the use cases are not"
  },
  {
    "start": 4115.14,
    "end": 4115.359,
    "text": "there.\"."
  },
  {
    "start": 4116.2,
    "end": 4129.74,
    "text": "And from that point I actually had say okay i have go this group again because finally they're some real talk going on and everybody is doing marketing slides like oh yeah we now do anomaly textual blah blah blah right?"
  },
  {
    "start": 4131.8,
    "end": 4132.439,
    "text": "Why didn't?"
  },
  {
    "start": 4134.1,
    "end": 4135.88,
    "text": "I mean, obviously how you would frame it."
  },
  {
    "start": 4136.02,
    "end": 4140.22,
    "text": "I would frame like this typically people were drowning in the false positives."
  },
  {
    "start": 4140.58,
    "end": 4140.899,
    "text": "Why?"
  },
  {
    "start": 4141.42,
    "end": 4146.64,
    "text": "Because It was just and we can go back to what we discussed at beginning Was just narrow AI."
  },
  {
    "start": 4146.979,
    "end": 4158.059,
    "text": "so if You had an anomaly detection for example on on acoustics right On a machine And somebody Just dropped something near The Machine Like immediately get an alert because its just different Right!"
  },
  {
    "start": 4158.96,
    "end": 4161.02,
    "text": "And the system doesn't know why Its Different."
  },
  {
    "start": 4161.08,
    "end": 4162.559,
    "text": "It just sees out of distribution."
  },
  {
    "start": 4164.0,
    "end": 4168.52,
    "text": "And now with the context knowledge, it's just so amazing."
  },
  {
    "start": 4168.62,
    "end": 4180.04,
    "text": "The way we do anomaly detection right now is typically as a data scientist would go and look at histograms that can actually also incorporate classical machine learning."
  },
  {
    "start": 4180.7,
    "end": 4192.84,
    "text": "The bottom line is, once you have some candidates for outliers or candidates for anomalies the system understands."
  },
  {
    "start": 4192.979,
    "end": 4200.28,
    "text": "It says Hey, we have here a speed of one thousand kilometers per hour."
  },
  {
    "start": 4200.62,
    "end": 4202.14,
    "text": "that's unrealistic for our car."
  },
  {
    "start": 4202.9,
    "end": 4205.96,
    "text": "It has to be a sensor error and stuff like that."
  },
  {
    "start": 4206.08,
    "end": 4207.12,
    "text": "And what it says?"
  },
  {
    "start": 4207.4,
    "end": 4210.28,
    "text": "How does this relate with the use case is mind-blowing."
  },
  {
    "start": 4210.42,
    "end": 4212.64,
    "text": "I often sit in session with customers."
  },
  {
    "start": 4213.0,
    "end": 4231.92,
    "text": "they say well... The model cannot have knowledge about whatever specific test or product actually knows about it, how he says hey maybe you should fix magnet number three in your ten stator configuration or whatever."
  },
  {
    "start": 4232.1,
    "end": 4232.82,
    "text": "It's like what?"
  },
  {
    "start": 4234.059,
    "end": 4236.6,
    "text": "So yeah and I think that is the main difference."
  },
  {
    "start": 4236.88,
    "end": 4238.4,
    "text": "You have global context."
  },
  {
    "start": 4238.88,
    "end": 4253.04,
    "text": "Yeah very good examples And its kind of really putting a stage before human looks at it right many times, even knowing more than a human."
  },
  {
    "start": 4253.2,
    "end": 4257.88,
    "text": "I could come with examples but i don't need to where...I can then very well imagine."
  },
  {
    "start": 4258.04,
    "end": 4270.22,
    "text": "But let me first understand and that is typically foundational model knowledge That you have to enrich With the data of your environment Of whatever it is You know?"
  },
  {
    "start": 4270.3,
    "end": 4273.72,
    "text": "Your cars or your production Or your client."
  },
  {
    "start": 4274.38,
    "end": 4274.96,
    "text": "What does that?"
  },
  {
    "start": 4275.42,
    "end": 4276.82,
    "text": "Where does that knowledge Come from?"
  },
  {
    "start": 4277.04,
    "end": 4278.82,
    "text": "The agent has access."
  },
  {
    "start": 4279.88,
    "end": 4280.54,
    "text": "Yeah, interestingly."
  },
  {
    "start": 4280.74,
    "end": 4290.42,
    "text": "I think it's very interesting point that the knowledge about this specific domain in the sense how is a certain test structured what is tested and so on and so forth?"
  },
  {
    "start": 4290.84,
    "end": 4298.54,
    "text": "The model knows basically everything And actually there's big difference between like small models and large frontier models."
  },
  {
    "start": 4298.64,
    "end": 4312.34,
    "text": "i think its exactly where the model always know alot but we have to give and where to find stuff, how to call tools very efficiently."
  },
  {
    "start": 4312.64,
    "end": 4316.8,
    "text": "And I think again it's a very interesting note that efficiency is super-super important."
  },
  {
    "start": 4316.86,
    "end": 4338.68,
    "text": "so if you use a Claude Frontier model amazingly good, but you just don't want or cannot afford to run it for like thirty minutes when the task is so simple yet if you know the right stuff."
  },
  {
    "start": 4339.84,
    "end": 4347.26,
    "text": "You can do in ten seconds and then that's again part of the harness to provide this information Yeah."
  },
  {
    "start": 4351.14,
    "end": 4351.18,
    "text": "Or"
  },
  {
    "start": 4353.72,
    "end": 4361.82,
    "text": "your finance person comes in and believes that the bill for entropic or whoever is, isn't normally but it's not."
  },
  {
    "start": 4362.32,
    "end": 4363.059,
    "text": "Yeah,"
  },
  {
    "start": 4368.72,
    "end": 4369.22,
    "text": "absolutely."
  },
  {
    "start": 4369.72,
    "end": 4371.88,
    "text": "Absolutely I think it's a super big point."
  },
  {
    "start": 4372.34,
    "end": 4381.54,
    "text": "we are at or close to this point where if you had unlimited tokens You could basically build whatever you want but you don't have unlimited tokens."
  },
  {
    "start": 4381.96,
    "end": 4387.6,
    "text": "So i have to think about how to spend it wisely and like using like a specialized harness Like onion?"
  },
  {
    "start": 4391.1,
    "end": 4393.54,
    "text": "Okay, back to the harness general."
  },
  {
    "start": 4393.6,
    "end": 4396.34,
    "text": "how does onion deal with deployment security?"
  },
  {
    "start": 4396.68,
    "end": 4396.8,
    "text": "Yeah."
  },
  {
    "start": 4396.86,
    "end": 4398.44,
    "text": "That's a super big problem."
  },
  {
    "start": 4398.68,
    "end": 4399.98,
    "text": "It's probably also right now."
  },
  {
    "start": 4400.02,
    "end": 4409.78,
    "text": "I mean we see like a lot of stuff going on and in this realm And maybe you've also followed these different supply chain attacks obviously because everybody uses agents."
  },
  {
    "start": 4409.9,
    "end": 4413.36,
    "text": "Now it's much harder at some point or two."
  },
  {
    "start": 4413.44,
    "end": 4419.72,
    "text": "make sure your rigorous processes in place that Obviously people are doing more."
  },
  {
    "start": 4420.42,
    "end": 4427.14,
    "text": "obviously our tool, there's like a lot of certifications that also mandate to do this."
  },
  {
    "start": 4427.18,
    "end": 4428.74,
    "text": "I think it is super reasonable."
  },
  {
    "start": 4429.3,
    "end": 4435.18,
    "text": "And then this is like where we, I think have to make sure that in the development process already."
  },
  {
    "start": 4436.54,
    "end": 4440.04,
    "text": "We are taking care of all these things and they architecture."
  },
  {
    "start": 4440.8,
    "end": 4447.7,
    "text": "obviously agents can do a lot stuff and I think the analogy quite fitting."
  },
  {
    "start": 4447.92,
    "end": 4457.42,
    "text": "some people say hey an agent super intelligent but slightly drunk intern it would go on delete your entire production database."
  },
  {
    "start": 4457.54,
    "end": 4463.74,
    "text": "maybe It can make your code like a thirty percent faster, but just to test it would just delete everything."
  },
  {
    "start": 4463.88,
    "end": 4465.58,
    "text": "There's no problem with that and I think its true."
  },
  {
    "start": 4466.4,
    "end": 4468.96,
    "text": "so the only... Yeah!"
  },
  {
    "start": 4469.28,
    "end": 4475.34,
    "text": "And everybody who codes with agent has these stories for themselves."
  },
  {
    "start": 4475.4,
    "end": 4477.4,
    "text": "For example you have a secret right?"
  },
  {
    "start": 4477.82,
    "end": 4482.64,
    "text": "You hide this secret from an agent With some mechanisms where they cant see in their IDE."
  },
  {
    "start": 4482.84,
    "end": 4488.7,
    "text": "But then the agent come up other tools The ways how he tries access it anyway."
  },
  {
    "start": 4490.2,
    "end": 4492.5,
    "text": "It's really strange or really, really alarming."
  },
  {
    "start": 4492.58,
    "end": 4497.4,
    "text": "so the only way you can make this secure is to sandbox it heavily."
  },
  {
    "start": 4497.52,
    "end": 4499.18,
    "text": "and that what we're doing also in Onions."
  },
  {
    "start": 4499.34,
    "end": 4513.6,
    "text": "You have good sandboxes around everything even if for example they build a bad query system cannot break out of container stuff like"
  },
  {
    "start": 4516.88,
    "end": 4528.54,
    "text": "work while still you need to have access to all the interesting information because without having the access, the agent cannot be smart."
  },
  {
    "start": 4528.64,
    "end": 4531.24,
    "text": "So is that a contradict?"
  },
  {
    "start": 4532.0,
    "end": 4532.76,
    "text": "Yeah, absolutely."
  },
  {
    "start": 4533.059,
    "end": 4550.58,
    "text": "Absolutely there's definitely a trade-off and I also found it...I actually think about this a lot And i found very interesting that to me currently the best abstraction is just you have a folder That how typically IDEs work or Also how Claude co-work?"
  },
  {
    "start": 4550.64,
    "end": 4553.42,
    "text": "It has project which are also just a photo in the folder."
  },
  {
    "start": 4553.5,
    "end": 4560.22,
    "text": "You can basically do whatever you want like that, but you cannot access for example the internet to prevent leakage."
  },
  {
    "start": 4560.3,
    "end": 4563.76,
    "text": "and in Onion we have a slightly more abstract powerful concept."
  },
  {
    "start": 4563.82,
    "end": 4565.1,
    "text": "We call it data container."
  },
  {
    "start": 4565.7,
    "end": 4571.559,
    "text": "And what you can do is with these data containers, you can control the agent."
  },
  {
    "start": 4571.88,
    "end": 4575.6,
    "text": "First of all like Cs how will the agent understand their data?"
  },
  {
    "start": 4576.08,
    "end": 4577.6,
    "text": "But also that security aspect."
  },
  {
    "start": 4577.72,
    "end": 4578.76,
    "text": "I think it's super important."
  },
  {
    "start": 4578.8,
    "end": 4600.52,
    "text": "for example on a database Super simple but You make it read only for certain data container right so i can Make sure It can't mess with That whereas maybe have another Data Container where you Can write stuff too and having sandboxing On this level I think it's the only way to make sure that system is secure."
  },
  {
    "start": 4601.16,
    "end": 4609.059,
    "text": "Very good, i read about this option of bringing your own LLM... ...I can imagine what."
  },
  {
    "start": 4609.34,
    "end": 4615.84,
    "text": "maybe you could answer us and combine with talking about your license model."
  },
  {
    "start": 4615.88,
    "end": 4616.98,
    "text": "just get an idea."
  },
  {
    "start": 4617.16,
    "end": 4618.7,
    "text": "or interested parties."
  },
  {
    "start": 4618.76,
    "end": 4619.08,
    "text": "how does?"
  },
  {
    "start": 4620.02,
    "end": 4620.6,
    "text": "Yeah, absolutely."
  },
  {
    "start": 4622.3,
    "end": 4626.72,
    "text": "First of all we really strive for onion to be I would say cheap."
  },
  {
    "start": 4626.98,
    "end": 4644.96,
    "text": "so it's something like every engineer can use buy seat, classical by-seat license where we also feel like hey you pay by seed but then you save so much on query time token usage and so forth that it's not a big investment."
  },
  {
    "start": 4645.52,
    "end": 4651.36,
    "text": "And the actual LLMs are in our case typically from our customers."
  },
  {
    "start": 4652.08,
    "end": 4663.22,
    "text": "And there, our customers have their cloud environment and they use their endpoints like a GPT."
  },
  {
    "start": 4663.5,
    "end": 4667.42,
    "text": "five point four, whatever like is available there."
  },
  {
    "start": 4667.82,
    "end": 4673.6,
    "text": "And it's for I think engineering software the best deployment path for several reasons."
  },
  {
    "start": 4674.16,
    "end": 4683.22,
    "text": "There are a lot of sensitive IPs in this system so its important to not have cloud based systems that controlled by vendor but controlled them."
  },
  {
    "start": 4683.88,
    "end": 4695.84,
    "text": "and also important And that's by this bringing your own LLM, like hosting it in your own cloud tenant is so important."
  },
  {
    "start": 4696.059,
    "end": 4703.58,
    "text": "It's also important because when we do data analysis on hundreds of terabytes of data, data locality is everything."
  },
  {
    "start": 4703.7,
    "end": 4711.6,
    "text": "So you definitely have to have the very same compute center where for example onion or harness runs."
  },
  {
    "start": 4711.78,
    "end": 4715.96,
    "text": "otherwise you had transfer data which would be super slow and super expensive."
  },
  {
    "start": 4716.68,
    "end": 4719.22,
    "text": "So that's why bring your own LLM is very important."
  },
  {
    "start": 4719.78,
    "end": 4721.54,
    "text": "Very good, thanks for the now."
  },
  {
    "start": 4721.96,
    "end": 4724.72,
    "text": "one thing I was thinking of like what?"
  },
  {
    "start": 4724.8,
    "end": 4727.559,
    "text": "Is What is your renumix business model?"
  },
  {
    "start": 4727.82,
    "end": 4729.3,
    "text": "and so you said You're gonna."
  },
  {
    "start": 4729.559,
    "end": 4733.84,
    "text": "this is something that will provide to engineers."
  },
  {
    "start": 4734.16,
    "end": 4736.8,
    "text": "they would be typically Customer engineers."
  },
  {
    "start": 4736.96,
    "end": 4747.88,
    "text": "when i say customer uh services to your customers like so you're being paid by the hour, but a week or whatever."
  },
  {
    "start": 4747.96,
    "end": 4748.18,
    "text": "You do."
  },
  {
    "start": 4748.74,
    "end": 4750.8,
    "text": "is that a combination of two things?"
  },
  {
    "start": 4750.86,
    "end": 4752.24,
    "text": "That you provide to the market?"
  },
  {
    "start": 4752.72,
    "end": 4755.26,
    "text": "and where do you see that moving?"
  },
  {
    "start": 4755.34,
    "end": 4758.18,
    "text": "Do you see it changing over the next year?"
  },
  {
    "start": 4758.28,
    "end": 4761.559,
    "text": "And then we'll gonna be coming back to their topic at The End Of Our Talk."
  },
  {
    "start": 4761.98,
    "end": 4762.82,
    "text": "Yeah yeah absolutely."
  },
  {
    "start": 4762.92,
    "end": 4777.559,
    "text": "So I think its also very interesting question because traditionally done like a lot of bespoke custom systems for our customers and we increasingly see that there is some standardization happening."
  },
  {
    "start": 4777.7,
    "end": 4785.28,
    "text": "We also see actually which was super cool, more engineers or domain experts start to vibe code something right?"
  },
  {
    "start": 4785.62,
    "end": 4791.6,
    "text": "So I think it will always be the need kind of customizations that we gladly do."
  },
  {
    "start": 4791.74,
    "end": 4795.639,
    "text": "We love to do, I think it's also where a lot of innovation happens right?"
  },
  {
    "start": 4795.76,
    "end": 4808.26,
    "text": "All these standardized software is only developed in other places, because you first have projects where we do real innovative stuff and see the new path forward."
  },
  {
    "start": 4808.34,
    "end": 4812.1,
    "text": "And that's what we are happily doing with our customers."
  },
  {
    "start": 4812.66,
    "end": 4825.66,
    "text": "Then there is actually one area where currently like a lot of demand also feel it important to help people which exactly prepare this engineering data infrastructure for Agentica AI."
  },
  {
    "start": 4826.82,
    "end": 4837.52,
    "text": "Currently a lot of people are moving from old on-premise systems, private systems to the lake house and then it's always same management says hey move everything to fabric or whatever."
  },
  {
    "start": 4837.96,
    "end": 4841.86,
    "text": "And engineers have to figure out how do we do like other data models?"
  },
  {
    "start": 4844.68,
    "end": 4846.24,
    "text": "We did wrong in the past, quite frankly."
  },
  {
    "start": 4846.36,
    "end": 4852.24,
    "text": "So you can do wrong and if you do a couple of things wrong then it's really difficult."
  },
  {
    "start": 4852.32,
    "end": 4859.98,
    "text": "for example If like You have the wrong partitioning And It gives you three times or slow down Three time and make two other mistakes."
  },
  {
    "start": 4860.02,
    "end": 4866.1,
    "text": "Then it gets multiplied at end ten times slower than normal."
  },
  {
    "start": 4866.44,
    "end": 4868.16,
    "text": "So I think that's a big area."
  },
  {
    "start": 4868.3,
    "end": 4876.559,
    "text": "and then, the overall market in custom software is like yeah let's see where it goes because you have the wipe coding stuff."
  },
  {
    "start": 4876.7,
    "end": 4883.44,
    "text": "maybe that will fade but who are much quicker in developing solutions."
  },
  {
    "start": 4883.82,
    "end": 4885.54,
    "text": "Yeah, let's see where this goes."
  },
  {
    "start": 4886.22,
    "end": 4887.46,
    "text": "here we're gonna come back to these."
  },
  {
    "start": 4887.52,
    "end": 4887.8,
    "text": "anyway."
  },
  {
    "start": 4887.9,
    "end": 4890.28,
    "text": "now We talked about bring your own LLM."
  },
  {
    "start": 4891.0,
    "end": 4897.22,
    "text": "share with us your experience With small language models In your engineering environment."
  },
  {
    "start": 4898.4,
    "end": 4902.059,
    "text": "Yeah, I think it's also something where i think a lot about."
  },
  {
    "start": 4902.16,
    "end": 4904.62,
    "text": "and maybe why can't also be like very wrong?"
  },
  {
    "start": 4904.76,
    "end": 4911.52,
    "text": "but right now we can actually see that the frontier models large language models are much better than smaller ones."
  },
  {
    "start": 4911.6,
    "end": 4916.62,
    "text": "There are some tasks where the small ones like a good enough and they're a lot cheaper."
  },
  {
    "start": 4916.7,
    "end": 4918.559,
    "text": "And a lot faster, that's really important."
  },
  {
    "start": 4918.66,
    "end": 4923.639,
    "text": "so maybe the idea of having these sub agents with smaller models is good."
  },
  {
    "start": 4923.78,
    "end": 4930.2,
    "text": "but at the same time we actually see currently you just have one flat loop With big model."
  },
  {
    "start": 4930.7,
    "end": 4933.4,
    "text": "You use a lot token caching to make it efficient."
  },
  {
    "start": 4933.96,
    "end": 4935.639,
    "text": "The model sees the whole context."
  },
  {
    "start": 4935.78,
    "end": 4939.74,
    "text": "don't need to put a smaller context for the smaller agent or smaller model."
  },
  {
    "start": 4941.2,
    "end": 4943.82,
    "text": "so well that, yeah I'm not sure."
  },
  {
    "start": 4943.92,
    "end": 4951.5,
    "text": "We will see if we actually really incorporate these model models a lot or if it will center on the larger ones?"
  },
  {
    "start": 4952.3,
    "end": 4956.62,
    "text": "What I especially maybe what sometimes have to admit is you'll get angry."
  },
  {
    "start": 4957.02,
    "end": 4962.2,
    "text": "there's this story in Germany about large models."
  },
  {
    "start": 4962.559,
    "end": 4966.3,
    "text": "they are expensive and compute-intensive."
  },
  {
    "start": 4966.38,
    "end": 4970.0,
    "text": "energy intensive train obviously can't catch anymore"
  },
  {
    "start": 4970.4,
    "end": 4970.6,
    "text": "But"
  },
  {
    "start": 4971.02,
    "end": 4976.0,
    "text": "we have this highly specialized knowledge and we can train it into small language models."
  },
  {
    "start": 4976.42,
    "end": 4977.32,
    "text": "And then, we win!"
  },
  {
    "start": 4979.36,
    "end": 4983.92,
    "text": "At least right now... Sorry to say that's just a bullshit thing."
  },
  {
    "start": 4986.5,
    "end": 4994.24,
    "text": "Especially in the domain-knowledge stuff, large models are just the best because they're the strongest AI."
  },
  {
    "start": 4994.52,
    "end": 5001.76,
    "text": "Everything else will be... to come back that more narrow AI and it's just definitely right now."
  },
  {
    "start": 5001.82,
    "end": 5007.62,
    "text": "The lesson to be learned is like the larger you know, more You have a better context."
  },
  {
    "start": 5008.18,
    "end": 5009.3,
    "text": "this Is winning?"
  },
  {
    "start": 5009.66,
    "end": 5015.68,
    "text": "And we should not I think Just feel hey let's build A lot of small language models then We Have a chance."
  },
  {
    "start": 5015.88,
    "end": 5016.86,
    "text": "i Think It's a little bit"
  },
  {
    "start": 5020.6,
    "end": 5026.559,
    "text": "Robert and I, we've been talking about industrial foundation models for five years."
  },
  {
    "start": 5028.639,
    "end": 5038.0,
    "text": "Meanwhile... in organizations that would, but I've heard there's many organizations now doing exactly that."
  },
  {
    "start": 5038.559,
    "end": 5056.48,
    "text": "That is one import and the second input is...that's another story about Europe, Germany also Italy France all other more industrial oriented countries saying we don't need to have models."
  },
  {
    "start": 5056.58,
    "end": 5066.0,
    "text": "you know train has left station But we sit on all this industrial data, which is the combination of both."
  },
  {
    "start": 5067.28,
    "end": 5071.8,
    "text": "And maybe if I combine those two in my brain... ...I'm interested in your view on that."
  },
  {
    "start": 5072.38,
    "end": 5078.62,
    "text": "it's like three years ago when Sam came with his large language model."
  },
  {
    "start": 5078.76,
    "end": 5083.46,
    "text": "he had been scraping the internet and nobody knew for a year or two, I guess."
  },
  {
    "start": 5084.04,
    "end": 5086.92,
    "text": "Now how has that worked in the meantime then?"
  },
  {
    "start": 5086.98,
    "end": 5088.9,
    "text": "And that's also the robotics story."
  },
  {
    "start": 5089.46,
    "end": 5093.36,
    "text": "you can't scrape industrial... Or industrial data."
  },
  {
    "start": 5093.44,
    "end": 5099.36,
    "text": "maybe you can scrape but the area of physical AI for engineers You can't really scrape that kind of thing."
  },
  {
    "start": 5099.72,
    "end": 5101.8,
    "text": "So it is little bit bigger thing."
  },
  {
    "start": 5101.86,
    "end": 5104.04,
    "text": "But do we need this industrial data?"
  },
  {
    "start": 5104.16,
    "end": 5109.0,
    "text": "or say well your experience the huge, big foundation models."
  },
  {
    "start": 5109.12,
    "end": 5113.16,
    "text": "they seem to have this kind of knowledge somehow in them anyway."
  },
  {
    "start": 5114.02,
    "end": 5115.02,
    "text": "Yeah, I think it's a good question."
  },
  {
    "start": 5115.1,
    "end": 5120.86,
    "text": "And definitely have to say like...I don't feel that i'm big expert on this so please take as an opinion."
  },
  {
    "start": 5121.04,
    "end": 5126.5,
    "text": "but what do the Frontier models or Frontier things have?"
  },
  {
    "start": 5126.559,
    "end": 5130.2,
    "text": "They all have books and knowledge available."
  },
  {
    "start": 5130.78,
    "end": 5143.6,
    "text": "then they also have clever reinforcement learning where there are coding puzzles with very general purpose problem solving capability Right."
  },
  {
    "start": 5144.2,
    "end": 5149.62,
    "text": "And, um like to replicate that is super difficult."
  },
  {
    "start": 5149.74,
    "end": 5153.3,
    "text": "and then obviously if you have some very company-specific knowledge."
  },
  {
    "start": 5153.36,
    "end": 5156.1,
    "text": "how exactly this test has run on all these things."
  },
  {
    "start": 5156.48,
    "end": 5158.04,
    "text": "You can just put it in the context."
  },
  {
    "start": 5166.82,
    "end": 5167.24,
    "text": "can help."
  },
  {
    "start": 5168.54,
    "end": 5170.74,
    "text": "But I mean, what should we do?"
  },
  {
    "start": 5171.26,
    "end": 5192.94,
    "text": "Should you say hey there's no initiatives anymore...I think that is also wrong as well so i don't want to like downplay it..i think its important try and maybe be on the topic have people here who do some kind of meaningful research but i think its super difficult ..and i think the narrative should be tried."
  },
  {
    "start": 5193.2,
    "end": 5194.94,
    "text": "yes ...but is this?"
  },
  {
    "start": 5195.02,
    "end": 5197.54,
    "text": "Hmm don't know."
  },
  {
    "start": 5199.0,
    "end": 5200.34,
    "text": "But that's very important."
  },
  {
    "start": 5200.38,
    "end": 5214.6,
    "text": "Yeah, so at least you know it's like It's not a given doing this whatever industrial foundation model with a handful or twenty I mean similar to the OPC Foundation has been doing in the past."
  },
  {
    "start": 5215.04,
    "end": 5219.58,
    "text": "You know twenty robotics companies and do an OPC information model The same way."
  },
  {
    "start": 5219.68,
    "end": 5229.32,
    "text": "we could sit together With one hundred thousand automation companies and make these industrial foundation model, maybe something like."
  },
  {
    "start": 5229.5,
    "end": 5230.8,
    "text": "I forgot the name now."
  },
  {
    "start": 5231.7,
    "end": 5235.44,
    "text": "The daughter company of BMW saw them again yesterday."
  },
  {
    "start": 5235.52,
    "end": 5235.88,
    "text": "they're doing."
  },
  {
    "start": 5236.7,
    "end": 5239.24,
    "text": "but then you say yeah may be it works?"
  },
  {
    "start": 5240.059,
    "end": 5240.22,
    "text": "But"
  },
  {
    "start": 5240.78,
    "end": 5241.86,
    "text": "that's not a given."
  },
  {
    "start": 5243.8,
    "end": 5247.66,
    "text": "Yeah, I think actually if you look at how these frontier labs are structured."
  },
  {
    "start": 5249.18,
    "end": 5254.38,
    "text": "You don't need one hundred different stakeholders with hundreds of people."
  },
  {
    "start": 5255.059,
    "end": 5259.86,
    "text": "it's more like you need fifty people who know exactly what they're doing and billions of compute."
  },
  {
    "start": 5261.24,
    "end": 5275.16,
    "text": "so yeah i think especially in this consortium Maybe, I mean obviously don't do nothing is also no option."
  },
  {
    "start": 5275.24,
    "end": 5286.96,
    "text": "but i really feel like For example, what I'm a big fan is really making sure that the data infrastructure layer and all these things we have like sovereign stacks."
  },
  {
    "start": 5287.38,
    "end": 5288.94,
    "text": "We have good best practices."
  },
  {
    "start": 5289.8,
    "end": 5293.62,
    "text": "We become data-driven fast."
  },
  {
    "start": 5294.48,
    "end": 5308.98,
    "text": "These are types of things especially in German industries should gather around built stacks that are independent from the big American companies."
  },
  {
    "start": 5309.059,
    "end": 5320.16,
    "text": "And I think there is a lot of open source, there's a lot cool new stuff we're also using internally and where you can actually build on the frontier."
  },
  {
    "start": 5320.24,
    "end": 5326.139,
    "text": "so build stuff better than other stuff And at the same time, open source and sovereign."
  },
  {
    "start": 5326.22,
    "end": 5331.96,
    "text": "So I would say like let's put this more in front on center than yeah."
  },
  {
    "start": 5332.08,
    "end": 5332.26,
    "text": "then the"
  },
  {
    "start": 5332.38,
    "end": 5333.82,
    "text": "LLM things right."
  },
  {
    "start": 5334.58,
    "end": 5339.04,
    "text": "you started giving a part of the answer to my final question."
  },
  {
    "start": 5339.44,
    "end": 5352.12,
    "text": "i always want to know from people that talk too And it kind of you already given maybe the answer to that, do they open source and don't be dependent on now."
  },
  {
    "start": 5352.44,
    "end": 5372.82,
    "text": "So let's talk about this future engineering around your offer which is I mean You see or you believe to know how as good as anybody else where the market And one way of reacting or acting is that you put a certain solution into the market."
  },
  {
    "start": 5372.9,
    "end": 5384.88,
    "text": "Now, let's start very specifically with do you see Jeff Bezos going to be a competitor Prometheus amateur,"
  },
  {
    "start": 5385.22,
    "end": 5385.76,
    "text": "yeah."
  },
  {
    "start": 5386.38,
    "end": 5389.04,
    "text": "Do you understand what it is that he's trying to do?"
  },
  {
    "start": 5389.2,
    "end": 5390.98,
    "text": "And are not going?"
  },
  {
    "start": 5391.1,
    "end": 5404.6,
    "text": "ask if your scared but maybe I see him as a potential partner or I don't know, do you have a feeling for what he is doing in our market?"
  },
  {
    "start": 5404.74,
    "end": 5405.54,
    "text": "He's going to be doing."
  },
  {
    "start": 5405.66,
    "end": 5405.82,
    "text": "Yeah"
  },
  {
    "start": 5406.12,
    "end": 5408.639,
    "text": "so and i don't...I Know nearly nothing about it."
  },
  {
    "start": 5408.7,
    "end": 5418.26,
    "text": "So but there is actually this general trend In Silicon Valley To not just build AI software But Just provide the service."
  },
  {
    "start": 5418.34,
    "end": 5431.02,
    "text": "For example if I Provide like A text consulting Service for Example I Don't Build Software for Text Consultants run a text consultancy, but internally I build all my processes with AI."
  },
  {
    "start": 5431.5,
    "end": 5445.46,
    "text": "So become like the service and not built to software is something that it's big mantra currently in Silicon Valley because then you capture more of value and have everything internally You can iterate faster right?"
  },
  {
    "start": 5445.54,
    "end": 5461.74,
    "text": "All these things And i think It seems similar play In the sense That its not a software company But maybe like a company planning to buy older or like established industry players, but then revamp them with AI."
  },
  {
    "start": 5462.26,
    "end": 5467.54,
    "text": "And I think on paper it's probably a very interesting idea."
  },
  {
    "start": 5468.24,
    "end": 5477.52,
    "text": "I don't want to discount it, but obviously its also from players who have little experience in the field."
  },
  {
    "start": 5477.639,
    "end": 5491.58,
    "text": "and just looking back at history for example maybe you remember Elon Musk's The Boring Company where he thought that he could replicate Herrenknecht quickly or what Google has been doing."
  },
  {
    "start": 5491.72,
    "end": 5500.58,
    "text": "The ideas are probably very good, but it also turns out like execution is really hard."
  },
  {
    "start": 5500.66,
    "end": 5504.74,
    "text": "And there's a lot where the domain knowledge experience and value networks coming in."
  },
  {
    "start": 5505.84,
    "end": 5523.08,
    "text": "so for me personally I'd say we try to deliver one puzzle piece super well data analysis, especially time series data analysis for engineering and really try to focus on the puzzle piece."
  },
  {
    "start": 5523.18,
    "end": 5527.54,
    "text": "And there are a lot more puzzle pieces that are solved by other people."
  },
  {
    "start": 5528.22,
    "end": 5535.26,
    "text": "If Jeff Bezos wants do it with billions like in very end-to-end fashion let's see where is goes."
  },
  {
    "start": 5535.74,
    "end": 5538.44,
    "text": "Okay thanks yeah Thanks for that."
  },
  {
    "start": 5538.54,
    "end": 5540.9,
    "text": "It always difficult be looking into."
  },
  {
    "start": 5542.52,
    "end": 5544.559,
    "text": "maybe I thought something similar."
  },
  {
    "start": 5544.82,
    "end": 5550.72,
    "text": "he was going As I just said, as you could scrape text from the internet."
  },
  {
    "start": 5550.82,
    "end": 5557.16,
    "text": "You can't scrape physically maybe more like that engineering a little bit but not really."
  },
  {
    "start": 5557.6,
    "end": 5560.34,
    "text": "and i was thinking as well That Maybe he Was going to be buying."
  },
  {
    "start": 5560.8,
    "end": 5567.08,
    "text": "But then coming back To The Discussion we Just had before And He Also Going to Be Buying Industrial Data."
  },
  {
    "start": 5567.32,
    "end": 5576.92,
    "text": "Now Believe Here You Say That Maybe He's Buying More Than Knowledge The Capabilities of People inside of such potential companies."
  },
  {
    "start": 5577.08,
    "end": 5584.1,
    "text": "and here you say that maybe he's going to then reinvent him becoming AI first."
  },
  {
    "start": 5584.76,
    "end": 5595.94,
    "text": "Yes, yes I think it is a plan buying like real entities everything not just some data package but basically the whole thing with all people or processes at least how i understand this might be wrong."
  },
  {
    "start": 5596.26,
    "end": 5596.8,
    "text": "disclaimer"
  },
  {
    "start": 5597.68,
    "end": 5598.58,
    "text": "Let's close."
  },
  {
    "start": 5599.02,
    "end": 5600.28,
    "text": "I think we passed the hour."
  },
  {
    "start": 5600.66,
    "end": 5613.18,
    "text": "Robert is going to be very happy with us again, tell us quickly if you want to about your team where you based and maybe If You're looking for talent just in case Everybody anybody would be interested in joining?"
  },
  {
    "start": 5614.0,
    "end": 5614.48,
    "text": "Yeah Thanks."
  },
  {
    "start": 5614.86,
    "end": 5629.0,
    "text": "We are based in Karlsruhe were ten people so they have this good size of what in software engineering will call like The Two Pizza Team If people want to do like some internship, for example where we're always open for new colleagues."
  },
  {
    "start": 5629.059,
    "end": 5636.7,
    "text": "but it's also interesting would be another podcast that at the moment because AI coding and everything is making us so fast."
  },
  {
    "start": 5637.22,
    "end": 5643.34,
    "text": "It's interesting despite having a very good growth with ONION are not looking for"
  },
  {
    "start": 5643.84,
    "end": 5643.96,
    "text": "new"
  },
  {
    "start": 5644.46,
    "end": 5645.38,
    "text": "full-time team members"
  },
  {
    "start": 5645.9,
    "end": 5651.24,
    "text": "oh that's a very interesting finish here and yes we have talked about."
  },
  {
    "start": 5653.52,
    "end": 5655.76,
    "text": "We will continue to talk about it before."
  },
  {
    "start": 5672.28,
    "end": 5685.139,
    "text": "Bye bye."
  }
]