[
  {
    "start": 0.0,
    "end": 2.2,
    "text": "AI is completely stupid."
  },
  {
    "start": 3.36,
    "end": 16.74,
    "text": "It has no clue of anything, you're just trying to direct the information... ...to an area where maybe someone trained or gave that it could use."
  },
  {
    "start": 18.12,
    "end": 31.08,
    "text": "You can then start forgetting chatGPT, Klot, Gemini And forget all those tools because The data is not under your control."
  },
  {
    "start": 32.54,
    "end": 40.42,
    "text": "People talk that they are using agentic systems and have already implemented it, but actually they're talking about agents."
  },
  {
    "start": 41.88,
    "end": 46.88,
    "text": "It's no artificial intelligence really."
  },
  {
    "start": 47.34,
    "end": 49.16,
    "text": "There IS NO INTELLIGENCE in there!"
  },
  {
    "start": 49.6,
    "end": 52.48,
    "text": "It just gathering information and handing over to you."
  },
  {
    "start": 53.48,
    "end": 55.72,
    "text": "Welcome to Software Testing Unleashed."
  },
  {
    "start": 55.88,
    "end": 61.76,
    "text": "My name is Richie and I have another episode from the Quality Land Conference in Vilnius for you!"
  },
  {
    "start": 63.08,
    "end": 92.64,
    "text": "Hi Nisha, great to have you on the show here."
  },
  {
    "start": 93.3,
    "end": 94.42,
    "text": "Thank You for the invitation."
  },
  {
    "start": 94.58,
    "end": 95.6,
    "text": "it's an honor to be Here."
  },
  {
    "start": 95.84,
    "end": 97.92,
    "text": "yeah we always meet in different places."
  },
  {
    "start": 98.02,
    "end": 99.14,
    "text": "We are now he and will news."
  },
  {
    "start": 100.08,
    "end": 101.46,
    "text": "I stick to be conference here."
  },
  {
    "start": 101.78,
    "end": 101.9,
    "text": "Yeah"
  },
  {
    "start": 102.22,
    "end": 103.76,
    "text": "And we're both speakers here too?"
  },
  {
    "start": 103.98,
    "end": 108.72,
    "text": "And so i grabbed opportunity To get you into my podcast here."
  },
  {
    "start": 109.6,
    "end": 109.9,
    "text": "thank ya."
  },
  {
    "start": 110.04,
    "end": 114.58,
    "text": "i mean last time we met at The other part of the world yeah In the south."
  },
  {
    "start": 115.32,
    "end": 116.82,
    "text": "yeah more issues Yeah,"
  },
  {
    "start": 117.12,
    "end": 119.38,
    "text": "but usually we live nearby."
  },
  {
    "start": 119.5,
    "end": 120.52,
    "text": "You are in Switzerland?"
  },
  {
    "start": 120.72,
    "end": 121.6,
    "text": "I'm in Germany so..."
  },
  {
    "start": 122.62,
    "end": 124.54,
    "text": "But we never meet in Germany or Switzerland."
  },
  {
    "start": 125.06,
    "end": 125.34,
    "text": "That's"
  },
  {
    "start": 126.44,
    "end": 127.54,
    "text": "true!"
  },
  {
    "start": 127.58,
    "end": 143.2,
    "text": "We want to talk about agentic systems today and i think it is a very important topic because lots of people are talking about that aspects and also definitions for that, what's about it?"
  },
  {
    "start": 143.66,
    "end": 145.14,
    "text": "And so maybe we start with that."
  },
  {
    "start": 145.78,
    "end": 146.68,
    "text": "What is an agentic"
  },
  {
    "start": 146.76,
    "end": 147.14,
    "text": "system?\"."
  },
  {
    "start": 147.98,
    "end": 150.3,
    "text": "So... Maybe we need to make it clear."
  },
  {
    "start": 150.4,
    "end": 158.64,
    "text": "a lot of people talk they are using agentics systems but actually talking about agents."
  },
  {
    "start": 159.36,
    "end": 164.38,
    "text": "They have some agents doing somethings But its not fully agentic."
  },
  {
    "start": 166.44,
    "end": 174.44,
    "text": "When we talk about a genetic system, then you're talking about agents orchestrating agents by themselves."
  },
  {
    "start": 175.24,
    "end": 179.72,
    "text": "So there are no humans that actually giving the tasks."
  },
  {
    "start": 180.26,
    "end": 182.74,
    "text": "The agents are performing themselfs."
  },
  {
    "start": 183.3,
    "end": 185.48,
    "text": "they're goal-based and not task based."
  },
  {
    "start": 187.0,
    "end": 189.54,
    "text": "people just get confused with it."
  },
  {
    "start": 191.18,
    "end": 192.94,
    "text": "maybe its time to clarify."
  },
  {
    "start": 195.3,
    "end": 198.96,
    "text": "it's a full ecosystem containing agents."
  },
  {
    "start": 200.52,
    "end": 208.92,
    "text": "And I think that the human in there is not as important than agent, so agent you give a task and he does it then its finished?"
  },
  {
    "start": 209.38,
    "end": 210.86,
    "text": "Well... Not really!"
  },
  {
    "start": 210.98,
    "end": 229.16,
    "text": "So.. There was first intention having a full AI system doing everything for you so that humans just can relax monitor look at everything running perfectly and you just enjoy what the AI does, and you trust the AI."
  },
  {
    "start": 229.94,
    "end": 230.84,
    "text": "And it's not the case."
  },
  {
    "start": 230.9,
    "end": 245.8,
    "text": "so we completely failed on that one because we realized that The AI is doing a great job but there are some points where You need to be the judge for the AI."
  },
  {
    "start": 246.36,
    "end": 253.06,
    "text": "So the human in-the loop is essential But you need to define where to put them in the right manner."
  },
  {
    "start": 253.66,
    "end": 257.38,
    "text": "And if you want a human too, overwrite or not?"
  },
  {
    "start": 259.519,
    "end": 264.34,
    "text": "When we try give an example for their community and who are testers?"
  },
  {
    "start": 264.4,
    "end": 265.34,
    "text": "quality people do."
  },
  {
    "start": 265.42,
    "end": 265.7,
    "text": "they have."
  },
  {
    "start": 267.42,
    "end": 269.64,
    "text": "use case of us for an agentic system."
  },
  {
    "start": 269.78,
    "end": 271.54,
    "text": "in our quality testing"
  },
  {
    "start": 272.28,
    "end": 273.86,
    "text": "branch We actually have one."
  },
  {
    "start": 273.96,
    "end": 286.14,
    "text": "so... In our company we had full agentics systems where you have all the tasks that a tester does, but it's AI doing."
  },
  {
    "start": 286.74,
    "end": 290.32,
    "text": "The human is only in loop to make sure they're still working their right way."
  },
  {
    "start": 290.94,
    "end": 299.52,
    "text": "So let us say we have a development department and a company which develops everything into house."
  },
  {
    "start": 301.2,
    "end": 302.76,
    "text": "Normally how do this work?"
  },
  {
    "start": 302.88,
    "end": 308.2,
    "text": "You have somebody who writes requirements either HL or waterfall matter."
  },
  {
    "start": 309.68,
    "end": 315.02,
    "text": "Then the testers need to review it, they need to make sure that everything is consistent."
  },
  {
    "start": 316.02,
    "end": 318.9,
    "text": "Afterwards you hand over to the developers."
  },
  {
    "start": 318.98,
    "end": 323.1,
    "text": "in the meantime the testors write task cases and now think about it."
  },
  {
    "start": 323.92,
    "end": 341.12,
    "text": "You have a system where The requirement engineer wrote the requirements And once he submits on Girol let's say then the agentic system recognizes, hey there is a new requirement."
  },
  {
    "start": 341.88,
    "end": 356.2,
    "text": "I will go ahead and pull it and review it or you as a tester can say pulled latest requirements that are in the scope of let's say release twenty-five."
  },
  {
    "start": 357.12,
    "end": 367.78,
    "text": "Then AI takes their requirements starts reviewing Okay, there are some things that aren't correct here."
  },
  {
    "start": 367.88,
    "end": 368.46,
    "text": "We need to fix"
  },
  {
    "start": 368.62,
    "end": 368.72,
    "text": "it.\"."
  },
  {
    "start": 369.68,
    "end": 371.92,
    "text": "It shows the output of the human."
  },
  {
    "start": 373.22,
    "end": 379.36,
    "text": "The tester has now a responsibility to check if he meets or not."
  },
  {
    "start": 380.16,
    "end": 384.32,
    "text": "If they meet then he can just approve and continue."
  },
  {
    "start": 385.26,
    "end": 386.74,
    "text": "Then what should we do?"
  },
  {
    "start": 387.88,
    "end": 396.64,
    "text": "go ahead and write new test cases so that your test set increases from two hundred to five hundred, six hundred to eight hundred."
  },
  {
    "start": 397.14,
    "end": 399.18,
    "text": "And then you're testing months-and-months."
  },
  {
    "start": 400.46,
    "end": 405.66,
    "text": "actually no what it does is checks the coverage."
  },
  {
    "start": 406.4,
    "end": 411.46,
    "text": "So Is there any case can cover those requirements?"
  },
  {
    "start": 416.14,
    "end": 423.0,
    "text": "map the requirement to those test cases and Make sure that the coverage is still Align."
  },
  {
    "start": 423.48,
    "end": 429.02,
    "text": "if not The agent will then tell That it's responsible for the coverage."
  },
  {
    "start": 429.56,
    "end": 433.18,
    "text": "That there are some more test cases to be created."
  },
  {
    "start": 434.72,
    "end": 435.9,
    "text": "Do you agree or not?"
  },
  {
    "start": 436.4,
    "end": 440.68,
    "text": "So again, the tester comes in to do loop And checks."
  },
  {
    "start": 440.92,
    "end": 442.06,
    "text": "does it make sense or not?"
  },
  {
    "start": 442.2,
    "end": 448.86,
    "text": "because one thing the AI doesn't understand is it doesn't understand the business rules."
  },
  {
    "start": 449.64,
    "end": 454.96,
    "text": "That's one thing that AI will never understand because we humans think completely different."
  },
  {
    "start": 455.92,
    "end": 463.08,
    "text": "and in that moment, the tester needs to understand his role."
  },
  {
    "start": 465.1,
    "end": 473.78,
    "text": "He cannot be a junior so he need an expert actually evaluate what does this system do or not."
  },
  {
    "start": 474.74,
    "end": 478.38,
    "text": "And if everything is fine, then you can continue."
  },
  {
    "start": 478.5,
    "end": 490.12,
    "text": "If not, you can add more information in because maybe the information that the AI has is outdated or it has some leaks and there are holes."
  },
  {
    "start": 490.66,
    "end": 494.6,
    "text": "so It gives them more information to regenerate it."
  },
  {
    "start": 495.26,
    "end": 502.2,
    "text": "and at the end You have test case which was well written for a manual tester."
  },
  {
    "start": 503.02,
    "end": 504.42,
    "text": "You could say, okay stop there."
  },
  {
    "start": 504.86,
    "end": 508.02,
    "text": "Now the tester takes over it does manual test and gone."
  },
  {
    "start": 508.4,
    "end": 509.32,
    "text": "but what now?"
  },
  {
    "start": 509.66,
    "end": 511.82,
    "text": "the agent the agentic system does?"
  },
  {
    "start": 512.039,
    "end": 517.659,
    "text": "is It hands-over not a whole task to The automation engineer?"
  },
  {
    "start": 518.52,
    "end": 521.28,
    "text": "So you have an agent that is specialized in automations?"
  },
  {
    "start": 522.039,
    "end": 532.48,
    "text": "And That agent will ask shall I automate your test cases or do you want keep it manually and It takes the test case's and starts automating its?"
  },
  {
    "start": 533.4,
    "end": 539.18,
    "text": "And what is interesting, it automates the way how we do as well."
  },
  {
    "start": 540.2,
    "end": 551.18,
    "text": "Because when I automate and try to go into properties figure out their application so i need give also information make sure that its right application that the AI testing?"
  },
  {
    "start": 552.54,
    "end": 553.56,
    "text": "Is this the right?"
  },
  {
    "start": 553.76,
    "end": 560.92,
    "text": "elements clicking again human in a loop needs to judge if they are doing thing or not."
  },
  {
    "start": 561.7,
    "end": 564.02,
    "text": "If not, they can override at any time."
  },
  {
    "start": 565.58,
    "end": 567.06,
    "text": "if They cannot overwrite it."
  },
  {
    "start": 568.06,
    "end": 571.78,
    "text": "Then the AI needs to figure out how to fix that issue."
  },
  {
    "start": 572.82,
    "end": 573.76,
    "text": "What is interesting?"
  },
  {
    "start": 574.1,
    "end": 588.9,
    "text": "Is That once the test case has been written It gets automatically submitted To the release so that you can be automatically executed and You get right away after the execution results."
  },
  {
    "start": 590.06,
    "end": 594.14,
    "text": "now Maybe it worked, maybe it failed."
  },
  {
    "start": 594.92,
    "end": 606.54,
    "text": "If it fails then again another agent comes into play the buck analyze agent and Then one starts now investigating why did that fail?"
  },
  {
    "start": 607.0,
    "end": 609.0,
    "text": "Is is because element changed?"
  },
  {
    "start": 609.58,
    "end": 615.16,
    "text": "Did they move from left-hand side to right hand side or color change properties changes?"
  },
  {
    "start": 615.98,
    "end": 624.78,
    "text": "And there's actually a risk if you trust AI that it fixes by itself, maybe its fixing in the wrong way."
  },
  {
    "start": 625.54,
    "end": 642.04,
    "text": "Because may be developer did a mistake and changed ID numberings And now AI is identifying an element which looks right for the AI but not It's just so crucial."
  },
  {
    "start": 642.64,
    "end": 652.42,
    "text": "human needs to have control at any time To fix or continue running it as is."
  },
  {
    "start": 653.4,
    "end": 657.98,
    "text": "And once the test case has run, we could say okay done!"
  },
  {
    "start": 658.46,
    "end": 660.06,
    "text": "We can release and deploy it."
  },
  {
    "start": 660.64,
    "end": 662.26,
    "text": "but what about performance testing?"
  },
  {
    "start": 663.94,
    "end": 673.88,
    "text": "So if you have an agent that says ok great now with functional testing... ...we've got a great coverage But let's do some performance testing."
  },
  {
    "start": 674.82,
    "end": 679.88,
    "text": "Let's convert those automated test cases into performance tests because they're not the same."
  },
  {
    "start": 680.14,
    "end": 681.24,
    "text": "They're completely different."
  },
  {
    "start": 681.74,
    "end": 691.64,
    "text": "The scope is different, and it starts building the performance test cases... ...and after it runs in the right system."
  },
  {
    "start": 692.5,
    "end": 694.92,
    "text": "there you need to also indicate which system you want to run."
  },
  {
    "start": 696.32,
    "end": 702.04,
    "text": "It gives you also the bottlenecks And alerts where some issues that needs to be fixed."
  },
  {
    "start": 703.0,
    "end": 729.46,
    "text": "So now all of the tasks a human tester would have done an AI is doing, and we tester are actually now in the space where we should be from beginning on but never had time to do it because were focused on fixing test cases reviewing trying explain requirements engineers into business what they need to."
  },
  {
    "start": 730.66,
    "end": 734.04,
    "text": "Now can focus on our way of testing right thing?"
  },
  {
    "start": 735.6,
    "end": 737.78,
    "text": "Do you trust that the AI?"
  },
  {
    "start": 739.34,
    "end": 743.02,
    "text": "Are we just blindly saying, hey it's quick."
  },
  {
    "start": 743.28,
    "end": 745.8,
    "text": "It looks nice so let's go for it?"
  },
  {
    "start": 745.96,
    "end": 753.08,
    "text": "Yeah yeah...it is a very interesting system to build up such agents in an agentic system."
  },
  {
    "start": 753.5,
    "end": 757.8,
    "text": "So you mentioned some slices of this agent."
  },
  {
    "start": 760.46,
    "end": 762.26,
    "text": "What's the number of agents you have in there?"
  },
  {
    "start": 762.48,
    "end": 766.46,
    "text": "One for their reviewer, one from writing test cases checking coverage."
  },
  {
    "start": 766.72,
    "end": 769.38,
    "text": "How much are they totaled to your testing process?"
  },
  {
    "start": 770.24,
    "end": 770.7,
    "text": "around about?"
  },
  {
    "start": 771.74,
    "end": 773.54,
    "text": "Between eight and twelve."
  },
  {
    "start": 773.94,
    "end": 779.72,
    "text": "Depending on how many features or functionalities we want to hand over... ...to the AI."
  },
  {
    "start": 779.84,
    "end": 788.24,
    "text": "if you're doing it with a banking system then don't give everything to the AI because all regulatory is contained."
  },
  {
    "start": 790.86,
    "end": 793.36,
    "text": "If you do it in a retail, then its something completely different."
  },
  {
    "start": 793.4,
    "end": 799.0,
    "text": "Then we can give even more because they don't need to bother on GDPR and data protection."
  },
  {
    "start": 799.36,
    "end": 799.74,
    "text": "Yeah yeah."
  },
  {
    "start": 800.6,
    "end": 808.06,
    "text": "And when your build up this system You mentioned that some of these agents are running autonomously."
  },
  {
    "start": 808.88,
    "end": 814.04,
    "text": "when they see there is a new requirement to start, how autonomous do you use it in practice?"
  },
  {
    "start": 814.14,
    "end": 820.1,
    "text": "Are all watching everywhere and then starting or push the button now."
  },
  {
    "start": 823.6,
    "end": 842.64,
    "text": "So that means once the requirements are ready in Jira Which requirements or which release to focus on?"
  },
  {
    "start": 843.36,
    "end": 845.84,
    "text": "otherwise it will pull all different types of?"
  },
  {
    "start": 847.16,
    "end": 854.1,
    "text": "Requirements that you actually don't want to focus because the The content is the wrong one."
  },
  {
    "start": 854.22,
    "end": 859.54,
    "text": "Because maybe it's next release, so the release after this being prepared mm-hmm"
  },
  {
    "start": 859.74,
    "end": 860.48,
    "text": "okay I understand."
  },
  {
    "start": 860.94,
    "end": 864.08,
    "text": "and When we?"
  },
  {
    "start": 864.2,
    "end": 870.98,
    "text": "when we go into the agents the different agents What are the main Parts you give them?"
  },
  {
    "start": 871.06,
    "end": 871.84,
    "text": "so we have a lot of."
  },
  {
    "start": 872.02,
    "end": 875.48,
    "text": "we can gave them skills and tools And all the stuff."
  },
  {
    "start": 875.74,
    "end": 877.06,
    "text": "how what is the package?"
  },
  {
    "start": 877.12,
    "end": 879.24,
    "text": "Of such an agent in your in-your system."
  },
  {
    "start": 879.36,
    "end": 881.14,
    "text": "So content is key."
  },
  {
    "start": 881.88,
    "end": 883.46,
    "text": "You need to give the right content."
  },
  {
    "start": 884.28,
    "end": 901.2,
    "text": "The question Is do you want To give the content directly to the agent or Do you Want to Have A container where the Content is In There, and it pulls It and has a version control to check when was the last time that it checked their information."
  },
  {
    "start": 901.98,
    "end": 904.3,
    "text": "And how is actually working right now?"
  },
  {
    "start": 904.54,
    "end": 905.42,
    "text": "So, thats one part."
  },
  {
    "start": 905.5,
    "end": 908.16,
    "text": "Thats business context from company."
  },
  {
    "start": 909.26,
    "end": 914.02,
    "text": "then you have skills not skills by giving system prompt."
  },
  {
    "start": 914.76,
    "end": 916.44,
    "text": "The system prompt is crucial."
  },
  {
    "start": 917.22,
    "end": 920.36,
    "text": "You need know how set up the system prompt."
  },
  {
    "start": 922.28,
    "end": 924.92,
    "text": "Trust me, it's not the first system prompt that you build."
  },
  {
    "start": 925.46,
    "end": 927.02,
    "text": "That will work out well."
  },
  {
    "start": 928.3,
    "end": 947.74,
    "text": "It takes some iterations to get this system prompt in a way so as it should And thats actually crucial part and also decide on how much freedom and access do we want give which areas?"
  },
  {
    "start": 950.54,
    "end": 958.0,
    "text": "When I look at the bubble of AI now, a lot things often get over-engineered."
  },
  {
    "start": 958.44,
    "end": 966.66,
    "text": "You mentioned something like before performance testing and when you talked about it that performances start where everything is ready to agent takesover."
  },
  {
    "start": 967.44,
    "end": 969.38,
    "text": "do we need really an agent for this?"
  },
  {
    "start": 969.5,
    "end": 980.98,
    "text": "Because if i have my CICD pipeline already then automatically can start my performance test task of the agent in this part?"
  },
  {
    "start": 981.4,
    "end": 985.64,
    "text": "So, that task is not only running it."
  },
  {
    "start": 985.72,
    "end": 998.38,
    "text": "Because yes then agreed you can have it in the CICD pipeline and its done but what we need to do Is You build functional test cases That are running autonomous."
  },
  {
    "start": 1000.26,
    "end": 1007.4,
    "text": "But those automated test cases... ...you cannot use for performance testing because the scope is completely different."
  },
  {
    "start": 1008.06,
    "end": 1023.08,
    "text": "On the automated test cases, you are controlling a system under test GUI and your triggering like human being is working on the application or you're triggering something during ABI."
  },
  {
    "start": 1025.16,
    "end": 1036.4,
    "text": "By performance testing... You actually forget that this whole part of happening in front and you look at the communications going on back-and-forward."
  },
  {
    "start": 1036.68,
    "end": 1049.5,
    "text": "You're looking if, let's say not only one person is in the application but now we have two hundred people that are using their applications."
  },
  {
    "start": 1049.84,
    "end": 1065.84,
    "text": "That scaling needs to be done by changing test cases And either give it an AI which can convert them Or you have a performance expert that does it."
  },
  {
    "start": 1066.4,
    "end": 1067.46,
    "text": "The difference is the time."
  },
  {
    "start": 1068.1,
    "end": 1078.7,
    "text": "Okay, so that's... ...the task for the performance agent to transform test cases based on changes and not just running them?"
  },
  {
    "start": 1079.06,
    "end": 1079.52,
    "text": "Exactly!"
  },
  {
    "start": 1080.12,
    "end": 1083.06,
    "text": "burn a lot of tokens for nothing to just run it there."
  },
  {
    "start": 1084.1,
    "end": 1089.28,
    "text": "When we talk about token and usage agents, when they're dealing all the day?"
  },
  {
    "start": 1091.02,
    "end": 1094.34,
    "text": "It's always matter-of-costs in all those contexts."
  },
  {
    "start": 1094.56,
    "end": 1102.98,
    "text": "so how do you deal with that between the agents giving their information and dealing with money behind them?"
  },
  {
    "start": 1103.48,
    "end": 1108.2,
    "text": "So... There is one issue in Switzerland The data protection."
  },
  {
    "start": 1108.24,
    "end": 1108.82,
    "text": "Oh,"
  },
  {
    "start": 1108.96,
    "end": 1112.08,
    "text": "that's not an issue."
  },
  {
    "start": 1112.12,
    "end": 1122.74,
    "text": "No it is not but makes some things difficult because we have companies and domains."
  },
  {
    "start": 1124.22,
    "end": 1131.84,
    "text": "you are not allowed to use applications on US servers or even outside of Switzerland."
  },
  {
    "start": 1133.0,
    "end": 1136.64,
    "text": "How many AI systems do we have in Switzerland running on cloud?"
  },
  {
    "start": 1139.18,
    "end": 1139.64,
    "text": "I know one."
  },
  {
    "start": 1141.18,
    "end": 1152.74,
    "text": "So it's quite a challenge to first understand the company, what are their needs and what kind of rules do they need to follow?"
  },
  {
    "start": 1153.82,
    "end": 1168.76,
    "text": "And with that you can then start forgetting chatGPT, Cloud, Gemini... You can forget all those tools because the data is not under your control."
  },
  {
    "start": 1171.24,
    "end": 1174.2,
    "text": "Even though you think it's under your controls, its NOT!"
  },
  {
    "start": 1175.92,
    "end": 1189.16,
    "text": "You need to have the agents and AI on a server in Switzerland running so that I can have full control over what kind of information they are using."
  },
  {
    "start": 1189.98,
    "end": 1198.52,
    "text": "So we use Mistral or Ulama as the AI in the back, or even DeepSeq."
  },
  {
    "start": 1199.98,
    "end": 1211.08,
    "text": "Because you can switch however you like to and if it comes to data protection You need to run locally."
  },
  {
    "start": 1211.42,
    "end": 1234.54,
    "text": "There is no other chance right now than running locally And invest a lot into a server Or have a server farm somewhere that is specialized on AI systems, so it has enough GPUs and capable of running those systems because you cannot use an irregular server."
  },
  {
    "start": 1234.6,
    "end": 1240.54,
    "text": "You need to have the graphic performance to do those AI system."
  },
  {
    "start": 1242.4,
    "end": 1248.0,
    "text": "from older gamers all over Switzerland, the machines to build their own data center."
  },
  {
    "start": 1250.28,
    "end": 1264.56,
    "text": "Okay yeah well one very specific question what comes in my mind when I'm using for my business an agentic system too a different way for my back office and i gave my agents different names uh...to address them."
  },
  {
    "start": 1264.7,
    "end": 1269.8,
    "text": "so I have my personal assistant Anna and she's dealing with other stuff around it."
  },
  {
    "start": 1271.66,
    "end": 1273.28,
    "text": "Do you have names for this agent?"
  },
  {
    "start": 1273.4,
    "end": 1274.1,
    "text": "I have javas."
  },
  {
    "start": 1274.16,
    "end": 1274.96,
    "text": "Maybe"
  },
  {
    "start": 1275.0,
    "end": 1277.38,
    "text": "in a while, yeah okay that's obvious!"
  },
  {
    "start": 1277.48,
    "end": 1279.02,
    "text": "Yeah so...I"
  },
  {
    "start": 1279.34,
    "end": 1283.58,
    "text": "actually call it Javas and uh i like it."
  },
  {
    "start": 1284.08,
    "end": 1289.22,
    "text": "um i used to give all my agents' names mm-hmm And then i got confused."
  },
  {
    "start": 1289.6,
    "end": 1290.74,
    "text": "what names did i give?"
  },
  {
    "start": 1291.04,
    "end": 1291.94,
    "text": "which agent?"
  },
  {
    "start": 1292.08,
    "end": 1294.04,
    "text": "So..i stopped at."
  },
  {
    "start": 1294.34,
    "end": 1299.4,
    "text": "Ah Okay Yes You addressed them by task."
  },
  {
    "start": 1301.84,
    "end": 1318.98,
    "text": "yes When we look on your whole agentic system and someone who's watching this episode now, thinks okay I want to start this one too maybe with my own AI systems or with US systems."
  },
  {
    "start": 1321.04,
    "end": 1323.66,
    "text": "Which of these agents do you think is the best to start?"
  },
  {
    "start": 1325.16,
    "end": 1331.68,
    "text": "It depends if you want to understand technical backgrounds with it."
  },
  {
    "start": 1332.4,
    "end": 1334.12,
    "text": "There are some tools that you can use."
  },
  {
    "start": 1336.32,
    "end": 1342.54,
    "text": "I started to wait make, make dot com To just understand how those interactions work."
  },
  {
    "start": 1343.74,
    "end": 1348.04,
    "text": "Then i started using an eight n to run it also locally."
  },
  {
    "start": 1349.22,
    "end": 1366.24,
    "text": "then i started looking into cloud code because everybody is talking about cloud code now and Also chat gpt And you need to figure out what kind of system does your needs cover."
  },
  {
    "start": 1366.38,
    "end": 1369.82,
    "text": "Now, for a company perspective I don't use any of them."
  },
  {
    "start": 1371.38,
    "end": 1382.72,
    "text": "You need to have a system that is capable of doing it and right now we used from Tricentis and UiPath The systems because those two are very powerful."
  },
  {
    "start": 1383.42,
    "end": 1390.9,
    "text": "They did great job to build a system that meets the needs for testers especially."
  },
  {
    "start": 1392.1,
    "end": 1397.7,
    "text": "And pointing on the task, which one would you suggest start with?"
  },
  {
    "start": 1398.08,
    "end": 1405.14,
    "text": "The requirement review or the Test Case Generator... Which is good way of getting into these topics?"
  },
  {
    "start": 1405.44,
    "end": 1428.5,
    "text": "The requirements' reviews are actually great because then we can see if they're understanding the right way like AI is doing And if you are having your thoughts and impressions giving into the expectation of the requirement, which AI is not doing to figure out does it match or not."
  },
  {
    "start": 1429.06,
    "end": 1448.1,
    "text": "Then a test case creation is also very good because that's simple task where we can just make sure that functional test cases have been created and to understand if the knowledge that you learned, let's say from an ISTQB foundation level is still there."
  },
  {
    "start": 1448.3,
    "end": 1450.36,
    "text": "And do have a same understanding?"
  },
  {
    "start": 1450.84,
    "end": 1456.08,
    "text": "Or are AI starting hallucinating or doing some stuff it doesn't meet?"
  },
  {
    "start": 1457.14,
    "end": 1462.22,
    "text": "If they don't know what should actually be done then where does this issue come in?"
  },
  {
    "start": 1462.86,
    "end": 1471.36,
    "text": "Is because we forgot to give them content... the major issue that a lot of people think, they think I'm using an AI."
  },
  {
    "start": 1471.76,
    "end": 1477.52,
    "text": "The AI is all-known source and it can do everything!"
  },
  {
    "start": 1477.98,
    "end": 1481.24,
    "text": "And actually its not... AI is completely stupid."
  },
  {
    "start": 1482.4,
    "end": 1484.7,
    "text": "It has no clue about anything."
  },
  {
    "start": 1485.54,
    "end": 1500.38,
    "text": "You're just trying to direct information into areas where maybe someone trained or gave them that it could use, but if you're not giving your content then there the issue starts."
  },
  {
    "start": 1500.66,
    "end": 1510.02,
    "text": "Yeah yeah when you look at the agentic system now today May the sixth I think is a day."
  },
  {
    "start": 1511.1,
    "end": 1512.1,
    "text": "what are they boundaries?"
  },
  {
    "start": 1512.6,
    "end": 1513.46,
    "text": "What where do say?"
  },
  {
    "start": 1513.52,
    "end": 1521.72,
    "text": "okay therefore we cannot use the agentics system right now and We need to wait till next generation or maybe It's not feasible in their future."
  },
  {
    "start": 1521.94,
    "end": 1527.88,
    "text": "so It has a lot of capabilities to do recurring tasks."
  },
  {
    "start": 1528.46,
    "end": 1529.6,
    "text": "That's one thing that it can't Do."
  },
  {
    "start": 1530.62,
    "end": 1530.72,
    "text": "what?"
  },
  {
    "start": 1530.76,
    "end": 1531.44,
    "text": "It cannot do."
  },
  {
    "start": 1531.6,
    "end": 1537.78,
    "text": "it cannot think like a human, it will never Think Like A Human even though we think that they Will."
  },
  {
    "start": 1538.12,
    "end": 1538.9,
    "text": "They Were Startling."
  },
  {
    "start": 1538.98,
    "end": 1540.92,
    "text": "so there is One Big Boundary."
  },
  {
    "start": 1542.5,
    "end": 1548.3,
    "text": "It Will Never Have The Possibility To Check the Coverage in A way that We Humans Do."
  },
  {
    "start": 1549.58,
    "end": 1556.78,
    "text": "our way of judging something is completely different than an AI, because the AI has just zeros and ones."
  },
  {
    "start": 1557.42,
    "end": 1557.88,
    "text": "Nothing more!"
  },
  {
    "start": 1558.94,
    "end": 1561.48,
    "text": "And that's why it's not able to think."
  },
  {
    "start": 1562.26,
    "end": 1565.88,
    "text": "even though you THINK they can't think."
  },
  {
    "start": 1566.32,
    "end": 1575.18,
    "text": "They just grab information and think, oh Richie likes this information so I will give him this last time he liked one So i'll go again."
  },
  {
    "start": 1576.68,
    "end": 1581.68,
    "text": "it's not thinking, no artificial intelligence really."
  },
  {
    "start": 1582.14,
    "end": 1587.28,
    "text": "There is no intelligence in there It just gathering information and handing over to you."
  },
  {
    "start": 1588.56,
    "end": 1594.4,
    "text": "This will definitely change the future but I cannot tell when And how?"
  },
  {
    "start": 1594.86,
    "end": 1594.98,
    "text": "Yeah"
  },
  {
    "start": 1596.9,
    "end": 1597.0,
    "text": "yeah."
  },
  {
    "start": 1597.12,
    "end": 1607.08,
    "text": "so i think that The thinking dot-dot-dot is always misleading because its'nt thinking doing something with information, but not thinking at all."
  },
  {
    "start": 1608.36,
    "end": 1613.26,
    "text": "Nishan thank you very much that you were here in the show and shared your insights and your experience with it."
  },
  {
    "start": 1613.38,
    "end": 1620.44,
    "text": "I think It's very valuable And to think about for everyone For its own test process where we can put in agents?"
  },
  {
    "start": 1620.66,
    "end": 1626.12,
    "text": "How We Can build up an agentic system which supports us as a tester and As a developer team."
  },
  {
    "start": 1626.78,
    "end": 1628.6,
    "text": "so Thank You very Much That you shared it."
  },
  {
    "start": 1628.66,
    "end": 1631.5,
    "text": "i wish you A great conference and a Great Talk Later on."
  },
  {
    "start": 1631.78,
    "end": 1632.46,
    "text": "yeah Hopefully"
  },
  {
    "start": 1633.1,
    "end": 1633.58,
    "text": "See you soon."
  },
  {
    "start": 1633.96,
    "end": 1634.76,
    "text": "Thank you very much for"
  },
  {
    "start": 1635.74,
    "end": 1635.8,
    "text": "the"
  },
  {
    "start": 1636.72,
    "end": 1637.36,
    "text": "time."
  }
]