[
  {
    "start": 0.08,
    "end": 3.74,
    "text": "At this year's World Economic Forum in\nDavos, Dario"
  },
  {
    "start": 3.78,
    "end": 7.64,
    "text": "Amodei predicted that within the next 12"
  },
  {
    "start": 7.72,
    "end": 11.38,
    "text": "months or so, AI would be able to write"
  },
  {
    "start": 11.6,
    "end": 15.4,
    "text": "all code, fully automated, on its own,\nessentially."
  },
  {
    "start": 15.41,
    "end": 19.28,
    "text": "And it's worth paying attention to what\nthis man has to say and I'll"
  },
  {
    "start": 19.38,
    "end": 22.92,
    "text": "also share my thoughts on it\nand why I think that you should take a"
  },
  {
    "start": 23.06,
    "end": 27.02,
    "text": "nuanced look at that.\nNot just because Dario, of course,"
  },
  {
    "start": 27.06,
    "end": 30.67,
    "text": "is the CEO of Anthropic which\nis one of the most important players in"
  },
  {
    "start": 30.74,
    "end": 34.54,
    "text": "generative AI space, especially\nwhen it comes to models related to"
  },
  {
    "start": 34.56,
    "end": 38.14,
    "text": "coding.\nBut it's also worth paying attention"
  },
  {
    "start": 38.22,
    "end": 42.14,
    "text": "year, in the beginning of 2025,\nhe predicted"
  },
  {
    "start": 42.15,
    "end": 45.88,
    "text": "that AI would be able to write 90% of all\ncode"
  },
  {
    "start": 45.92,
    "end": 49.8,
    "text": "within the next three or six months\nand depending on how"
  },
  {
    "start": 49.88,
    "end": 52.7,
    "text": "you look at that, he\nwas not entirely wrong."
  },
  {
    "start": 52.72,
    "end": 56.4,
    "text": "Now, sure, this timeline is probably,\nor was"
  },
  {
    "start": 56.5,
    "end": 60.42,
    "text": "probably a bit too aggressive.\nIt maybe wasn't three or"
  },
  {
    "start": 60.58,
    "end": 64.54,
    "text": "six months, more like six to nine months,\nnot sure."
  },
  {
    "start": 64.599,
    "end": 68.5,
    "text": "And, of course,\nit totally depends on your field, so in"
  },
  {
    "start": 68.54,
    "end": 72.13,
    "text": "which area you're programming,\nalso in the programming language you're"
  },
  {
    "start": 72.16,
    "end": 76.02,
    "text": "It will depend on your company,\nthe policies you have in your company."
  },
  {
    "start": 76.04,
    "end": 79.26,
    "text": "And of course, it also depends on, on you,\nyour"
  },
  {
    "start": 79.38,
    "end": 82.5,
    "text": "preferences, your experiences with AI."
  },
  {
    "start": 82.64,
    "end": 86.2,
    "text": "I can say that for me,\nI am at a point I would say since"
  },
  {
    "start": 86.48,
    "end": 90.44,
    "text": "maybe December, November, where for some"
  },
  {
    "start": 90.48,
    "end": 93.66,
    "text": "projects, most of my projects,\nAI does write"
  },
  {
    "start": 94.26,
    "end": 98.04,
    "text": "80, 90% of all code for me probably."
  },
  {
    "start": 98.06,
    "end": 102.04,
    "text": "But this is now where things get\ninteresting."
  },
  {
    "start": 102.2,
    "end": 105.93,
    "text": "I'm not talking about wipe coding\nand I'm not"
  },
  {
    "start": 105.96,
    "end": 109.61,
    "text": "talking about AI doing it all on its own\nand I'm also not"
  },
  {
    "start": 109.64,
    "end": 113.24,
    "text": "talking about AI getting it all right."
  },
  {
    "start": 113.3,
    "end": 116.4,
    "text": "Instead,\nthings are way more difficult than that."
  },
  {
    "start": 116.42,
    "end": 120.2,
    "text": "So it does absolutely not do that"
  },
  {
    "start": 120.32,
    "end": 123.85,
    "text": "on its own. It also is"
  },
  {
    "start": 124.0,
    "end": 127.89,
    "text": "not about wipe coding here. I mean,"
  },
  {
    "start": 127.92,
    "end": 131.84,
    "text": "you can probably wipe code\nbut I've shared my view on that before"
  },
  {
    "start": 131.88,
    "end": 135.5,
    "text": "me, by the way,\njust to make sure we all got the same"
  },
  {
    "start": 135.6,
    "end": 138.64,
    "text": "really means that you don't look at the\ncode at all, that you don't care about the"
  },
  {
    "start": 138.7,
    "end": 141.66,
    "text": "code,\nthat you don't necessarily even know how"
  },
  {
    "start": 141.7,
    "end": 145.62,
    "text": "So yeah, this is also not about wipe"
  },
  {
    "start": 145.72,
    "end": 149.63,
    "text": "coding here. Instead, I would say we are"
  },
  {
    "start": 149.68,
    "end": 153.58,
    "text": "at a point where you can hand off many\ntasks"
  },
  {
    "start": 153.66,
    "end": 157.56,
    "text": "to AI with a good plan. So"
  },
  {
    "start": 157.62,
    "end": 160.46,
    "text": "if you have a good plan, AI can"
  },
  {
    "start": 160.68,
    "end": 164.6,
    "text": "implement it. So I think we're at that"
  },
  {
    "start": 164.72,
    "end": 168.68,
    "text": "point right now. You,\nyou can definitely do that, at least"
  },
  {
    "start": 168.69,
    "end": 172.66,
    "text": "for some tasks. But of course, you wanna"
  },
  {
    "start": 172.72,
    "end": 176.58,
    "text": "review that code and, uh, I know there are"
  },
  {
    "start": 176.6,
    "end": 180.32,
    "text": "people that tell you you don't need to do\nthat anymore. Might work for them."
  },
  {
    "start": 180.34,
    "end": 184.22,
    "text": "Definitely not working for me and\nif you're shipping anything, if you're"
  },
  {
    "start": 184.36,
    "end": 187.82,
    "text": "shipping anything to customers,\nyou're responsible for that."
  },
  {
    "start": 187.96,
    "end": 191.92,
    "text": "You, as a developer, you can't say, \"Oh,\nthe AI got that wrong.\" No, it's"
  },
  {
    "start": 192.04,
    "end": 195.92,
    "text": "your responsibility\nand I won't take responsibility in code"
  },
  {
    "start": 196.58,
    "end": 199.98,
    "text": "I haven't reviewed, I don't understand."
  },
  {
    "start": 200.08,
    "end": 203.32,
    "text": "Also, still, AI makes many"
  },
  {
    "start": 203.38,
    "end": 207.32,
    "text": "mistakes, so I need to fix mistakes"
  },
  {
    "start": 207.78,
    "end": 210.64,
    "text": "or steer AI in the right"
  },
  {
    "start": 211.5,
    "end": 212.34,
    "text": "direction."
  },
  {
    "start": 214.62,
    "end": 218.2,
    "text": "And that is also very important. That's,\nof course,"
  },
  {
    "start": 218.24,
    "end": 222.0,
    "text": "important to the planning part\nor related to the"
  },
  {
    "start": 222.04,
    "end": 224.58,
    "text": "planning part,\nthat you steer AI in the right direction."
  },
  {
    "start": 224.6,
    "end": 228.03,
    "text": "But it's also related to what you do with\ncode AI gives you."
  },
  {
    "start": 228.06,
    "end": 231.76,
    "text": "So very far away from AI"
  },
  {
    "start": 232.14,
    "end": 236.04,
    "text": "generating 90% of the code and\nthat means I have"
  },
  {
    "start": 236.08,
    "end": 240.04,
    "text": "no work to do. Far away from that.\nIt just means that I"
  },
  {
    "start": 240.08,
    "end": 243.56,
    "text": "got a very fast typer who can"
  },
  {
    "start": 243.62,
    "end": 247.46,
    "text": "implement my plans,\nbut those plans need to be good"
  },
  {
    "start": 247.96,
    "end": 251.24,
    "text": "and the output is kind of varying in\nquality."
  },
  {
    "start": 251.32,
    "end": 255.24,
    "text": "Still, for me,\nit's very likely faster than"
  },
  {
    "start": 255.28,
    "end": 258.92,
    "text": "it all from scratch.\nIf I instead try to build good plans,"
  },
  {
    "start": 259.0,
    "end": 262.98,
    "text": "split that up into smaller chunks,\nlet AI write the code for me and"
  },
  {
    "start": 263.02,
    "end": 264.8,
    "text": "then review and fine-tune that code."
  },
  {
    "start": 264.84,
    "end": 268.43,
    "text": "And by the way, with planning,\nI mean really detailed"
  },
  {
    "start": 268.43,
    "end": 272.36,
    "text": "plans where I also break down the exact\nlibraries I wanna use, the"
  },
  {
    "start": 272.42,
    "end": 275.68,
    "text": "patterns I wanna use,\nthe architecture of the software I wanna"
  },
  {
    "start": 275.76,
    "end": 279.74,
    "text": "step-by-step. So it's not a,\na rough plan or a general plan; it's"
  },
  {
    "start": 279.78,
    "end": 283.5,
    "text": "a very detailed plan.\nThat works for me and therefore, for"
  },
  {
    "start": 283.56,
    "end": 287.38,
    "text": "me, with, uh,\nwith these restrictions here in mind, I"
  },
  {
    "start": 287.42,
    "end": 290.4,
    "text": "would say, yeah, it can probably write"
  },
  {
    "start": 290.66,
    "end": 294.469,
    "text": "90%-ish of the code. But that does not"
  },
  {
    "start": 294.5,
    "end": 296.76,
    "text": "mean that it does 90% of the work on its\nown."
  },
  {
    "start": 296.78,
    "end": 300.55,
    "text": "Now,\nthat's important context because that's"
  },
  {
    "start": 300.6,
    "end": 303.89,
    "text": "to last year's statement by Dario."
  },
  {
    "start": 303.92,
    "end": 307.82,
    "text": "So yeah, I would say we are kind of there,\nbut not in the sense of AI doing"
  },
  {
    "start": 307.86,
    "end": 311.38,
    "text": "it all on its own. Now,\nwhat about this year's"
  },
  {
    "start": 311.48,
    "end": 315.42,
    "text": "statement? Though, in this year,\nDario essentially said that AI will be"
  },
  {
    "start": 315.48,
    "end": 319.46,
    "text": "able to do what software engineers do\nand write software fully on its own,"
  },
  {
    "start": 319.48,
    "end": 322.36,
    "text": "end-to-end, within six to 12 months."
  },
  {
    "start": 322.4,
    "end": 325.92,
    "text": "That's this year's statement. Now,\nI fully recommend"
  },
  {
    "start": 325.96,
    "end": 329.92,
    "text": "watching this entire talk.\nIt's very interesting, but I have some"
  },
  {
    "start": 329.94,
    "end": 333.85,
    "text": "thoughts about this statement.\nAnd obviously, just to make"
  },
  {
    "start": 333.86,
    "end": 337.46,
    "text": "and obvious, I'm in no way smarter\nor more capable"
  },
  {
    "start": 337.94,
    "end": 341.28,
    "text": "in judging the performance of AI models\nthan Dario."
  },
  {
    "start": 341.32,
    "end": 345.02,
    "text": "But I'm also not the CEO of a company\nthat needs to"
  },
  {
    "start": 345.12,
    "end": 348.71,
    "text": "sell these AI models\nand I can share from my own"
  },
  {
    "start": 348.74,
    "end": 352.31,
    "text": "experiences. Now, as I said,\nI would agree with last year's"
  },
  {
    "start": 352.36,
    "end": 356.02,
    "text": "statement to some degree,\nbut with many caveats, with many"
  },
  {
    "start": 356.3,
    "end": 360.24,
    "text": "restrictions. The AI\nis definitely not writing 90% of my code"
  },
  {
    "start": 360.28,
    "end": 364.04,
    "text": "its own. So naturally,\nI have a very hard time"
  },
  {
    "start": 364.1,
    "end": 367.976,
    "text": "imagining... that this will come true,\nand I have a very hard time"
  },
  {
    "start": 368.016,
    "end": 371.916,
    "text": "imagining that this will come true,\nnot just within the next 6 or 12"
  },
  {
    "start": 371.996,
    "end": 375.616,
    "text": "months, but anywhere in the near future."
  },
  {
    "start": 375.626,
    "end": 379.016,
    "text": "Now, I totally see that AI is"
  },
  {
    "start": 379.036,
    "end": 382.736,
    "text": "capable of building software on its"
  },
  {
    "start": 382.856,
    "end": 386.046,
    "text": "own in a loop,\nwith the Rolf loop that's getting a lot of"
  },
  {
    "start": 386.076,
    "end": 389.956,
    "text": "hype, uh, related to Claude code.\nI totally see"
  },
  {
    "start": 390.016,
    "end": 393.916,
    "text": "that. But the full work of a software"
  },
  {
    "start": 393.956,
    "end": 396.916,
    "text": "engineer includes the tasks I outlined\nhere"
  },
  {
    "start": 397.576,
    "end": 401.396,
    "text": "and arguably more than that.\nIt includes building a"
  },
  {
    "start": 401.456,
    "end": 404.776,
    "text": "good plan, defining which architecture,\nwhich"
  },
  {
    "start": 404.916,
    "end": 408.396,
    "text": "patterns, which technologies to use,\nreviewing the"
  },
  {
    "start": 408.456,
    "end": 412.256,
    "text": "code, and also, of course, analyzing the"
  },
  {
    "start": 412.296,
    "end": 416.216,
    "text": "code, fixing problems with the code,\ntaking responsibility for that"
  },
  {
    "start": 416.296,
    "end": 420.036,
    "text": "code.\nThat is something I have a very hard time"
  },
  {
    "start": 420.636,
    "end": 424.436,
    "text": "within the near future because right now,\nthe AI I can"
  },
  {
    "start": 424.556,
    "end": 427.776,
    "text": "use is too far away from that. It's a"
  },
  {
    "start": 427.876,
    "end": 431.836,
    "text": "talented, fast writer\nthat makes a lot of mistakes along"
  },
  {
    "start": 431.856,
    "end": 434.816,
    "text": "the way and that needs very clear\nguidance."
  },
  {
    "start": 434.836,
    "end": 438.636,
    "text": "And that going to models\nthat can do it all on its own, that can"
  },
  {
    "start": 438.676,
    "end": 442.336,
    "text": "plan out entire architectures cleanly on\ntheir"
  },
  {
    "start": 442.416,
    "end": 445.666,
    "text": "own, that are capable of using the latest"
  },
  {
    "start": 445.856,
    "end": 449.756,
    "text": "technologies,\nof writing error-free secure code totally"
  },
  {
    "start": 449.816,
    "end": 453.336,
    "text": "on their own without review\nor only with review by other"
  },
  {
    "start": 453.396,
    "end": 457.176,
    "text": "AIs,\nthat is something I have a very hard time"
  },
  {
    "start": 457.516,
    "end": 461.476,
    "text": "seeing when I look at today's models\nand also when I look at the"
  },
  {
    "start": 461.536,
    "end": 464.106,
    "text": "progress of models over the last years."
  },
  {
    "start": 464.156,
    "end": 467.676,
    "text": "Because, yeah, that progress, of course,\nhas been remarkable."
  },
  {
    "start": 467.716,
    "end": 471.556,
    "text": "It has been steady and, and good,\nand especially the tools I shared that"
  },
  {
    "start": 471.596,
    "end": 475.576,
    "text": "before got way better.\nSo I shared it before"
  },
  {
    "start": 475.656,
    "end": 479.356,
    "text": "with models. I'm not sure\nif we're the raw model intelligence,"
  },
  {
    "start": 479.416,
    "end": 482.806,
    "text": "we're still on a linear,\nleave alone exponential"
  },
  {
    "start": 482.856,
    "end": 486.836,
    "text": "trajectory. For the tools,\nI would say we definitely are on some"
  },
  {
    "start": 486.896,
    "end": 490.876,
    "text": "linear trajectory. But I have a hard"
  },
  {
    "start": 490.956,
    "end": 494.745,
    "text": "time seeing that being enough to get us to\nthat full"
  },
  {
    "start": 494.796,
    "end": 498.266,
    "text": "automation anytime soon. Now,"
  },
  {
    "start": 498.316,
    "end": 502.076,
    "text": "naturally, these CEOs have other interests"
  },
  {
    "start": 502.136,
    "end": 506.025,
    "text": "than,\nthan me and you could say I have the"
  },
  {
    "start": 506.076,
    "end": 509.636,
    "text": "us developers also because I also sell\nprogramming"
  },
  {
    "start": 509.676,
    "end": 513.656,
    "text": "courses. But let's be totally honest,\nthat is not my role"
  },
  {
    "start": 513.696,
    "end": 517.666,
    "text": "and nothing I will be able to do.\nI'm just sharing here what"
  },
  {
    "start": 517.716,
    "end": 521.525,
    "text": "my experience with AI has been,\nand I'm very open-minded, I would"
  },
  {
    "start": 521.536,
    "end": 523.805,
    "text": "say, for AI. I'm using it a lot,\nas I said."
  },
  {
    "start": 523.836,
    "end": 527.546,
    "text": "It does write 90% of my code.\nIt's just far away from"
  },
  {
    "start": 527.596,
    "end": 531.496,
    "text": "that full automation. But yeah,\nplease let me know your"
  },
  {
    "start": 531.516,
    "end": 535.376,
    "text": "thoughts on this as well,\nyour experiences with AI,"
  },
  {
    "start": 535.436,
    "end": 538.616,
    "text": "working because you might not be using AI\nat all or only for"
  },
  {
    "start": 539.176,
    "end": 543.096,
    "text": "small isolated tasks\nand you might still write most of the code"
  },
  {
    "start": 543.316,
    "end": 547.146,
    "text": "or you're using it for everything,\nyou're not even looking on the c- uh,"
  },
  {
    "start": 547.196,
    "end": 549.536,
    "text": "code, and you're having great results."
  },
  {
    "start": 549.656,
    "end": 552.805,
    "text": "I'm interested to learn more about that,\nso please let me know."
  },
  {
    "start": 552.856,
    "end": 555.055,
    "text": "And yeah, have a great time"
  }
]