[
  {
    "startMs": 1309,
    "text": "[music]"
  },
  {
    "startMs": 12560,
    "text": ">> How is the conference for all of you so"
  },
  {
    "startMs": 14280,
    "text": "far?"
  },
  {
    "startMs": 16080,
    "text": "Great. Awesome."
  },
  {
    "startMs": 18280,
    "text": "Um well, I think this was the best most"
  },
  {
    "startMs": 22400,
    "text": "productive year for so many of us."
  },
  {
    "startMs": 25560,
    "text": "I'm Lena. A few days ago, I solved a"
  },
  {
    "startMs": 28640,
    "text": "production incident on a trail near a"
  },
  {
    "startMs": 31480,
    "text": "waterfall."
  },
  {
    "startMs": 32960,
    "text": "My friend ran"
  },
  {
    "startMs": 34680,
    "text": "18 agents while riding his bike."
  },
  {
    "startMs": 37960,
    "text": "We're literally drowning in abundance."
  },
  {
    "startMs": 40800,
    "text": "We have more output, more speed, more"
  },
  {
    "startMs": 44240,
    "text": "leverage than any of us have ever had."
  },
  {
    "startMs": 47360,
    "text": "So, why do we have this feeling like the"
  },
  {
    "startMs": 50560,
    "text": "ground underneath is moving too fast?"
  },
  {
    "startMs": 54520,
    "text": "One of the engineers that I met at this"
  },
  {
    "startMs": 56000,
    "text": "conference"
  },
  {
    "startMs": 57520,
    "text": "said yesterday that it feels like"
  },
  {
    "startMs": 60360,
    "text": "the opportunity cost for not working"
  },
  {
    "startMs": 62600,
    "text": "9:00 a.m. to 9:00 p.m. 6 days a week is"
  },
  {
    "startMs": 65360,
    "text": "too high right now."
  },
  {
    "startMs": 67600,
    "text": "So, we're all token maxing. We're all"
  },
  {
    "startMs": 69680,
    "text": "working all the time."
  },
  {
    "startMs": 72400,
    "text": "But the same abundance that made you"
  },
  {
    "startMs": 74400,
    "text": "fast, it also made everyone else fast."
  },
  {
    "startMs": 77920,
    "text": "So, now everyone can build everything."
  },
  {
    "startMs": 81560,
    "text": "Your competitor can build your feature"
  },
  {
    "startMs": 83440,
    "text": "this afternoon, too."
  },
  {
    "startMs": 85520,
    "text": "So, the cost of the average just went to"
  },
  {
    "startMs": 87960,
    "text": "zero and so did its value."
  },
  {
    "startMs": 91040,
    "text": "A year ago, the superpower, as we were"
  },
  {
    "startMs": 93120,
    "text": "told, was to be good at using AI."
  },
  {
    "startMs": 97360,
    "text": "But models got so good"
  },
  {
    "startMs": 99880,
    "text": "and they got so easy and everybody now"
  },
  {
    "startMs": 102920,
    "text": "is a lot more skilled at using AI and"
  },
  {
    "startMs": 105720,
    "text": "everybody's pointing AI at the same"
  },
  {
    "startMs": 108200,
    "text": "goals."
  },
  {
    "startMs": 109520,
    "text": "Cuz AI gives everyone the same answer"
  },
  {
    "startMs": 112400,
    "text": "because everybody is asking the same"
  },
  {
    "startMs": 114160,
    "text": "question."
  },
  {
    "startMs": 115360,
    "text": "It It on data and data is a record of"
  },
  {
    "startMs": 117960,
    "text": "what has already happened."
  },
  {
    "startMs": 120200,
    "text": "So, when you point AI at tasks like,"
  },
  {
    "startMs": 123240,
    "text": "\"Tell me what users want. Make more"
  },
  {
    "startMs": 125040,
    "text": "money. What should we build? Make this"
  },
  {
    "startMs": 127479,
    "text": "viral.\""
  },
  {
    "startMs": 128640,
    "text": "It answers from the common knowledge."
  },
  {
    "startMs": 131440,
    "text": "Very competent, very confident, but also"
  },
  {
    "startMs": 133720,
    "text": "very identical to what it tells your"
  },
  {
    "startMs": 135520,
    "text": "competitor."
  },
  {
    "startMs": 137480,
    "text": "To see something that data doesn't show"
  },
  {
    "startMs": 139480,
    "text": "yet, we need to have a vision, a point"
  },
  {
    "startMs": 141600,
    "text": "of view, a read on where it's going, and"
  },
  {
    "startMs": 144680,
    "text": "then use all that automation to execute"
  },
  {
    "startMs": 147040,
    "text": "it."
  },
  {
    "startMs": 148600,
    "text": "AI is a really smart convergence"
  },
  {
    "startMs": 151120,
    "text": "machine."
  },
  {
    "startMs": 152320,
    "text": "So, if you leave it alone, it makes"
  },
  {
    "startMs": 154480,
    "text": "everything the same."
  },
  {
    "startMs": 157560,
    "text": "There is one decision, though, that AI"
  },
  {
    "startMs": 159200,
    "text": "can't and shouldn't make for you. It is"
  },
  {
    "startMs": 161680,
    "text": "to decide what to point at."
  },
  {
    "startMs": 164920,
    "text": "So, the job, the new job for everyone of"
  },
  {
    "startMs": 167000,
    "text": "us is deciding what it makes, being the"
  },
  {
    "startMs": 170040,
    "text": "reason the right people choose your"
  },
  {
    "startMs": 172240,
    "text": "version over the identical-looking rest."
  },
  {
    "startMs": 175920,
    "text": "But also, I'm sure many of you uh walked"
  },
  {
    "startMs": 178000,
    "text": "around the Expo Hall at this conference,"
  },
  {
    "startMs": 180320,
    "text": "and there are so many amazing products,"
  },
  {
    "startMs": 182600,
    "text": "so many tools and vendors."
  },
  {
    "startMs": 184680,
    "text": "They're all solving important problems."
  },
  {
    "startMs": 187040,
    "text": "But what Why do they all sound the same?"
  },
  {
    "startMs": 191519,
    "text": "So, when anyone can build anything, what"
  },
  {
    "startMs": 194160,
    "text": "makes me different? What makes you"
  },
  {
    "startMs": 196200,
    "text": "different? Why should anyone pick your"
  },
  {
    "startMs": 198440,
    "text": "version, your product? Um I call this"
  },
  {
    "startMs": 201760,
    "text": "work uh signal layer."
  },
  {
    "startMs": 204280,
    "text": "And there are two There's two halves to"
  },
  {
    "startMs": 206600,
    "text": "solving it, to getting this right. So,"
  },
  {
    "startMs": 209360,
    "text": "that's how we will walk through it."
  },
  {
    "startMs": 211600,
    "text": "The first half is knowing your signal,"
  },
  {
    "startMs": 214680,
    "text": "being able to define it very clearly."
  },
  {
    "startMs": 217120,
    "text": "What you're building and why it's yours"
  },
  {
    "startMs": 219360,
    "text": "and not the average. So, that's the"
  },
  {
    "startMs": 221400,
    "text": "build side, the code, the product, the"
  },
  {
    "startMs": 223680,
    "text": "road map."
  },
  {
    "startMs": 225320,
    "text": "And the second half is emitting that"
  },
  {
    "startMs": 227760,
    "text": "signal without distortion. So, making"
  },
  {
    "startMs": 230640,
    "text": "sure that your customers um making sure"
  },
  {
    "startMs": 233960,
    "text": "what your customers come to believe"
  },
  {
    "startMs": 235880,
    "text": "about you actually matches what you"
  },
  {
    "startMs": 238200,
    "text": "believe and what you've built."
  },
  {
    "startMs": 240440,
    "text": "That's the ship side the content and go"
  },
  {
    "startMs": 242720,
    "text": "to market engineering."
  },
  {
    "startMs": 245400,
    "text": "And I've had an unusual vantage point in"
  },
  {
    "startMs": 247560,
    "text": "this. I've built products as an"
  },
  {
    "startMs": 249280,
    "text": "engineer. I created my own as a founder."
  },
  {
    "startMs": 252400,
    "text": "I brought other people's products to"
  },
  {
    "startMs": 254440,
    "text": "market. So three very different jobs"
  },
  {
    "startMs": 257320,
    "text": "with one identical challenge. The signal"
  },
  {
    "startMs": 260840,
    "text": "doesn't always survive. So let's start"
  },
  {
    "startMs": 263680,
    "text": "with the build side."
  },
  {
    "startMs": 265840,
    "text": "So what do we even work on? Everything"
  },
  {
    "startMs": 268200,
    "text": "is implementable."
  },
  {
    "startMs": 270960,
    "text": "Two years ago the best autonomous coding"
  },
  {
    "startMs": 273800,
    "text": "agents you know solved on the fraction"
  },
  {
    "startMs": 276160,
    "text": "of the tasks on the standard software"
  },
  {
    "startMs": 278600,
    "text": "benchmark"
  },
  {
    "startMs": 279880,
    "text": "and now the best agents are in the high"
  },
  {
    "startMs": 281919,
    "text": "eighties. So we nearly tripled the"
  },
  {
    "startMs": 283840,
    "text": "amount of"
  },
  {
    "startMs": 285320,
    "text": "writing and shipping barely moved a"
  },
  {
    "startMs": 288240,
    "text": "third."
  },
  {
    "startMs": 289520,
    "text": "The benchmark was measuring the part of"
  },
  {
    "startMs": 291480,
    "text": "software engineering"
  },
  {
    "startMs": 293200,
    "text": "that has a greater"
  },
  {
    "startMs": 295760,
    "text": "and shipping is where all the ungraded"
  },
  {
    "startMs": 297720,
    "text": "parts come back in. So"
  },
  {
    "startMs": 301400,
    "text": "here is the rule underneath it. Anything"
  },
  {
    "startMs": 303919,
    "text": "that you can measure you can train"
  },
  {
    "startMs": 306000,
    "text": "against as Sarah Guo puts it."
  },
  {
    "startMs": 309320,
    "text": "A compiler is a free grader. A test"
  },
  {
    "startMs": 312040,
    "text": "suite is a free grader."
  },
  {
    "startMs": 314080,
    "text": "And the instant a task can grade itself"
  },
  {
    "startMs": 317040,
    "text": "you can grind a model against you know"
  },
  {
    "startMs": 320520,
    "text": "that grade until it wins. Automation"
  },
  {
    "startMs": 323919,
    "text": "of code was first because it's the most"
  },
  {
    "startMs": 326160,
    "text": "checkable thing that we have."
  },
  {
    "startMs": 328880,
    "text": "So implementation is converging for free"
  },
  {
    "startMs": 331520,
    "text": "for everyone at the same time"
  },
  {
    "startMs": 333880,
    "text": "and the most buildable thing and the"
  },
  {
    "startMs": 335680,
    "text": "most valuable thing are almost never the"
  },
  {
    "startMs": 338080,
    "text": "same thing."
  },
  {
    "startMs": 340400,
    "text": "So the model will build whatever you"
  },
  {
    "startMs": 342120,
    "text": "point it at but it will tell you nothing"
  },
  {
    "startMs": 344560,
    "text": "about where to point. Anything visible"
  },
  {
    "startMs": 347360,
    "text": "is replicatable."
  },
  {
    "startMs": 350120,
    "text": "So now, some people when they hear"
  },
  {
    "startMs": 351560,
    "text": "everything is implementable, they panic."
  },
  {
    "startMs": 354680,
    "text": "Um but we can flip the question."
  },
  {
    "startMs": 356920,
    "text": "The pointing is actually the job. It has"
  },
  {
    "startMs": 359880,
    "text": "always been the job. We just had so much"
  },
  {
    "startMs": 362440,
    "text": "implementation work in the way"
  },
  {
    "startMs": 365480,
    "text": "um that we never had to get good at it."
  },
  {
    "startMs": 368360,
    "text": "So, how do you decide where to point"
  },
  {
    "startMs": 369960,
    "text": "that?"
  },
  {
    "startMs": 371400,
    "text": "Paul Graham shared some wisdom on this."
  },
  {
    "startMs": 374840,
    "text": "Where you find something that people"
  },
  {
    "startMs": 376160,
    "text": "genuinely want is by feeling the need"
  },
  {
    "startMs": 378680,
    "text": "yourself."
  },
  {
    "startMs": 380720,
    "text": "Build something you and your friends"
  },
  {
    "startMs": 382200,
    "text": "need because the market hasn't formed"
  },
  {
    "startMs": 384320,
    "text": "yet, surveys can't see it, and your own"
  },
  {
    "startMs": 387240,
    "text": "need is the only signal that isn't a"
  },
  {
    "startMs": 389680,
    "text": "crap signal."
  },
  {
    "startMs": 391480,
    "text": "And the best ideas may sound genuinely"
  },
  {
    "startMs": 394360,
    "text": "lame at first, like a guy uh strapped"
  },
  {
    "startMs": 397360,
    "text": "with a with a camera on his head live"
  },
  {
    "startMs": 399800,
    "text": "streaming his his life. That sounds"
  },
  {
    "startMs": 402000,
    "text": "really ridiculous, but it became Twitch."
  },
  {
    "startMs": 404600,
    "text": "Um and the convergence machine doesn't"
  },
  {
    "startMs": 407000,
    "text": "really, you know, propose proactively"
  },
  {
    "startMs": 409400,
    "text": "these"
  },
  {
    "startMs": 410440,
    "text": "uh weird, specific, genuinely"
  },
  {
    "startMs": 412240,
    "text": "embarrassing ideas."
  },
  {
    "startMs": 414040,
    "text": "But even with the Twitch example,"
  },
  {
    "startMs": 416560,
    "text": "it worked, but there were a thousand"
  },
  {
    "startMs": 419720,
    "text": "other similar startups I start startup"
  },
  {
    "startMs": 421840,
    "text": "ideas that didn't."
  },
  {
    "startMs": 423919,
    "text": "So, the weird specific signal is"
  },
  {
    "startMs": 425919,
    "text": "necessary, but it is not sufficient."
  },
  {
    "startMs": 430080,
    "text": "It's also really tempting to say that"
  },
  {
    "startMs": 432919,
    "text": "we just need to have good judgment and"
  },
  {
    "startMs": 435160,
    "text": "good taste and call it safe. But taste"
  },
  {
    "startMs": 439160,
    "text": "is really just preference under"
  },
  {
    "startMs": 441000,
    "text": "feedback, and preference under feedback"
  },
  {
    "startMs": 443640,
    "text": "is exactly what these systems can learn."
  },
  {
    "startMs": 446480,
    "text": "Anything you can demonstrate enough"
  },
  {
    "startMs": 447919,
    "text": "times"
  },
  {
    "startMs": 449200,
    "text": "uh with a better or worse signal"
  },
  {
    "startMs": 451520,
    "text": "attached, the machine can eventually"
  },
  {
    "startMs": 453960,
    "text": "imitate. So, broad good taste is not"
  },
  {
    "startMs": 456960,
    "text": "really a differentiator."
  },
  {
    "startMs": 459560,
    "text": "What actually resists"
  },
  {
    "startMs": 461840,
    "text": "training is more narrow and more"
  },
  {
    "startMs": 464000,
    "text": "durable. So, two things."
  },
  {
    "startMs": 466520,
    "text": "Taste and judgement about what hasn't"
  },
  {
    "startMs": 468920,
    "text": "happened yet."
  },
  {
    "startMs": 470480,
    "text": "Because there's no data for an event"
  },
  {
    "startMs": 471880,
    "text": "that hasn't occurred. And then taste and"
  },
  {
    "startMs": 474160,
    "text": "judgement embedded in a relationship"
  },
  {
    "startMs": 476360,
    "text": "that the model can't observe."
  },
  {
    "startMs": 478440,
    "text": "What this customer in this situation"
  },
  {
    "startMs": 480840,
    "text": "with this history that you share"
  },
  {
    "startMs": 483240,
    "text": "actually needs."
  },
  {
    "startMs": 485680,
    "text": "The model has read everything ever"
  },
  {
    "startMs": 487640,
    "text": "written about your customer, but it has"
  },
  {
    "startMs": 489840,
    "text": "never actually met them."
  },
  {
    "startMs": 492320,
    "text": "So, if broad judgement is not safe, and"
  },
  {
    "startMs": 494680,
    "text": "the AI just handed everyone the ability"
  },
  {
    "startMs": 497360,
    "text": "to build anything, what's left to be"
  },
  {
    "startMs": 499840,
    "text": "good at?"
  },
  {
    "startMs": 501760,
    "text": "Richard Hamming uh spent his career"
  },
  {
    "startMs": 504200,
    "text": "studying why some scientists did great"
  },
  {
    "startMs": 507200,
    "text": "work and others who were just as smart"
  },
  {
    "startMs": 510240,
    "text": "didn't."
  },
  {
    "startMs": 512200,
    "text": "He found that the great ones worked on"
  },
  {
    "startMs": 514520,
    "text": "important problems."
  },
  {
    "startMs": 517520,
    "text": "And the problem isn't important because"
  },
  {
    "startMs": 519400,
    "text": "it just sounds impressive."
  },
  {
    "startMs": 522080,
    "text": "It's important when you have a"
  },
  {
    "startMs": 523479,
    "text": "reasonable attack on it. For example,"
  },
  {
    "startMs": 525920,
    "text": "time travel is consequential, he would"
  },
  {
    "startMs": 528200,
    "text": "say, but it's not important because"
  },
  {
    "startMs": 530240,
    "text": "nobody has an attack on it."
  },
  {
    "startMs": 532640,
    "text": "So, Hamming would tell you to keep 10 or"
  },
  {
    "startMs": 534720,
    "text": "20 ideas"
  },
  {
    "startMs": 536600,
    "text": "um"
  },
  {
    "startMs": 537440,
    "text": "on important problems in the back of"
  },
  {
    "startMs": 539000,
    "text": "your mind so that when you finally have"
  },
  {
    "startMs": 541200,
    "text": "an attack, a new tool, a new angle I"
  },
  {
    "startMs": 544120,
    "text": "think that only you noticed, then you go"
  },
  {
    "startMs": 546240,
    "text": "for it."
  },
  {
    "startMs": 548040,
    "text": "But in Hamming's world, the rarest thing"
  },
  {
    "startMs": 550360,
    "text": "was having an attack. And AI just handed"
  },
  {
    "startMs": 553240,
    "text": "everyone an attack on everything."
  },
  {
    "startMs": 555960,
    "text": "So, the rarest thing is knowing which"
  },
  {
    "startMs": 557839,
    "text": "problem is actually worth attacking. And"
  },
  {
    "startMs": 560040,
    "text": "that judgement comes from being a real"
  },
  {
    "startMs": 562040,
    "text": "person, close to a real domain, with"
  },
  {
    "startMs": 564920,
    "text": "your own battle scars, your weirdly"
  },
  {
    "startMs": 567640,
    "text": "specific experience, the thing that you"
  },
  {
    "startMs": 569520,
    "text": "care about more than is reasonable."
  },
  {
    "startMs": 572320,
    "text": "And you don't actually need to be first."
  },
  {
    "startMs": 574320,
    "text": "You just need to be genuinely close to a"
  },
  {
    "startMs": 575960,
    "text": "problem you actually understand where"
  },
  {
    "startMs": 578680,
    "text": "your insight is in the delta between"
  },
  {
    "startMs": 580680,
    "text": "what AI has been trained on and what"
  },
  {
    "startMs": 583320,
    "text": "should exist."
  },
  {
    "startMs": 585200,
    "text": "So, let's say you did it. You found that"
  },
  {
    "startMs": 587600,
    "text": "sweet spot problem that"
  },
  {
    "startMs": 590080,
    "text": "the one that you had an honest attack"
  },
  {
    "startMs": 591800,
    "text": "on, that you built this thing. It's"
  },
  {
    "startMs": 593640,
    "text": "genuinely good. It's genuinely yours,"
  },
  {
    "startMs": 595880,
    "text": "not the average."
  },
  {
    "startMs": 598120,
    "text": "You can still lose uh because knowing"
  },
  {
    "startMs": 601200,
    "text": "your signal is only half the job. The"
  },
  {
    "startMs": 603880,
    "text": "other half is getting it from your head"
  },
  {
    "startMs": 607040,
    "text": "into the head of a person that it was"
  },
  {
    "startMs": 608560,
    "text": "meant for."
  },
  {
    "startMs": 609880,
    "text": "And it's about reaching the right"
  },
  {
    "startMs": 611000,
    "text": "people."
  },
  {
    "startMs": 612200,
    "text": "And what do most of us"
  },
  {
    "startMs": 613920,
    "text": "do for that?"
  },
  {
    "startMs": 615720,
    "text": "We make content. So, let's talk about"
  },
  {
    "startMs": 618080,
    "text": "what uh AI convergence machine does to"
  },
  {
    "startMs": 620760,
    "text": "that."
  },
  {
    "startMs": 623280,
    "text": "What happened to the internet in the"
  },
  {
    "startMs": 624480,
    "text": "last 2 years? Open any feed,"
  },
  {
    "startMs": 627800,
    "text": "everything has started to sound the"
  },
  {
    "startMs": 629120,
    "text": "same."
  },
  {
    "startMs": 630760,
    "text": "The same LinkedIn posts, the same, you"
  },
  {
    "startMs": 632640,
    "text": "know, three bullet points and a bold"
  },
  {
    "startMs": 634440,
    "text": "takeaway, and"
  },
  {
    "startMs": 636400,
    "text": "the same blog post that uh says nothing"
  },
  {
    "startMs": 638960,
    "text": "but actually looks very polished."
  },
  {
    "startMs": 642120,
    "text": "Um your readers can now pattern match AI"
  },
  {
    "startMs": 645360,
    "text": "in just half a second. So, if a model"
  },
  {
    "startMs": 648080,
    "text": "could have written your post from a"
  },
  {
    "startMs": 650000,
    "text": "one-line prompt, your reader brain just"
  },
  {
    "startMs": 652680,
    "text": "skips it for the same reason."
  },
  {
    "startMs": 656800,
    "text": "So, AI has really learned the algorithm."
  },
  {
    "startMs": 658720,
    "text": "It has learned the format that performs."
  },
  {
    "startMs": 660839,
    "text": "It has learned what gets clicks, and"
  },
  {
    "startMs": 663280,
    "text": "everyone wants to hand the machine a"
  },
  {
    "startMs": 665680,
    "text": "paragraph and say, you know, \"Make it"
  },
  {
    "startMs": 667160,
    "text": "viral. Make me rich.\" It fills every gap"
  },
  {
    "startMs": 670600,
    "text": "that you leave with sameness."
  },
  {
    "startMs": 673560,
    "text": "So, what do you put in and what do you"
  },
  {
    "startMs": 675560,
    "text": "let it fill in?"
  },
  {
    "startMs": 677200,
    "text": "Cuz these there there are two different"
  },
  {
    "startMs": 679240,
    "text": "ways to use this thing, and they look"
  },
  {
    "startMs": 681760,
    "text": "very identical from the outside. One is"
  },
  {
    "startMs": 684520,
    "text": "you give it an average prompt, and gives"
  },
  {
    "startMs": 686400,
    "text": "you the average output."
  },
  {
    "startMs": 688360,
    "text": "And you ship one more indistinguishable"
  },
  {
    "startMs": 690480,
    "text": "drop into an ocean of indistinguishable"
  },
  {
    "startMs": 693080,
    "text": "drops."
  },
  {
    "startMs": 694320,
    "text": "So, you've automated your own"
  },
  {
    "startMs": 696520,
    "text": "irrelevance very efficiently."
  },
  {
    "startMs": 698880,
    "text": "And two, you can bring in the part that"
  },
  {
    "startMs": 700920,
    "text": "it can't have, your specific point of"
  },
  {
    "startMs": 702760,
    "text": "view,"
  },
  {
    "startMs": 703920,
    "text": "the real story that you were actually in"
  },
  {
    "startMs": 705560,
    "text": "the room for, and then let the machine"
  },
  {
    "startMs": 707640,
    "text": "do the converging work, the formatting,"
  },
  {
    "startMs": 709960,
    "text": "the drafting, the algorithm"
  },
  {
    "startMs": 711200,
    "text": "optimization, the cleanup around the"
  },
  {
    "startMs": 713600,
    "text": "core that it"
  },
  {
    "startMs": 715040,
    "text": "could have never generated."
  },
  {
    "startMs": 718000,
    "text": "The signal distorts on the way out. So,"
  },
  {
    "startMs": 721520,
    "text": "you can have the signal perfectly clear"
  },
  {
    "startMs": 723760,
    "text": "for you and still watch it fall apart"
  },
  {
    "startMs": 726280,
    "text": "between your brain and your users'"
  },
  {
    "startMs": 728440,
    "text": "understanding of it."
  },
  {
    "startMs": 730600,
    "text": "And in my experience, it breaks in three"
  },
  {
    "startMs": 732680,
    "text": "places."
  },
  {
    "startMs": 734080,
    "text": "And there are fixes for each, but"
  },
  {
    "startMs": 736520,
    "text": "they're different depending on product,"
  },
  {
    "startMs": 738560,
    "text": "the type, and the size of the company."
  },
  {
    "startMs": 742040,
    "text": "One of them is source distortion, which"
  },
  {
    "startMs": 744520,
    "text": "is very common in startups."
  },
  {
    "startMs": 746920,
    "text": "Founders, actually, they usually know"
  },
  {
    "startMs": 749160,
    "text": "the signal so well"
  },
  {
    "startMs": 751240,
    "text": "that they always have this accidental"
  },
  {
    "startMs": 753760,
    "text": "gift of compressing it past legibility."
  },
  {
    "startMs": 757400,
    "text": "They often assume the context that the"
  },
  {
    "startMs": 759560,
    "text": "audience doesn't have, and the room"
  },
  {
    "startMs": 761760,
    "text": "hears something technically very cool,"
  },
  {
    "startMs": 764320,
    "text": "but they"
  },
  {
    "startMs": 765280,
    "text": "doesn't they don't really understand why"
  },
  {
    "startMs": 766839,
    "text": "it matters."
  },
  {
    "startMs": 768240,
    "text": "I helped this one YC company with uh"
  },
  {
    "startMs": 771360,
    "text": "this exact thing recently."
  },
  {
    "startMs": 773440,
    "text": "Absolutely brilliant founders, genuinely"
  },
  {
    "startMs": 775640,
    "text": "new product, but every pitch that they"
  },
  {
    "startMs": 778000,
    "text": "started um was you know starting with"
  },
  {
    "startMs": 781000,
    "text": "architecture, with the clever parts,"
  },
  {
    "startMs": 783200,
    "text": "with things that they were very proud"
  },
  {
    "startMs": 785240,
    "text": "of. But it really landed as noise"
  },
  {
    "startMs": 787640,
    "text": "because the customer pain has been"
  },
  {
    "startMs": 789520,
    "text": "deleted from the whole story."
  },
  {
    "startMs": 792480,
    "text": "So, we rewrote the opening to include"
  },
  {
    "startMs": 794640,
    "text": "the thing that the user hated, and this"
  },
  {
    "startMs": 797080,
    "text": "product actually killed. So, the same"
  },
  {
    "startMs": 799520,
    "text": "product, the same week, the next"
  },
  {
    "startMs": 801839,
    "text": "conversations converted into pilots, and"
  },
  {
    "startMs": 804760,
    "text": "then we turned that into repeatable GTM"
  },
  {
    "startMs": 807080,
    "text": "system."
  },
  {
    "startMs": 808920,
    "text": "Organization distortion is another type"
  },
  {
    "startMs": 811760,
    "text": "of distortion that almost every big"
  },
  {
    "startMs": 813600,
    "text": "company has."
  },
  {
    "startMs": 815400,
    "text": "As signal travels through layers of"
  },
  {
    "startMs": 817800,
    "text": "management, through legal, through"
  },
  {
    "startMs": 819720,
    "text": "sales, through every department, at"
  },
  {
    "startMs": 822040,
    "text": "every hand handoff, it gets rewound"
  },
  {
    "startMs": 824760,
    "text": "towards the average."
  },
  {
    "startMs": 826520,
    "text": "And this really doesn't come from"
  },
  {
    "startMs": 827800,
    "text": "incompetence, it comes from investment."
  },
  {
    "startMs": 830960,
    "text": "So, hand a founder and the person three"
  },
  {
    "startMs": 833480,
    "text": "layers down the same task and the same"
  },
  {
    "startMs": 836520,
    "text": "AI, and you get two different things."
  },
  {
    "startMs": 839520,
    "text": "Um the founder really sweats the"
  },
  {
    "startMs": 841280,
    "text": "unaverageable details because the"
  },
  {
    "startMs": 843320,
    "text": "outcome is really theirs and they're"
  },
  {
    "startMs": 845680,
    "text": "personally invested and affected by it."
  },
  {
    "startMs": 849200,
    "text": "And others just ship it to spec, they"
  },
  {
    "startMs": 851080,
    "text": "close Jira tickets, they were asked for,"
  },
  {
    "startMs": 854040,
    "text": "you know, something like compliance, not"
  },
  {
    "startMs": 856280,
    "text": "as much conviction."
  },
  {
    "startMs": 858360,
    "text": "So, a long delegation chain plus"
  },
  {
    "startMs": 861240,
    "text": "convergence machine is really a factory"
  },
  {
    "startMs": 863800,
    "text": "for automating the signal out of your"
  },
  {
    "startMs": 866280,
    "text": "own company."
  },
  {
    "startMs": 867960,
    "text": "So, the first instinct usually is to add"
  },
  {
    "startMs": 870080,
    "text": "more process, which adds layers,"
  },
  {
    "startMs": 873040,
    "text": "bureaucracy, and slows everything down."
  },
  {
    "startMs": 875640,
    "text": "And we don't want that. Um"
  },
  {
    "startMs": 877839,
    "text": "to fix this, we have to help take the"
  },
  {
    "startMs": 879920,
    "text": "signal back"
  },
  {
    "startMs": 881360,
    "text": "and reattach it to the outcome like a"
  },
  {
    "startMs": 883160,
    "text": "founder and add the very thin signal"
  },
  {
    "startMs": 886160,
    "text": "layer to your go-to-market engineering,"
  },
  {
    "startMs": 888320,
    "text": "where its only job is to validate and"
  },
  {
    "startMs": 890120,
    "text": "carry the original intent across the"
  },
  {
    "startMs": 893120,
    "text": "handoffs intact."
  },
  {
    "startMs": 895760,
    "text": "Machine distortion is another way you"
  },
  {
    "startMs": 897800,
    "text": "can lose signal. You write one careful"
  },
  {
    "startMs": 900080,
    "text": "launch,"
  },
  {
    "startMs": 901240,
    "text": "your claim, your evidence, and your"
  },
  {
    "startMs": 902920,
    "text": "scope is very clear, but then of course"
  },
  {
    "startMs": 905480,
    "text": "AI remixes it"
  },
  {
    "startMs": 907560,
    "text": "um into a tweet, into a sales deck, into"
  },
  {
    "startMs": 909800,
    "text": "a partner one-pager. For example, you"
  },
  {
    "startMs": 912360,
    "text": "might have had this one very narrow eval"
  },
  {
    "startMs": 915080,
    "text": "that scored 94%"
  },
  {
    "startMs": 917560,
    "text": "but it was repeated enough times that"
  },
  {
    "startMs": 919440,
    "text": "your customers actually heard it as a"
  },
  {
    "startMs": 921600,
    "text": "promise."
  },
  {
    "startMs": 923840,
    "text": "So, we see the same through line. Your"
  },
  {
    "startMs": 925440,
    "text": "signal has to survive the trip"
  },
  {
    "startMs": 927920,
    "text": "undistorted."
  },
  {
    "startMs": 930360,
    "text": "And this is something you can engineer."
  },
  {
    "startMs": 933360,
    "text": "So, we need a thin signal layer, a small"
  },
  {
    "startMs": 935400,
    "text": "deliberate function whose job is to make"
  },
  {
    "startMs": 938200,
    "text": "sure that what your users take away is"
  },
  {
    "startMs": 940760,
    "text": "still the specific thing you meant."
  },
  {
    "startMs": 943640,
    "text": "Say you're building a monitoring tool."
  },
  {
    "startMs": 945880,
    "text": "There are 12 other tools in this"
  },
  {
    "startMs": 948040,
    "text": "category, but yours does something"
  },
  {
    "startMs": 950080,
    "text": "different. It tells you what not to wake"
  },
  {
    "startMs": 952680,
    "text": "up for, for example. It stays quiet on"
  },
  {
    "startMs": 955560,
    "text": "the noise, so when you"
  },
  {
    "startMs": 957720,
    "text": "get paged at night, you believe it. So,"
  },
  {
    "startMs": 960880,
    "text": "that quiet, that trust earned by silence"
  },
  {
    "startMs": 963200,
    "text": "is your signal."
  },
  {
    "startMs": 965200,
    "text": "So, first, say it in one sentence with"
  },
  {
    "startMs": 967240,
    "text": "the limit built in. Definitely don't say"
  },
  {
    "startMs": 970320,
    "text": "intelligent AI-native observability"
  },
  {
    "startMs": 972880,
    "text": "platform."
  },
  {
    "startMs": 974760,
    "text": "Say something like uh stays quiet on"
  },
  {
    "startMs": 976800,
    "text": "anything it can't tie to a real user"
  },
  {
    "startMs": 979360,
    "text": "impact and shows you everything it"
  },
  {
    "startMs": 981960,
    "text": "silenced so you can overrule it. The"
  },
  {
    "startMs": 984480,
    "text": "promise and the scope are welded"
  },
  {
    "startMs": 986200,
    "text": "together here."
  },
  {
    "startMs": 988080,
    "text": "Then make sure that the limit can't be"
  },
  {
    "startMs": 989760,
    "text": "edited out. So, in the product, every"
  },
  {
    "startMs": 992400,
    "text": "suppressed alert is visible. In the"
  },
  {
    "startMs": 994880,
    "text": "launch, statements like 90% fewer pages"
  },
  {
    "startMs": 999040,
    "text": "uh live next to statements like every"
  },
  {
    "startMs": 1001400,
    "text": "silence is visible and reversible. So,"
  },
  {
    "startMs": 1003760,
    "text": "when AI chops your launch into a tweet,"
  },
  {
    "startMs": 1006800,
    "text": "it can keep the impressive number, but"
  },
  {
    "startMs": 1009560,
    "text": "also remove the part that"
  },
  {
    "startMs": 1011960,
    "text": "keeps the part that uh keeps the product"
  },
  {
    "startMs": 1013960,
    "text": "honest."
  },
  {
    "startMs": 1015520,
    "text": "And before you scale it, check what"
  },
  {
    "startMs": 1017240,
    "text": "people actually heard. So, give a readme"
  },
  {
    "startMs": 1020040,
    "text": "to an SRE who has never seen your"
  },
  {
    "startMs": 1022440,
    "text": "project and ask a person to describe the"
  },
  {
    "startMs": 1025480,
    "text": "product back to you. The gap between"
  },
  {
    "startMs": 1028040,
    "text": "what they say and what you meant is the"
  },
  {
    "startMs": 1030160,
    "text": "distortion that you were about to"
  },
  {
    "startMs": 1032160,
    "text": "broadcast."
  },
  {
    "startMs": 1034040,
    "text": "And it's a very lightweight signal"
  },
  {
    "startMs": 1035360,
    "text": "layer, and a lot of it is buildable, So,"
  },
  {
    "startMs": 1037560,
    "text": "you can automate more of the checking"
  },
  {
    "startMs": 1039400,
    "text": "and the catching and the surveying than"
  },
  {
    "startMs": 1041280,
    "text": "most people realize."
  },
  {
    "startMs": 1043480,
    "text": "So, step back and ask what all of this,"
  },
  {
    "startMs": 1046800,
    "text": "the building, the shipping, the"
  },
  {
    "startMs": 1048240,
    "text": "undistorted signal, is actually for."
  },
  {
    "startMs": 1051320,
    "text": "It's for one thing of getting a human or"
  },
  {
    "startMs": 1054720,
    "text": "increasingly an agent to choose you and"
  },
  {
    "startMs": 1057520,
    "text": "rely on you when they have an infinite"
  },
  {
    "startMs": 1060920,
    "text": "identical-looking alternatives. So,"
  },
  {
    "startMs": 1062920,
    "text": "that's trust. Trust is the one thing"
  },
  {
    "startMs": 1065360,
    "text": "that's left with no greater. There's no"
  },
  {
    "startMs": 1068040,
    "text": "benchmark for it, no reward signal. It"
  },
  {
    "startMs": 1070680,
    "text": "can't be entirely automated because it's"
  },
  {
    "startMs": 1073120,
    "text": "granted slowly through relationship with"
  },
  {
    "startMs": 1076360,
    "text": "consent. For example, doctors who"
  },
  {
    "startMs": 1079640,
    "text": "open one particular tool every morning,"
  },
  {
    "startMs": 1082040,
    "text": "they didn't have that habit trained into"
  },
  {
    "startMs": 1083960,
    "text": "them."
  },
  {
    "startMs": 1085280,
    "text": "And what happens if we get this wrong?"
  },
  {
    "startMs": 1088040,
    "text": "Getting your signal wrong is actually"
  },
  {
    "startMs": 1090240,
    "text": "not neutral. It's negative."
  },
  {
    "startMs": 1092880,
    "text": "Producing averageness is not free. You"
  },
  {
    "startMs": 1095280,
    "text": "actually pay for it in tokens, in infra,"
  },
  {
    "startMs": 1098200,
    "text": "in the salaried hours of good people,"
  },
  {
    "startMs": 1101040,
    "text": "you know, with with customers that"
  },
  {
    "startMs": 1103440,
    "text": "take a look at your product once, decide"
  },
  {
    "startMs": 1106200,
    "text": "once, and never come back. So, every"
  },
  {
    "startMs": 1108440,
    "text": "generic post teaches them that your name"
  },
  {
    "startMs": 1111320,
    "text": "isn't worth the click. So, you spend"
  },
  {
    "startMs": 1113960,
    "text": "real money to make yourself harder to"
  },
  {
    "startMs": 1115760,
    "text": "choose."
  },
  {
    "startMs": 1117360,
    "text": "So, back to the main question."
  },
  {
    "startMs": 1119440,
    "text": "We got faster,"
  },
  {
    "startMs": 1121800,
    "text": "but the speed is not where the value"
  },
  {
    "startMs": 1123360,
    "text": "went."
  },
  {
    "startMs": 1124520,
    "text": "Uh the value moved up to deciding what"
  },
  {
    "startMs": 1126800,
    "text": "is worth building, what is worth saying,"
  },
  {
    "startMs": 1129080,
    "text": "what deserves trust. And where does the"
  },
  {
    "startMs": 1131240,
    "text": "thing that you actually"
  },
  {
    "startMs": 1133080,
    "text": "um that that you meant survives the trip"
  },
  {
    "startMs": 1135920,
    "text": "to the people that it was for."
  },
  {
    "startMs": 1138120,
    "text": "So, you don't need to be first. You need"
  },
  {
    "startMs": 1139760,
    "text": "a real problem and enough conviction to"
  },
  {
    "startMs": 1142600,
    "text": "carry the signal clearly through to, you"
  },
  {
    "startMs": 1145680,
    "text": "know, right people to find it."
  },
  {
    "startMs": 1148160,
    "text": "So, when you can build anything, you"
  },
  {
    "startMs": 1150280,
    "text": "should build trust."
  },
  {
    "startMs": 1152120,
    "text": "Have the strongest conviction, define"
  },
  {
    "startMs": 1154320,
    "text": "the signal yourself, protect it from"
  },
  {
    "startMs": 1156440,
    "text": "distortion, and use AI aggressively for"
  },
  {
    "startMs": 1159360,
    "text": "everything else."
  },
  {
    "startMs": 1161160,
    "text": "Thank you. Let's connect and happy to"
  },
  {
    "startMs": 1162960,
    "text": "chat with you afterwards."
  },
  {
    "startMs": 1164800,
    "text": "Thank you."
  },
  {
    "startMs": 1166568,
    "text": ">> [applause]"
  },
  {
    "startMs": 1180323,
    "text": "[music]"
  }
]
