[
  {
    "startMs": 1309,
    "text": "[music]"
  },
  {
    "startMs": 12720,
    "text": ">> Good afternoon. My name is Alon Blum. I"
  },
  {
    "startMs": 15640,
    "text": "am a software engineer at Figma."
  },
  {
    "startMs": 17680,
    "text": "And in my talk today, we're going to"
  },
  {
    "startMs": 19920,
    "text": "talk about how we've adopted or are"
  },
  {
    "startMs": 22280,
    "text": "adopting"
  },
  {
    "startMs": 24880,
    "text": "agent into our workflow at Figma"
  },
  {
    "startMs": 27080,
    "text": "while maintaining"
  },
  {
    "startMs": 28760,
    "text": "high quality for our code base."
  },
  {
    "startMs": 31960,
    "text": "So, as you may know, Figma is the"
  },
  {
    "startMs": 35160,
    "text": "browser-based"
  },
  {
    "startMs": 36920,
    "text": "editor where design and engineering and"
  },
  {
    "startMs": 39600,
    "text": "now AI agent collaborate together to"
  },
  {
    "startMs": 42360,
    "text": "ship code."
  },
  {
    "startMs": 43920,
    "text": "Uh this"
  },
  {
    "startMs": 45840,
    "text": "Figma has"
  },
  {
    "startMs": 47240,
    "text": "pivoted very strongly from being a"
  },
  {
    "startMs": 48840,
    "text": "traditional tool to an AI-first tool."
  },
  {
    "startMs": 52160,
    "text": "But in this talk, I'm not going to talk"
  },
  {
    "startMs": 53640,
    "text": "about our product. I'm going to talk"
  },
  {
    "startMs": 54840,
    "text": "more about our internal organization and"
  },
  {
    "startMs": 57360,
    "text": "how our engineering org has been"
  },
  {
    "startMs": 60600,
    "text": "adopting AI agents."
  },
  {
    "startMs": 64440,
    "text": "Um what we we found internally is both"
  },
  {
    "startMs": 68200,
    "text": "organizations, companies, and individual"
  },
  {
    "startMs": 71400,
    "text": "there's kind of a three-act"
  },
  {
    "startMs": 73760,
    "text": "process of AI adoption."
  },
  {
    "startMs": 76560,
    "text": "You start with picking up something,"
  },
  {
    "startMs": 78800,
    "text": "whether it was a lot of the people in"
  },
  {
    "startMs": 80600,
    "text": "this room who have been doing using our"
  },
  {
    "startMs": 83160,
    "text": "AI pal and have been using AI for a"
  },
  {
    "startMs": 85320,
    "text": "while and they picked up something and"
  },
  {
    "startMs": 87400,
    "text": "got some simple things to work very"
  },
  {
    "startMs": 89120,
    "text": "well."
  },
  {
    "startMs": 90240,
    "text": "10x faster."
  },
  {
    "startMs": 92400,
    "text": "Then you start applying those same"
  },
  {
    "startMs": 94400,
    "text": "practices to bigger problems and AI"
  },
  {
    "startMs": 97000,
    "text": "fails pretty badly at that, gives you"
  },
  {
    "startMs": 99280,
    "text": "bad stuff,"
  },
  {
    "startMs": 100800,
    "text": "lots of bugs,"
  },
  {
    "startMs": 102240,
    "text": "and the trust that you build breaks"
  },
  {
    "startMs": 104520,
    "text": "down."
  },
  {
    "startMs": 105760,
    "text": "And then"
  },
  {
    "startMs": 106760,
    "text": "from that point, you start building the"
  },
  {
    "startMs": 108440,
    "text": "real skill, which is learning how to use"
  },
  {
    "startMs": 111880,
    "text": "AI correctly and put the right"
  },
  {
    "startMs": 113960,
    "text": "guardrails and the right prompting and"
  },
  {
    "startMs": 115560,
    "text": "the right context and all the stuff that"
  },
  {
    "startMs": 117240,
    "text": "we've been talking all day about here in"
  },
  {
    "startMs": 119480,
    "text": "all the talks in order to actually build"
  },
  {
    "startMs": 121400,
    "text": "a real scale."
  },
  {
    "startMs": 123800,
    "text": "And one thing that"
  },
  {
    "startMs": 126280,
    "text": "is happening internally as we we adopted"
  },
  {
    "startMs": 129360,
    "text": "whether teams or individuals"
  },
  {
    "startMs": 131800,
    "text": "the adoption is uneven. We have teams"
  },
  {
    "startMs": 134400,
    "text": "that are very AI forward and have"
  },
  {
    "startMs": 137280,
    "text": "already transformed their entire"
  },
  {
    "startMs": 138560,
    "text": "workflows and then we have teams that"
  },
  {
    "startMs": 140520,
    "text": "are still experimenting in the earlier"
  },
  {
    "startMs": 144000,
    "text": "act and or have lost confidence and they"
  },
  {
    "startMs": 146240,
    "text": "all need to work together in order to"
  },
  {
    "startMs": 147520,
    "text": "ship our product. Um"
  },
  {
    "startMs": 150640,
    "text": "So, they need to coexist"
  },
  {
    "startMs": 153080,
    "text": "in the organization and we need to find"
  },
  {
    "startMs": 154880,
    "text": "a way to support them and while bringing"
  },
  {
    "startMs": 156840,
    "text": "on everybody along for the journey and"
  },
  {
    "startMs": 159400,
    "text": "getting everybody to the third act of"
  },
  {
    "startMs": 161840,
    "text": "the story."
  },
  {
    "startMs": 164520,
    "text": "Aside from that main friction point, we"
  },
  {
    "startMs": 166720,
    "text": "have also noticed other friction points"
  },
  {
    "startMs": 168720,
    "text": "that happened"
  },
  {
    "startMs": 170200,
    "text": "as we adopt AI."
  },
  {
    "startMs": 173959,
    "text": "One thing that we've heard a lot from"
  },
  {
    "startMs": 175360,
    "text": "developers and managers have have been"
  },
  {
    "startMs": 177120,
    "text": "noticing is that reduced developer"
  },
  {
    "startMs": 178760,
    "text": "agency causes"
  },
  {
    "startMs": 180720,
    "text": "um"
  },
  {
    "startMs": 181360,
    "text": "engineers to lose some of their job"
  },
  {
    "startMs": 183239,
    "text": "satisfaction. So, if a lot of people"
  },
  {
    "startMs": 185480,
    "text": "used to take a lot of pride and"
  },
  {
    "startMs": 187320,
    "text": "enjoyment in writing code and getting"
  },
  {
    "startMs": 189160,
    "text": "into the flow"
  },
  {
    "startMs": 190519,
    "text": "and a lot of people feel like that's"
  },
  {
    "startMs": 192160,
    "text": "been lost or they're losing a lot of"
  },
  {
    "startMs": 193800,
    "text": "that"
  },
  {
    "startMs": 194880,
    "text": "element and getting into more of a"
  },
  {
    "startMs": 196640,
    "text": "prompt cycle where they just wait on"
  },
  {
    "startMs": 198480,
    "text": "output from AI and then speak to the AI"
  },
  {
    "startMs": 200280,
    "text": "that like not as much fun as they used"
  },
  {
    "startMs": 202000,
    "text": "to have and they're getting burned out."
  },
  {
    "startMs": 205320,
    "text": "Um we've noticed another interesting"
  },
  {
    "startMs": 207519,
    "text": "thing. It's actually our best engineer,"
  },
  {
    "startMs": 209080,
    "text": "the one that hold all their contacts in"
  },
  {
    "startMs": 210560,
    "text": "their brain. Um they end up getting out"
  },
  {
    "startMs": 213320,
    "text": "of the burden. What ends up happening is"
  },
  {
    "startMs": 215640,
    "text": "they they know where all the pitfalls"
  },
  {
    "startMs": 217239,
    "text": "are. They are like holding together with"
  },
  {
    "startMs": 220239,
    "text": "with like their mental duct tape all the"
  },
  {
    "startMs": 222440,
    "text": "places"
  },
  {
    "startMs": 223800,
    "text": "that agents are not working well and"
  },
  {
    "startMs": 225920,
    "text": "they're preventing all the really bad"
  },
  {
    "startMs": 227519,
    "text": "stuff from coming in or all they they"
  },
  {
    "startMs": 229640,
    "text": "have all the institutional contact that"
  },
  {
    "startMs": 232080,
    "text": "have never written down in their head"
  },
  {
    "startMs": 234080,
    "text": "and they get so much burden and and"
  },
  {
    "startMs": 236680,
    "text": "become bottlenecks and gets really"
  },
  {
    "startMs": 238120,
    "text": "frustrated. So, they actually end up"
  },
  {
    "startMs": 240120,
    "text": "being slowest to adopt because they see"
  },
  {
    "startMs": 241760,
    "text": "all the problem"
  },
  {
    "startMs": 243080,
    "text": "uh first hand."
  },
  {
    "startMs": 244520,
    "text": "That's another big big issue that we've"
  },
  {
    "startMs": 246160,
    "text": "seen."
  },
  {
    "startMs": 248000,
    "text": "Um and this one I'm sure everybody can"
  },
  {
    "startMs": 250840,
    "text": "resonate or in Sorry, I'm sure everybody"
  },
  {
    "startMs": 254000,
    "text": "here will resonate."
  },
  {
    "startMs": 255800,
    "text": "Um that all of a sudden all the design"
  },
  {
    "startMs": 257720,
    "text": "docs and all the Slack messages you"
  },
  {
    "startMs": 260359,
    "text": "know, this the emails have gotten three"
  },
  {
    "startMs": 262120,
    "text": "or four times as long and we've gotten"
  },
  {
    "startMs": 263960,
    "text": "two or three times as many emails"
  },
  {
    "startMs": 266200,
    "text": "and they say basically as much as they"
  },
  {
    "startMs": 268320,
    "text": "did before. So, communication has gotten"
  },
  {
    "startMs": 271360,
    "text": "quite inefficient and some of the"
  },
  {
    "startMs": 273280,
    "text": "markers of like what is high quality and"
  },
  {
    "startMs": 275240,
    "text": "important things versus not so much high"
  },
  {
    "startMs": 277200,
    "text": "quality"
  },
  {
    "startMs": 278320,
    "text": "um has become challenging to navigate."
  },
  {
    "startMs": 282520,
    "text": "Um so, I'm going to spend uh the next"
  },
  {
    "startMs": 285280,
    "text": "few minutes talking about some of the"
  },
  {
    "startMs": 287040,
    "text": "lessons that we've learned and how we've"
  },
  {
    "startMs": 289160,
    "text": "been trying to apply this. This is a"
  },
  {
    "startMs": 291120,
    "text": "journey. We have not come out through"
  },
  {
    "startMs": 292720,
    "text": "the other end, but we've seen some"
  },
  {
    "startMs": 294040,
    "text": "really interesting progress"
  },
  {
    "startMs": 295880,
    "text": "along a lot of these lines."
  },
  {
    "startMs": 300200,
    "text": "Um I think this"
  },
  {
    "startMs": 302840,
    "text": "uh a lot of the speakers here have"
  },
  {
    "startMs": 304040,
    "text": "touched upon this, but investing in"
  },
  {
    "startMs": 306000,
    "text": "verification is probably the highest"
  },
  {
    "startMs": 308000,
    "text": "value thing we can do in our code base."
  },
  {
    "startMs": 310400,
    "text": "Um anytime that we can lift a"
  },
  {
    "startMs": 314240,
    "text": "left shift anything in our workflow from"
  },
  {
    "startMs": 316800,
    "text": "a human needing to do it to an agent"
  },
  {
    "startMs": 318800,
    "text": "being able to verify it."
  },
  {
    "startMs": 320760,
    "text": "So, for example, when uh Playwright and"
  },
  {
    "startMs": 324760,
    "text": "MCP came out, instead of having humans"
  },
  {
    "startMs": 327440,
    "text": "navigate the code, now the agent can"
  },
  {
    "startMs": 328919,
    "text": "explore the code. That was a big win"
  },
  {
    "startMs": 331240,
    "text": "unlock for productivity in a lot of our"
  },
  {
    "startMs": 333280,
    "text": "team. That's really That's always a"
  },
  {
    "startMs": 336480,
    "text": "a big win for us."
  },
  {
    "startMs": 338320,
    "text": "The other thing is"
  },
  {
    "startMs": 340400,
    "text": "um"
  },
  {
    "startMs": 341520,
    "text": "it's even better if when you find"
  },
  {
    "startMs": 343560,
    "text": "something that the agent has found to be"
  },
  {
    "startMs": 345720,
    "text": "useful,"
  },
  {
    "startMs": 346880,
    "text": "take the time to take that and encode"
  },
  {
    "startMs": 349120,
    "text": "into a deterministic flow."
  },
  {
    "startMs": 351440,
    "text": "A deterministic flow that can be easily"
  },
  {
    "startMs": 353200,
    "text": "repeated is saved on tokens, save on"
  },
  {
    "startMs": 355160,
    "text": "time for the and then it also you also"
  },
  {
    "startMs": 357000,
    "text": "know that you're using the the LLM when"
  },
  {
    "startMs": 360960,
    "text": "it needs to reason, but when you have"
  },
  {
    "startMs": 363040,
    "text": "something that is already"
  },
  {
    "startMs": 365400,
    "text": "known and basically can be encoded into"
  },
  {
    "startMs": 367440,
    "text": "a test, spending that time always always"
  },
  {
    "startMs": 369760,
    "text": "pays dividends."
  },
  {
    "startMs": 371320,
    "text": "Um"
  },
  {
    "startMs": 372360,
    "text": "And another tip, if you tell your scale"
  },
  {
    "startMs": 375480,
    "text": "your agent to write the code that you're"
  },
  {
    "startMs": 377160,
    "text": "writing"
  },
  {
    "startMs": 378280,
    "text": "um like at the red to green to red to"
  },
  {
    "startMs": 381440,
    "text": "green at the TDD style,"
  },
  {
    "startMs": 384520,
    "text": "it almost always gives you better"
  },
  {
    "startMs": 385680,
    "text": "results because you set a goal, then you"
  },
  {
    "startMs": 388360,
    "text": "tell the agent to strive toward that"
  },
  {
    "startMs": 389960,
    "text": "goal, it will almost always give you"
  },
  {
    "startMs": 391960,
    "text": "better results than writing the code and"
  },
  {
    "startMs": 393400,
    "text": "then writing the test afterward because"
  },
  {
    "startMs": 394919,
    "text": "then it will fit the test to the code"
  },
  {
    "startMs": 397000,
    "text": "rather than fit the code to pass the"
  },
  {
    "startMs": 398960,
    "text": "verification criteria."
  },
  {
    "startMs": 402480,
    "text": "Um this is the testing pyramid that uh"
  },
  {
    "startMs": 404520,
    "text": "can't the classic testing pyramid from"
  },
  {
    "startMs": 406680,
    "text": "the"
  },
  {
    "startMs": 407440,
    "text": "previous"
  },
  {
    "startMs": 408840,
    "text": "uh just when you think about the testing"
  },
  {
    "startMs": 410200,
    "text": "themselves, which you had the end-to-end"
  },
  {
    "startMs": 412600,
    "text": "test and the integration test and the"
  },
  {
    "startMs": 414040,
    "text": "unit test. This is very similar."
  },
  {
    "startMs": 417280,
    "text": "Move as much as you can down to the"
  },
  {
    "startMs": 419120,
    "text": "deterministic analysis where that's"
  },
  {
    "startMs": 420919,
    "text": "linting, the compiler, um the unit test"
  },
  {
    "startMs": 424200,
    "text": "themselves."
  },
  {
    "startMs": 425520,
    "text": "Whatever that can come be covered"
  },
  {
    "startMs": 427240,
    "text": "easily, you can have engine agent do"
  },
  {
    "startMs": 429160,
    "text": "reviews on it based on on criteria, so"
  },
  {
    "startMs": 432880,
    "text": "um"
  },
  {
    "startMs": 433880,
    "text": "architectural standards that that have"
  },
  {
    "startMs": 435560,
    "text": "been easily encoded into the code base,"
  },
  {
    "startMs": 438200,
    "text": "you can move into the agent. And then"
  },
  {
    "startMs": 440560,
    "text": "only at the very top you need to have"
  },
  {
    "startMs": 442280,
    "text": "some sort of human review, which is"
  },
  {
    "startMs": 444160,
    "text": "usually around the functionality and"
  },
  {
    "startMs": 446040,
    "text": "this is the right thing to build."
  },
  {
    "startMs": 448240,
    "text": "That like only leave the human to do"
  },
  {
    "startMs": 450080,
    "text": "what the humans need to actually be"
  },
  {
    "startMs": 451600,
    "text": "involved in."
  },
  {
    "startMs": 454840,
    "text": "Um Another really important thing is the"
  },
  {
    "startMs": 457280,
    "text": "planning"
  },
  {
    "startMs": 458520,
    "text": "versus prompting. This is really tied"
  },
  {
    "startMs": 460160,
    "text": "into the giving agency back to"
  },
  {
    "startMs": 462640,
    "text": "developers and finding a replacement to"
  },
  {
    "startMs": 465520,
    "text": "the craft of writing code."
  },
  {
    "startMs": 468880,
    "text": "Um"
  },
  {
    "startMs": 470160,
    "text": "spending a lot of time writing the plan"
  },
  {
    "startMs": 472000,
    "text": "and then"
  },
  {
    "startMs": 473240,
    "text": "sending enough to the agent basically as"
  },
  {
    "startMs": 474800,
    "text": "a"
  },
  {
    "startMs": 475760,
    "text": "as an implementation that can be done"
  },
  {
    "startMs": 477680,
    "text": "automatically is something that we find"
  },
  {
    "startMs": 480080,
    "text": "to really"
  },
  {
    "startMs": 482640,
    "text": "kind of"
  },
  {
    "startMs": 483520,
    "text": "reintroduce the joy of of building back"
  },
  {
    "startMs": 486000,
    "text": "into the process."
  },
  {
    "startMs": 487680,
    "text": "And so it's not uncommon to spend a week"
  },
  {
    "startMs": 491360,
    "text": "writing a very detailed plan, making all"
  },
  {
    "startMs": 493440,
    "text": "the decisions, flushing it out,"
  },
  {
    "startMs": 494840,
    "text": "iterating, sending it out to teammates"
  },
  {
    "startMs": 496640,
    "text": "to review."
  },
  {
    "startMs": 498000,
    "text": "And then only when it's ready and you've"
  },
  {
    "startMs": 499680,
    "text": "flushed out all the decision, you can"
  },
  {
    "startMs": 501400,
    "text": "send it to the agent. The agent will"
  },
  {
    "startMs": 503400,
    "text": "um send it back to you when it's"
  },
  {
    "startMs": 505040,
    "text": "implemented."
  },
  {
    "startMs": 506360,
    "text": "And that that has been really successful"
  },
  {
    "startMs": 509560,
    "text": "also in accelerating and also"
  },
  {
    "startMs": 512760,
    "text": "really restoring some of the joy into"
  },
  {
    "startMs": 515479,
    "text": "the development process."
  },
  {
    "startMs": 518680,
    "text": "Uh so what makes a good plan? Um"
  },
  {
    "startMs": 521760,
    "text": "really important to start with a why at"
  },
  {
    "startMs": 523560,
    "text": "the top. It really helps preventing"
  },
  {
    "startMs": 525080,
    "text": "agent drift. If you have like a bold big"
  },
  {
    "startMs": 527360,
    "text": "section of kind of like it when you"
  },
  {
    "startMs": 528680,
    "text": "write a design doc, you want to have the"
  },
  {
    "startMs": 530240,
    "text": "executive summary. Put that in there for"
  },
  {
    "startMs": 532400,
    "text": "the agent. Otherwise, they'll start"
  },
  {
    "startMs": 533640,
    "text": "drifting over time and make sure that"
  },
  {
    "startMs": 535560,
    "text": "the agent don't go back and change that"
  },
  {
    "startMs": 537360,
    "text": "because they feel like it."
  },
  {
    "startMs": 539000,
    "text": "Uh so we start with a why."
  },
  {
    "startMs": 541400,
    "text": "Make sure that the plan can be broken"
  },
  {
    "startMs": 542960,
    "text": "down"
  },
  {
    "startMs": 544000,
    "text": "into small parts that can each be"
  },
  {
    "startMs": 546839,
    "text": "verified independently."
  },
  {
    "startMs": 548800,
    "text": "And my personal way of knowing what is a"
  },
  {
    "startMs": 552160,
    "text": "good size would I want to review that"
  },
  {
    "startMs": 554000,
    "text": "the PR that will correspond to that"
  },
  {
    "startMs": 555760,
    "text": "part? If it's going to be too big for me"
  },
  {
    "startMs": 557240,
    "text": "to want to review in one sitting, it's"
  },
  {
    "startMs": 558800,
    "text": "kind of like the test is"
  },
  {
    "startMs": 560320,
    "text": "I'm going to get need to get a cup of"
  },
  {
    "startMs": 561400,
    "text": "coffee before I read this."
  },
  {
    "startMs": 563160,
    "text": "That means it's too big and I'm going to"
  },
  {
    "startMs": 564839,
    "text": "want to have it broken down into pieces."
  },
  {
    "startMs": 567720,
    "text": "And then"
  },
  {
    "startMs": 569000,
    "text": "I make sure that each part can be"
  },
  {
    "startMs": 570880,
    "text": "validated independently cuz what I don't"
  },
  {
    "startMs": 572839,
    "text": "want to have is"
  },
  {
    "startMs": 574240,
    "text": "have five stages and then the first one"
  },
  {
    "startMs": 577320,
    "text": "is written but not validated, and then"
  },
  {
    "startMs": 578880,
    "text": "everything else is is built on top of"
  },
  {
    "startMs": 580880,
    "text": "all the assumptions. So, having kind of"
  },
  {
    "startMs": 583040,
    "text": "a validation gate or an exception"
  },
  {
    "startMs": 584880,
    "text": "criteria for each phase really helps and"
  },
  {
    "startMs": 588320,
    "text": "make the plan uh resilient to drift. And"
  },
  {
    "startMs": 592240,
    "text": "all and and there's all kind of"
  },
  {
    "startMs": 593920,
    "text": "technique on how to manage the contacts"
  },
  {
    "startMs": 595440,
    "text": "and"
  },
  {
    "startMs": 596440,
    "text": "doing a a software factory on top of"
  },
  {
    "startMs": 599160,
    "text": "that. But once you have the plan, you"
  },
  {
    "startMs": 600480,
    "text": "can use whatever loop uh you want or"
  },
  {
    "startMs": 603000,
    "text": "whatever workflow you want"
  },
  {
    "startMs": 605080,
    "text": "in order to implement it."
  },
  {
    "startMs": 607600,
    "text": "Uh"
  },
  {
    "startMs": 608160,
    "text": "this is a screenshot that I randomly"
  },
  {
    "startMs": 610040,
    "text": "picked up a plan, but this is what I"
  },
  {
    "startMs": 611680,
    "text": "usually look for. The executive summary"
  },
  {
    "startMs": 613800,
    "text": "at the top, the phases break it down,"
  },
  {
    "startMs": 615880,
    "text": "and then each one of them I would go"
  },
  {
    "startMs": 617200,
    "text": "into lots of details so that I can just"
  },
  {
    "startMs": 618960,
    "text": "fit it into a sub agent, and the sub"
  },
  {
    "startMs": 621120,
    "text": "agent can independently work on that and"
  },
  {
    "startMs": 623240,
    "text": "not have to worry about it. Um that's"
  },
  {
    "startMs": 626160,
    "text": "that's it. There are other workflows"
  },
  {
    "startMs": 627720,
    "text": "that would work or other structure to"
  },
  {
    "startMs": 629520,
    "text": "the plan. I find that"
  },
  {
    "startMs": 632040,
    "text": "part of the things that great about uh"
  },
  {
    "startMs": 634200,
    "text": "AI workflows is that everybody can set"
  },
  {
    "startMs": 636400,
    "text": "up the thing that works best for them."
  },
  {
    "startMs": 638040,
    "text": "Oh-oh."
  },
  {
    "startMs": 641160,
    "text": "No, thank you."
  },
  {
    "startMs": 643720,
    "text": "Everybody can very easily set up the"
  },
  {
    "startMs": 645600,
    "text": "workflow that work exactly for them for"
  },
  {
    "startMs": 647320,
    "text": "them. So, there's"
  },
  {
    "startMs": 649080,
    "text": "diminishing return in trying to"
  },
  {
    "startMs": 650200,
    "text": "centralize everybody on one thing, but"
  },
  {
    "startMs": 651920,
    "text": "as long as it works for their flow and"
  },
  {
    "startMs": 653640,
    "text": "other people can iterate with them, I"
  },
  {
    "startMs": 655680,
    "text": "find that it generally works very well."
  },
  {
    "startMs": 659800,
    "text": "And this is just an example kind of a"
  },
  {
    "startMs": 661680,
    "text": "brag of like this is uh could be a"
  },
  {
    "startMs": 663640,
    "text": "result from a plan. Um there are"
  },
  {
    "startMs": 666080,
    "text": "probably 20 PRs here. Some of them would"
  },
  {
    "startMs": 668600,
    "text": "be maybe 10 lines, and some of them"
  },
  {
    "startMs": 670240,
    "text": "would be 100 lines. There's probably"
  },
  {
    "startMs": 671560,
    "text": "nothing bigger than that, and that"
  },
  {
    "startMs": 673800,
    "text": "allows us to"
  },
  {
    "startMs": 675600,
    "text": "This is in the pre-AI world, this plan"
  },
  {
    "startMs": 678000,
    "text": "probably worked in that for a week. I"
  },
  {
    "startMs": 680400,
    "text": "aligned with the other with three other"
  },
  {
    "startMs": 681960,
    "text": "teams for another week on that, and then"
  },
  {
    "startMs": 683520,
    "text": "I just sent it to an agent to implement"
  },
  {
    "startMs": 685839,
    "text": "overnight, and it came back. This is"
  },
  {
    "startMs": 687680,
    "text": "probably from two plans, not one, but"
  },
  {
    "startMs": 689880,
    "text": "it's it's basically six weeks of of"
  },
  {
    "startMs": 692480,
    "text": "coding work just It's"
  },
  {
    "startMs": 694640,
    "text": "um"
  },
  {
    "startMs": 695440,
    "text": "only took 1 week, so that's where"
  },
  {
    "startMs": 697600,
    "text": "I got the 5x speed up. If I include the"
  },
  {
    "startMs": 701080,
    "text": "review cycle at the end that we always"
  },
  {
    "startMs": 702440,
    "text": "have to remember."
  },
  {
    "startMs": 705760,
    "text": "Um moving on from planning"
  },
  {
    "startMs": 709360,
    "text": "back to the issue that we had with the"
  },
  {
    "startMs": 711200,
    "text": "skeptics and the people who are burdened"
  },
  {
    "startMs": 713040,
    "text": "with the most work, make sure that"
  },
  {
    "startMs": 715400,
    "text": "you"
  },
  {
    "startMs": 716600,
    "text": "bring them in and take their feedback"
  },
  {
    "startMs": 718240,
    "text": "really seriously."
  },
  {
    "startMs": 719840,
    "text": "They're skeptic because they're seeing"
  },
  {
    "startMs": 721560,
    "text": "the the way you are lacking validation,"
  },
  {
    "startMs": 723960,
    "text": "where your tools fail. So, and their"
  },
  {
    "startMs": 727040,
    "text": "feedback is basically the road map of"
  },
  {
    "startMs": 728760,
    "text": "how to improve your agent"
  },
  {
    "startMs": 730720,
    "text": "interacting with the code base."
  },
  {
    "startMs": 732560,
    "text": "So, just make sure to bring them in"
  },
  {
    "startMs": 735080,
    "text": "rather than trying to"
  },
  {
    "startMs": 736960,
    "text": "um figure out how to make them use the"
  },
  {
    "startMs": 738880,
    "text": "AI. Just let's have them be in charge of"
  },
  {
    "startMs": 741680,
    "text": "the road map to"
  },
  {
    "startMs": 743280,
    "text": "make AI safe your organization, and they"
  },
  {
    "startMs": 745920,
    "text": "will come along once they see that"
  },
  {
    "startMs": 748120,
    "text": "that the improvement that they're making"
  },
  {
    "startMs": 750200,
    "text": "actually making their life better."
  },
  {
    "startMs": 753240,
    "text": "Um and as you can see, they'll not be"
  },
  {
    "startMs": 755440,
    "text": "shy about telling you what you need to"
  },
  {
    "startMs": 756880,
    "text": "fix. This is Latin hour sitting with a"
  },
  {
    "startMs": 759640,
    "text": "bunch of people, and"
  },
  {
    "startMs": 761400,
    "text": "this is the result of brainstorms."
  },
  {
    "startMs": 765320,
    "text": "Um another thing that's"
  },
  {
    "startMs": 767480,
    "text": "been really helpful with my team"
  },
  {
    "startMs": 769120,
    "text": "specifically, and we're working to adopt"
  },
  {
    "startMs": 771080,
    "text": "it"
  },
  {
    "startMs": 772000,
    "text": "in the broader organization as well, is"
  },
  {
    "startMs": 774720,
    "text": "to make sure that you have an"
  },
  {
    "startMs": 776360,
    "text": "attention-aware communication."
  },
  {
    "startMs": 778720,
    "text": "In the age of AI, human attention is a"
  },
  {
    "startMs": 780360,
    "text": "scarce resource. I think I've heard it"
  },
  {
    "startMs": 782000,
    "text": "for multiple talks, and a lot of people"
  },
  {
    "startMs": 783680,
    "text": "have"
  },
  {
    "startMs": 784600,
    "text": "have come to the same conclusion. You"
  },
  {
    "startMs": 786440,
    "text": "can't get more human attention. So,"
  },
  {
    "startMs": 788280,
    "text": "where you spend your time and what"
  },
  {
    "startMs": 789480,
    "text": "you're reading is really becomes really"
  },
  {
    "startMs": 791960,
    "text": "important."
  },
  {
    "startMs": 793360,
    "text": "Um so, since it's such a scarce"
  },
  {
    "startMs": 796160,
    "text": "resource,"
  },
  {
    "startMs": 797440,
    "text": "marking what was generated by AI versus"
  },
  {
    "startMs": 799960,
    "text": "what was written by human is really"
  },
  {
    "startMs": 801560,
    "text": "helpful to know how much time you need"
  },
  {
    "startMs": 803440,
    "text": "to spend reading this, and how much slop"
  },
  {
    "startMs": 806080,
    "text": "can you expect in this part of the"
  },
  {
    "startMs": 807400,
    "text": "communication?"
  },
  {
    "startMs": 809040,
    "text": "Um"
  },
  {
    "startMs": 811240,
    "text": "and that can building a"
  },
  {
    "startMs": 813240,
    "text": "new culture around that"
  },
  {
    "startMs": 814440,
    "text": "self-communication that really helps. Um"
  },
  {
    "startMs": 817800,
    "text": "So, for example,"
  },
  {
    "startMs": 819760,
    "text": "um the team that team that I work with,"
  },
  {
    "startMs": 821800,
    "text": "we've decided we always every PR"
  },
  {
    "startMs": 824280,
    "text": "description will start with something"
  },
  {
    "startMs": 825440,
    "text": "like that, something that I wrote by"
  },
  {
    "startMs": 827200,
    "text": "hand. It could be very short that I"
  },
  {
    "startMs": 828880,
    "text": "describe what this is in code and what"
  },
  {
    "startMs": 830680,
    "text": "this is doing. And then the AI"
  },
  {
    "startMs": 832520,
    "text": "description is going to come after that,"
  },
  {
    "startMs": 834160,
    "text": "which is I will probably read it. I will"
  },
  {
    "startMs": 836160,
    "text": "probably edit it to remove uh some wrong"
  },
  {
    "startMs": 838800,
    "text": "things, but they didn't write every line"
  },
  {
    "startMs": 840720,
    "text": "here, so they should be more suspicious"
  },
  {
    "startMs": 842840,
    "text": "and they should pay more attention to"
  },
  {
    "startMs": 844200,
    "text": "what I wrote in the top and they should"
  },
  {
    "startMs": 845560,
    "text": "override it. Things like that in Slack,"
  },
  {
    "startMs": 848240,
    "text": "in email, it's like leaning into the"
  },
  {
    "startMs": 851000,
    "text": "fact that everybody knows that you're"
  },
  {
    "startMs": 852720,
    "text": "using AI to to craft your communication,"
  },
  {
    "startMs": 856400,
    "text": "but just let them share about it, tell"
  },
  {
    "startMs": 857760,
    "text": "them what they should read and what you"
  },
  {
    "startMs": 860000,
    "text": "they should pay less attention to."
  },
  {
    "startMs": 862520,
    "text": "And I remember early on, maybe"
  },
  {
    "startMs": 865880,
    "text": "like earlier in this year, I tried to I"
  },
  {
    "startMs": 870200,
    "text": "had some senior engineers in our org"
  },
  {
    "startMs": 872600,
    "text": "that had kind of were very much AI"
  },
  {
    "startMs": 875440,
    "text": "skeptic and I tried to reach out to them"
  },
  {
    "startMs": 877240,
    "text": "to see what was the problem, what was"
  },
  {
    "startMs": 878640,
    "text": "going on. I said, \"I tried to run an"
  },
  {
    "startMs": 880640,
    "text": "analysis on some of the PR comments that"
  },
  {
    "startMs": 882720,
    "text": "you've run.\" And obviously"
  },
  {
    "startMs": 884640,
    "text": "I used the AI to do that."
  },
  {
    "startMs": 886640,
    "text": "And then I didn't distinguish very"
  },
  {
    "startMs": 888480,
    "text": "clearly what I wrote versus what they"
  },
  {
    "startMs": 891920,
    "text": "what AI generated. And they got very"
  },
  {
    "startMs": 894240,
    "text": "upset. They're like, \"Why is sending I"
  },
  {
    "startMs": 896000,
    "text": "did not expect somebody um that I"
  },
  {
    "startMs": 898839,
    "text": "respect this much to send me"
  },
  {
    "startMs": 901400,
    "text": "something that's clearly this sloppy.\""
  },
  {
    "startMs": 902960,
    "text": "And then like I I took immediately like"
  },
  {
    "startMs": 905720,
    "text": "I apologize. I realize I should have"
  },
  {
    "startMs": 907000,
    "text": "marked it clearly and marked my"
  },
  {
    "startMs": 908160,
    "text": "intention like this is what I wrote."
  },
  {
    "startMs": 910640,
    "text": "This is what the AI wrote and I need"
  },
  {
    "startMs": 911839,
    "text": "your feedback on that because I don't"
  },
  {
    "startMs": 913200,
    "text": "have the context to know if it is sloppy"
  },
  {
    "startMs": 914920,
    "text": "or not and that's what I'm asking you"
  },
  {
    "startMs": 916520,
    "text": "for, so lesson like that and change the"
  },
  {
    "startMs": 918960,
    "text": "culture is just as important as some the"
  },
  {
    "startMs": 921640,
    "text": "engineering challenges that we've been"
  },
  {
    "startMs": 924280,
    "text": "facing."
  },
  {
    "startMs": 928160,
    "text": "Um another thing that's really helpful"
  },
  {
    "startMs": 930600,
    "text": "around the adoption is"
  },
  {
    "startMs": 934160,
    "text": "um as you progress through adoption,"
  },
  {
    "startMs": 936040,
    "text": "there's a lot of very fancy tools and a"
  },
  {
    "startMs": 937840,
    "text": "lot of very fancy workflow that we've"
  },
  {
    "startMs": 940520,
    "text": "we've been implementing, but one of the"
  },
  {
    "startMs": 942480,
    "text": "really effective thing is just letting"
  },
  {
    "startMs": 945480,
    "text": "people use the AI where they're at. So,"
  },
  {
    "startMs": 948560,
    "text": "uh it help it really helps normalize"
  },
  {
    "startMs": 951440,
    "text": "uh the use of AI for everyday tasks and"
  },
  {
    "startMs": 953400,
    "text": "it helps reduce the friction."
  },
  {
    "startMs": 955480,
    "text": "And really one of the most powerful"
  },
  {
    "startMs": 956920,
    "text": "thing is being able to tag an agent in"
  },
  {
    "startMs": 959480,
    "text": "the Slack message with somebody and they"
  },
  {
    "startMs": 961440,
    "text": "can you just do this for me?"
  },
  {
    "startMs": 963440,
    "text": "And have the agents to close the loop in"
  },
  {
    "startMs": 965120,
    "text": "the thread."
  },
  {
    "startMs": 966480,
    "text": "Um that that's kind of thing is really"
  },
  {
    "startMs": 969120,
    "text": "powerful. And then you can go on top of"
  },
  {
    "startMs": 971160,
    "text": "that and have all this thing automated"
  },
  {
    "startMs": 973320,
    "text": "and do all kind of fancy things, but if"
  },
  {
    "startMs": 975320,
    "text": "you have a new conversation with"
  },
  {
    "startMs": 976560,
    "text": "somebody who's not fully bought in and"
  },
  {
    "startMs": 978280,
    "text": "then you can tag it in a non like"
  },
  {
    "startMs": 981640,
    "text": "non-passive-aggressive way. You can tag"
  },
  {
    "startMs": 983960,
    "text": "it and say, \"Let's try to see if the"
  },
  {
    "startMs": 985720,
    "text": "agent can get it this time.\""
  },
  {
    "startMs": 987600,
    "text": "And they close the loop and if it's a"
  },
  {
    "startMs": 988960,
    "text": "good experience, that really helps"
  },
  {
    "startMs": 990320,
    "text": "people try it out on their own"
  },
  {
    "startMs": 992839,
    "text": "in other cases."
  },
  {
    "startMs": 996240,
    "text": "And our journey continues. We're still"
  },
  {
    "startMs": 998720,
    "text": "learning even though we're shipping AI"
  },
  {
    "startMs": 1000720,
    "text": "externally, our AI adoption"
  },
  {
    "startMs": 1003520,
    "text": "um"
  },
  {
    "startMs": 1004480,
    "text": "we're experimenting with with so many"
  },
  {
    "startMs": 1006120,
    "text": "things all the time. Our automation"
  },
  {
    "startMs": 1008600,
    "text": "story is not"
  },
  {
    "startMs": 1010000,
    "text": "uh fully there yet. We're still trying"
  },
  {
    "startMs": 1012080,
    "text": "trying to figure out when we should use"
  },
  {
    "startMs": 1014200,
    "text": "how we can use cloud agent effectively"
  },
  {
    "startMs": 1016560,
    "text": "given all the dependencies we have for"
  },
  {
    "startMs": 1018080,
    "text": "some of our bell system."
  },
  {
    "startMs": 1020240,
    "text": "And so we are continuing to learn. It's"
  },
  {
    "startMs": 1022160,
    "text": "a culture shift, it's an engineering"
  },
  {
    "startMs": 1023880,
    "text": "shift and I don't know about you, but"
  },
  {
    "startMs": 1026560,
    "text": "for I've been I've been working in the"
  },
  {
    "startMs": 1028839,
    "text": "valley for the last 15 years and this is"
  },
  {
    "startMs": 1030720,
    "text": "the biggest change by orders of"
  },
  {
    "startMs": 1032480,
    "text": "magnitude of everything that I've seen"
  },
  {
    "startMs": 1034600,
    "text": "in term culture and technology."
  },
  {
    "startMs": 1037000,
    "text": "So, um we're all here together and we're"
  },
  {
    "startMs": 1039079,
    "text": "all figuring it out and that's that's"
  },
  {
    "startMs": 1041800,
    "text": "what I wanted to talk to you today."
  },
  {
    "startMs": 1043040,
    "text": "Thank you."
  },
  {
    "startMs": 1044736,
    "text": ">> [applause]"
  },
  {
    "startMs": 1060457,
    "text": "[music]"
  }
]
