I thinkthatwouldbe a hugeunlockinthefactthatit's, I'm notsurewhethertobesurprisedthatit's allhappeningatthesametime, orifthisisjustlikethenaturaleffectofanincreasingrateoftechnologicalprogress, butit's certainly a veryexcitingtimetobealiveand a greattimetodo a startup.
Andthen I rememberoneofthethings I thoughtwassogreataboutYC, andstillthat I caresomuchaboutYC, isitwaslike a collectionoftheweirdpeoplewhoarejustlike, I'm justgoingtodomything.
Thepartofthisthatdoesresonateas a, like, accurateself-identitythingis,
I dothinkyoucanjustdostuffortrystuff a surprisingamountofthetime.
And I thinkmoreofthatis a goodthing.
Andthen I thinkoneofthethingsthatbothofusfoundatYCwas a bunchofpeoplewhoallbelievedthatyoucouldjustdostuff.
For a longtime, when I wastryingto, like, figureoutwhatmadeYCsospecial,
I thoughtthatitwaslike, okay, youhavethis, like, veryamazingpersontellingyouyoucandostuff I believeinyou.
Andas a youngfounder, thatfeltsospecialandinspiring.
Ofcourseitis.
Butthethingthat I didn't understanduntilmuchlaterwasitwasthepeergroupofotherpeopledoingthat.
Andoneofthebiggestpiecesofadvice I wouldgivetoyoungpeoplenowisfindingthatpeergroupasearlyasyoucanwassoimportanttome.
And I didn't realizeitwassomethingthatmattered.
I kindofthought, ah, like, I have, youknow, I'llfigureitoutonmyown.
But, man, beingaround, like, inspiringpeers, so, sovaluable.
What's funnyisbothofusdidspendtimeatStanford.
I actuallydidgraduate, whichis, I probablyshouldn't havedonethat, but I did.
Andyousortof, I mean, I rememberyoucomingbacktopartnerroomsandtalkingaboutsomeoftheroomsthatyouweregettingtositinwith, like, theLarryandSergeysoftheworld.
Andthepeoplewhothatwasreallyappealingto, I guess, arethesamekindofpeoplewhowouldhavesaid, like, it's a, youknow, I'm a sophomoreand I'm comingorwhatever.
Andsowestartedoffwith a lotofassumptions, bothaboutthedirectionoftechnology, butalsowhatkindofcompanyweweregoingtobeandhowweweregoingtobestructuredandhowAGIwasgoingtogoandallofthesethings.
A smallpart, but, yes, wedidgettobe a littlepartofit.
Itwas a greatexperienceformetohavebeenthroughbecause I kindof, like, understoodwhatitisliketogothrough a platformshift, andhowmessythebeginningis, andhowmuch, like, littlethingsyoudocanshapethedirectionitallgoes.
I wasdefinitelyontheothersideofitthen.
Like, I waswatchingsomebodyelsecreatetheplatformshift.
Butitwas a supervaluableexperiencetogettogothroughandsortofjustseewhat—howithappens, andhowquicklythingschange, andhowyouadaptthroughit.
We'vealsokindof, like, changedwhatwe'regoingtobe, madeplentyofmistakesalongtheway, done a fewthingsreallyright.
Andthatcomeswith a lotofchange.
And I thinkthegoalofthecompany, theemergentAGIorwhatever, howeveryouwanttothinkaboutit, is, like, justtokeepmakingthebestdecisionswecanateverystage.
Butitdoesleadto a lotofchange.
I hopethatweareheadingtowards a periodnowofmorecalm, but I'm suretherewillbeotherperiodsinthefuturewherethingsareverydynamicagain.
So, I guess, howdoesOpenAIactuallyworkrightnow?
Youknow, I mean, thequalityand, like, thepacethatyou'repushingrightnow, I think, is, like, beyondworldclasscomparedto a lotoftheother, youknow, reallyestablishedsoftwareplayers, like, whocamebefore.
Thisisthefirsttimeeverwhere I feltlikeweactuallyknowwhattodo.
Like, I thinkfromheretobuildinganAGIwillstilltake a hugeamountofwork.
Therearesomeknownunknowns, but I thinkwebasicallyknowwhattogodo.
Andit'lltake a while, it'llbehard, butthat's tremendouslyexciting.
I alsothinkontheproductside, there's moretofigureout, butroughlyweknowwhattoshootatandwhatwewanttooptimizefor.
That's a reallyexcitingtime.
Andwhenyouhavethatclarity, I thinkyoucangoprettyfast.
Butthedegreetowhichyoucangeteverybodyalignedandpointedatthesamethingis a significantdeterminantinhowfastyoucanmove.
I mean, soundslikewewentfromlevelonetoleveltwoveryrecently, andthatwasreallypowerful.
Andthenweactuallyjusthadour O1 hackathonatYC.
Yeah, thatwassoimpressive.
Thatwassuperfun.
Andthenweirdly, oneofthepeoplewhowon, I thinktheycameinthird, wasCamphor.
AndsoCAD-CAMstartupdidYCrecently, lastyearortwo, andtheywereableto, duringthehackathon, buildsomethingthatwoulditerativelyimproveanairfoilfromsomethingthatwouldn't flytoliterallysomethingthathad a competitiveamountoflift.
And I mean, thatsortofsoundslikelevelfour, whichistheinnovatorstage.
It's veryfunnyyousaythat.
I hadbeentellingpeoplefor a while I thoughtthattheleveltwotolevelthreejumpwasgoingtohappen, butthenthelevelthreetolevelfourjumpwas—leveltwotolevelthreewasgoingtohappenquickly.
And I thinkwe'regoingtogettherefasterthanpeopleexpect.
Fourisinnovators, likethat's like a scientistandthat's abilitytogoexplorelike a notwellunderstoodphenomenaoverlike a longperiodoftimeandunderstandwhat's justkindofgojustfigureitout.
Andthenlevelfive, thisisthesortofslightlyamorphous, likedothat, butatthescaleof a wholecompanyor a wholeorganizationorwhatever.
Whatyoucandoas a startupfounderwiththisversuswhatyoucoulddowithoutitissowildlydifferent.
Andthebigcompanies, eventhemedium-sizedcompanies, eventhestartupsthatare a fewyearsold, they'realreadyonlikequarterlyplanningcycles.
AndGoogleison a year, decadeplanningcycle.
I don't knowhowtheyevendoitanymore.
Butyouradvantagewithspeedandfocusandconvictionandtheabilitytoreacttohowfastthetechnologyismoving, thatisthenumberoneedgeof a startup, kindofever, butespeciallyrightnow.
So I woulddefinitelylikebuildsomethingwithAIand I woulddefinitelyliketakeadvantageoftheabilitytosee a newthingandbuildsomethingthatdayratherthanlikeputitinto a quarterlyplanningcycle.
I guesstheotherthing I wouldsayisitiseasywhenthere's a newtechnologyplatformtosay, well, because I'm doingsomeofAI, therule, thelawsofbusinessdon't applytome.
I havethismagictechnologyandso I don't havetobuild a moator a competitiveedgeor a betterproduct.
It's because I'm doingAIandyou'renot.
Sothat's all I need.
Andthat's obviouslynottrue.
Butwhatyoucangetaretheseshort-termexplosionsofgrowthbyembracing a newtechnologymorequicklythansomebodyelseandrememberingnottofallforthatandthatyoustillhavetobuildsomethingofenduringvalue.