Julia Turc (@juliarturc) on X (original) (raw)

Explaining AI on YouTube • YC S24 Founder • Ex-Google Research • Eastern-European nihilist & American optimist

• 
  My computer science college professors used to look down on programming -- it's too volatile to deserve formal education (new paradigms and languages come out all the time). Instead, we did all kinds of theory: types, probability, queuing. Back then, it felt anachronistic and
• 
  Dear “15-18 yo founder”s sending me DMs, don’t. Go and hug your parents, fall in love, eat chocolate cereal for breakfast, read poetry. Nobody will give you back these years. And sure, do your homework and learn math and code if that feels fun. But stop building SaaS and
• 
  We need an official name for not-vibe-coding. My top candidates so far: · boomer coding · chewgy coding · trad coding · Coding with capital C Suggestions are welcome.
• 
  This baby bird deserves a hug and someone to tell him that he’s good enough. What. Are. We. Doing.
• 
  Everybody I know in SF works 12h a day and has no life already. What exactly does a lock in mean in this context?
• 
  Hot take: teaching linear algebra in isolation should be illegal. Determinants, traces, eigenvectors mean nothing without real-world context. 20 years later, I still have to google this s**** every time I encounter it, because whoever taught me this was drier than drywall.
• 
  "The user is absolutely right, I made a mess out of the codebase with my negligent changes. I am deeply sorry and am shutting myself down."
• 
  New sport in Silicon Valley: > meet VC at social event > have a thoughtful conversation about tech & AI > vc: so what are you building? > me: I make videos on YouTube > vc: looks down in disgust and suddenly needs to go to the restroom As a baseline, saying something like Google
• 
  Codex is so much better than Claude Code at figuring out when I want it to change code vs just talk about it. No more "DON'T YOU DARE CHANGE MY CODE AGAIN" all-caps rage.
• 
  Efficient BERT models from Google Research, now available at github.com/google-researc…! We hope our 24 BERT models with fewer layers and/or hidden sizes will enable research in resource-constrained institutions and encourage building more compact models.arxiv.org/abs/1908.08962
• 
  This is so painfully relatable. I find near-zero pleasure in coding anymore. You're trapped. Write by hand and know you're wasting time overall. Use AI and be crushed by wait times and context switch.
• 
  Replying to @waincarver
  Agreed, don't smoke hash. That's very different from "linear algebra and probability theory and philosophy and literature". Once you have that foundation, SaaS and vibe coding will take exactly 2-3 hours to pick up.
• 
  Replying to @vitaliidodonov and @vitddnv
  It's good for you, but not for them. The same way they're good for YC, but YC is not good for them.
• 
  New from Google Research: arxiv.org/abs/2106.16171To build multilingual NLP systems, a successful recipe is to pre-train on a multilingual corpus, and then fine-tune on labeled data in a single transfer language -- usually English. But is English best?