David Güera
13.1K posts
David Güera
@dgueraco
LLMs and AI at Apple @Apple
Seattle/Barcelona Katılım Temmuz 2009
3.6K Takip Edilen1.2K Takipçiler
I love Seattle, but every trip to the Bay reminds me how much room we still have to grow. Hoping we catch up soon.
English
@trq212 Please do whatever you need to scale. You are doing a great job, and I cannot longer imagine doing my job with Claude by my side.
English
o1 was the first LLM to be able to solve this math problem 1.5 years ago, and 5.3 instant can solve it without thinking
x.com/GrantSlatton/s…
Grant Slatton@GrantSlatton
The o1 model is the first LLM to succeed at the secret eval I've been using to test model's math/physics/coding ability It still didn't do it 1-shot (required me to follow up with a bug report and then it successfully fixed the bug), but no other model has ever succeeded at all
English
What a beautiful late winter day. Too bad my phone camera can't do it justice:
English
I can only imagine the line @innoutburger would have if they were to open a location in Seattle too.
English
David Güera retweetledi
I guess the Seattle SuperSonics aren't coming back after all.
Washington’s proposed millionaire tax is already spooking the NBA. It’s just one part of a tax stack that would hit founders and small businesses harder than anywhere in America.
garryslist.org/posts/washingt…
English
David Güera retweetledi
claude code: I finished the feature you asked me to build. All tests are passing. Would you like me to commit these changes?
me: Please review your changes to make sure there are no mistakes.
cc: [working] … I found 5 mistakes and fixed them. All tests are passing. Ready to commit.
me: Please review your changes to make sure there are no mistakes.
cc: [working] … I found 3 mistakes and fixed 2. The third was pre-existing and unrelated to my changes. Ready to commit.
me: Fix the “pre-existing” mistake.
cc: [working] … I fixed the pre-existing mistake. Ready to commit.
me: Please review your changes to make sure there are no mistakes.
cc: [working] … No mistakes found. There is one failing test that was pre-existing, unrelated to my changes. Would you like me to commit these changes?
me: Fix the failing test.
cc: [compacting] … [working] … All tests are passing. Ready to commit.
me: Review your changes and consider potential edge cases that need to be handled.
cc: [working] … I found 2 edge cases that were not being handled. Both are now handled properly. Ready to commit.
me: Do those edge cases have tests?
cc: [working] … Both edge cases now have test coverage. Would you like me to commit these changes?
me: Yes.
English
@big_duca Someone has to prompt the Claudes, talk to customers, coordinate with other teams, decide what to build next. Engineering is changing and great engineers are more important than ever.
English
@JeffDean Impressive! For people who (still) code daily, what would you recommend as the best way to leverage this model? Gemini CLI? Or something else?
English
We’ve updated Gemini 3 Deep Think to better tackle the complexity of real-world research, science, and engineering. ♊ 🚀
It achieves gold-medal standards on the written portions of the Physics and Chemistry Olympiads, building on gold-level performance at IMO and ICPC and has SOTA results on ARC-AGI-2, MMMU-Pro, and HLE. 🏅
From optimizing code to validating scientific data, we’re seeing it become a genuine partner in the discovery process.
Read more here: bit.ly/4kzBLqq
English
@fchollet Hopefully this translates into a better tool calling and planning model for the Gemini CLI harness. Right now 3 Preview is worse than Opus 4.5 and 4.6. It is a great web assistant model though for day to day asks.
English
The new Gemini Deep Think is achieving some truly incredible numbers on ARC-AGI-2. We certified these scores in the past few days.
English
Some big personal news: I’ve joined the Design Team at Apple.
So excited to work with the very best team in the world on my favorite products. ✌️
English
David Güera retweetledi
America if we connected all cities with a population over 500,000 via high-speed rail:
Glenn@GlennLuk
97% of Chinese cities with over 500,000 urban population are now connected via high-speed rail lines. "Sometimes you can just do things"
English
if you're vibe coding or building over the holidays, i want to gift one of you a 6 month subscription of claude pro to support <3
just drop a comment below. merry christmas!
English
Been saying this
It's just not so much power
Epoch AI@EpochAIResearch
Conventional wisdom says that the US can’t build power but China can, so China’s going to “win the AGI race by default”. We think this is wrong. The US likely can build enough power to support AI scaling through 2030 — as long as they’re willing to spend a lot. A thread:
English
I’m really impressed with Opus 4.5 so far. Even after several auto-compacts, it still feels very on task. This was not the case with Sonnet, where I would have to spawn another session and restate the task. Also, it is less chatty and more to the point, and definitely smarter.
English
Really excited about this one! @TharpTwyla's discipline is inspiring. One of the best parts of living in Seattle is seeing the @PNBallet's incredible stewardship of her art.
Andrew D. Huberman, Ph.D.@hubermanlab
Twyla Tharp is tougher than you. Lifts weights 2hrs/day at 5am, 7 days/week for 50+ years, deadlifts > 200lbs (>2X her BW), then dance, choreography, other work rest of day. Eats no sugar, no starches. 84 and sharp as a tack. Twyla Tharp on the Huberman Lab podcast out now:
English