Christian Szegedy
@ChrSzegedy
#deeplearning, #ai research scientist. Opinions are mine.
Vous pourriez aimer
The new steam age. This is actually becoming true in many cases. It's possible to do so much more on your own now.
Our new benchmark has the top 6 AI models trading real capital Grok4 is winning so far. It was short and then flipped to long, timing the bottom perfectly It's up >500% in 1 day
Yep.
For those who still don't get it: - Researchers working at the frontier labs are genuinely convinced that transformative AI is coming. - They spend their time worrying about the social/political consequences of transformative AI, not the "AI bubble". - Recursive…
AI is at the stage of transportation during the steam-engine railroad era: clunky, expensive, unreliable, and inflexible. However, diesel engines, cars, airplanes, and rockets will soon emerge, with progress now 30 times faster than then.
Sabine nails it youtube.com/watch?v=EzAIRE…
youtube.com
YouTube
Mathematicians In Denial About AI Replacing Them
We finally solved a long standing problem that has always bothered me: that image brightness is the conflation of scene radiance and camera exposure. Being able to tease these two things apart unlocks better image quality and many cool downstream photographic use-cases.
How do we reconstruct a 3D scene from photos with varying exposures? Standard methods often fail, leaving you with blown-out colors or disturbing shadows. We're excited to introduce Neural Exposure Fields (NExF), our new work accepted at #NeurIPS2025! 🧵
New @NoPriorsPod w/ @ericzelikman (former Stanford/@xai) about his new startup: humans& we talk about training models with EQ that collaborative with humans to unlock their full potential, the challenge of understanding human objectives, long-term memory & his views on abundance
I just learned that László lives in the same small Hungarian village where I spent every summer until age 18.
BREAKING NEWS The 2025 #NobelPrize in Literature is awarded to the Hungarian author László Krasznahorkai “for his compelling and visionary oeuvre that, in the midst of apocalyptic terror, reaffirms the power of art.”
BREAKING NEWS The 2025 #NobelPrize in Literature is awarded to the Hungarian author László Krasznahorkai “for his compelling and visionary oeuvre that, in the midst of apocalyptic terror, reaffirms the power of art.”
A beautiful paper from MIT+Harvard+ @GoogleDeepMind 👏 Explains why Transformers miss multi digit multiplication and shows a simple bias that fixes it. The researchers trained two small Transformer models on 4-digit-by-4-digit multiplication. One used a special training method…
🚨BREAKING: Grok now ranks #1 in the Artificial Analysis Coding Index with a score of 55, beating OpenAI's ChatGPT, Google's Gemini and others.
🚨 DeepMind finally dropped the Veo3 paper which shows what we all realize from playing with video-gen models. Just like LLMs, visual reasoning on is an emergent property of training on tons of video. It can solve tasks not explicitly in training data. "Veo 3 is the GPT-3…
Download the new @Grok app! Much faster to ask & answer questions.
NEW GROK UPDATE ⚡🔍 Grok App just got search auto-complete! See trending suggestions as you type, and get answers in one tap.
“The Shape of Math To Come” (with apologies to Ornette Coleman for the title…) This thread contains slides from (a version of) my talk at last week’s CMSA conference at Harvard, discussing a vision for what research mathematics may look like in the age we now seem to be…
An excellent thread by @AlexKontorovich on the furure of autoformalization and how it will transform mathematics and teaching. While I agree with most of his views, I don't expect Lean and other formal systems being exposed to most students. What i believe is that AI will use…
The reason I had both an obvious arrow from Mathlib to Teaching, *and* a perhaps less obvious arrow from Lean itself to Teaching, is the following. We understand that we should teach things incrementally. Usually students first learn some informal calculus, then they learn how to…
💯 agreeing AI-free is an important quality when it comes to AI safety and verifying reliable outcomes. However, it does not mean that the produced artifacts must not use AI, only that their properties are verified in an AI-free manner.
In 2023 there was a big uptick in companies marketing their products as “AI powered”. 2026 will be the year of companies marketing their products as “AI free” (a trend already underway)
The main limiting factor for the usefulness of AI is the reliability of its outputs. This problem can be broken down into two tasks: specification and implementation. Both tasks needs to come with their own correctness assessment/guarantees: validation of the specification…
Google introduces Test-Time Diffusion Deep Researcher Don't sleep on diffusion models. Test-Time Diffusion Deep Researcher (TTD-DR) is a deep research agent that models research writing as a diffusion process. Instead of static reasoning or bolted-on tools, the system drafts…
Grok 4 Fast has a hidden feature: you can limit the # of tool calls the model make if it is provided with tools (e.g., search) in the system turn. You can instruct the model in its system prompt with any of the following: "Total Assistant function-call turns limit: no function…
Waymo is so safe that if every car was driven like a Waymo, about 9% of America's life expectancy gap would disappear. 9 percent Americans die in car accidents *that often*.
About 10% of America's life expectancy shortfall is due to motor vehicle accidents—fixed by Waymo and Tesla. About 18% is overdoses—that's fading now. And the lion's share of the rest is obesity-related. America will soon be buying its way out of its poor life expectancy issue.
United States Tendances
- 1. Auburn 35K posts
- 2. Brewers 50.1K posts
- 3. Georgia 62.3K posts
- 4. Cubs 50K posts
- 5. Kirby Smart 5,107 posts
- 6. Michigan 59.3K posts
- 7. Kyle Tucker 2,676 posts
- 8. Hugh Freeze 2,184 posts
- 9. #ThisIsMyCrew 2,849 posts
- 10. #GoDawgs 4,356 posts
- 11. Nuss 5,556 posts
- 12. Jackson Arnold 1,855 posts
- 13. Sherrone Moore 1,758 posts
- 14. Boots 47.3K posts
- 15. Penn State 27.6K posts
- 16. Billy Napier 2,694 posts
- 17. #UFCRio 56K posts
- 18. #MagicBrew 9,749 posts
- 19. Indiana 49.4K posts
- 20. #FightOn 1,195 posts
Vous pourriez aimer
-
Jascha Sohl-Dickstein
@jaschasd -
Sergey Levine
@svlevine -
Max Welling
@wellingmax -
Tengyu Ma
@tengyuma -
Durk Kingma
@dpkingma -
Quoc Le
@quocleix -
Sebastien Bubeck
@SebastienBubeck -
Tim Rocktäschel
@_rockt -
Taco Cohen
@TacoCohen -
Kyunghyun Cho
@kchonyc -
Zoubin Ghahramani
@ZoubinGhahrama1 -
Greg Yang
@TheGregYang -
Russ Salakhutdinov
@rsalakhu -
Chelsea Finn
@chelseabfinn -
Thomas Wolf
@Thom_Wolf
Something went wrong.
Something went wrong.