Your curated collection of saved posts and media
Follow-up on non-English token-inefficiency with more model-language pairs: - Chinese is cheaper than English on major Chinese models - Gemini and Qwen provide least non-English tax - Anthropic has the highest tax by far; Kimi is next - Hindi is the worst-covered language here, despite its massive speaker base
The author achieved comparable performance with >10x less compute in this paper, but he needs compute to scale up this protein model desperately. Plz someone give him compute. https://t.co/cHPmPLkp3E
Not sure why there's a Graphic Content warning on the trailer, but here are some additional stills from behind-the-scenes. @elonmusk actually brought his own @Grimezsz poster to the set. https://t.co/zNeUDzo0O0
@elonmusk @Grimezsz The diner scene with @ericweinstein took a full 10 hours to get right. Somebody (not Eric) kept flubbing his lines. https://t.co/uzic3zxAHG
@elonmusk @Grimezsz @ericweinstein Unfortunately @nvidia CEO Jensen Huang was able to shoot for a day with us, but his performance was 🤌 https://t.co/hjwGLKF38v
@elonmusk @Grimezsz @ericweinstein @nvidia The intense romantic scenes with @ElizabethHolmes were a real highlight of the production! https://t.co/Wv0UOp9g2a
@elonmusk @Grimezsz @ericweinstein @nvidia @ElizabethHolmes At first @neiltyson passed on the script — said it was "too action-packed." But the second we signed Stephen Hawking, his agent called back insisting he was born to play the role! https://t.co/bATQLBLwxb
@elonmusk @Grimezsz @ericweinstein @nvidia @ElizabethHolmes @neiltyson In fact, @NYCMayor wasn't even officially cast in the film, but just happened to be strolling by when we were shooting that day and agreed to a quick cameo! https://t.co/8PmJzR6t3l
@elonmusk @Grimezsz @ericweinstein @nvidia @ElizabethHolmes @neiltyson @NYCMayor The director insisted on practical effects ONLY for the Wormhole Aliens. This was a huge pain in the producer's ass, but the results speak for themselves! https://t.co/eFnyZNFt51
@elonmusk @Grimezsz @ericweinstein @nvidia @ElizabethHolmes @neiltyson @NYCMayor The infamous sauna scenes with @BillGates were a real crew bonding experience in the sense that we all silently agreed never to speak of them again https://t.co/tRHbb47Pwa
🚨 🚨 🚨 Neil deGrasse Tyson reacts to AI or Die’s PI HARD 🚨 🚨 🚨 👇 https://t.co/M1jNoGEcwD
new site, same nonsense. https://t.co/RyaLLNSpkg
In the “Pi-Hard” movie trailer (https://t.co/3V3ZmHOn6W), is that the best Alien AI could come up with? A hairy lizard-monster dropped from a vortex in the sky over my City? That scene stretched credulity. Everybody knows Reptiles have no hair. https://t.co/YmPC2bNv3U
We’ve developed our own inference engine Runtime-Optimized Serving Engine (ROSE) to serve models ranging from embeddings to trillion-parameter LLMs. With CuTeDSL integrated into our inference engine, Perplexity can build the specialized GPU kernels faster to bring models up to peak performance on NVIDIA Hopper and Blackwell GPUs.
Personal Computer is now available to all users in a new Perplexity Mac app. Personal Computer is an advanced version of Perplexity Computer. It operates on any Mac, running tasks across your local files, native Mac apps, the web, and Perplexity’s secure servers. https://t.co/M1QMZJuOyJ
Perplexity Computer + realtime OHLCV data from stock exchanges (included with Computer) + Slack integration https://t.co/GSLOnUXRkp
spent the afternoon building an app with Kiro+Bedrock I am in disbelief... https://t.co/hChcuafjjo
not a single line of code written i never used most of the tools involved before this week https://t.co/j9eQcF9Jyi
@velvetmilkman0 @blue_clarity https://t.co/Mnn6fUU6XC
POV you’re my wife cracking the bathroom door open after I’ve texted you “toilet paper” https://t.co/2hqgbLUAHN
@zacglover @AlsieLC @nikitabier Works for me 😆 https://t.co/Q9oW0ZaLQ6
@jarrodwatts btw here’s mine https://t.co/2GWZjwfHTH
This GPT Image 2 prompt is going insanely viral right now. “Redraw the attached image in the most clumsy, scribbly, and utterly pathetic way possible. Use a white background, and make it look like it was drawn in MS Paint with a mouse. It should be vaguely similar but also not really, kind of matching but also off in a confusing, awkward way, with that low-quality pixel-by-pixel feel that really emphasizes how ridiculously bad it is. Actually, you know what, whatever, just draw it however you want.”
Fuck your water resources More datacenters https://t.co/0W5hHOBXzf
This GPT Image 2 prompt is going insanely viral right now. “Redraw the attached image in the most clumsy, scribbly, and utterly pathetic way possible. Use a white background, and make it look like it was drawn in MS Paint with a mouse. It should be vaguely similar but also not r
Continuing our IMO-gold journey, I’m delighted to share our #EMNLP2025 paper “Towards Robust Mathematical Reasoning”, which tells some of the key stories behind the success of our advanced Gemini #DeepThink at this year IMO. Finding the right north-star metrics was highly critical for our IMO effort and we did it with #IMOBench, a suite of advanced reasoning benchmarks for foundation models. More importantly, we encourage the community to go beyond short answers and showed that automatic grading of long-form answers is promising! Read on to see our project page, paper, and datasets in the thread 🙂
Very excited to share that an advanced version of Gemini Deep Think is the first to have achieved gold-medal level in the International Mathematical Olympiad! 🏆, solving five out of six problems perfectly, as verified by the IMO organizers! It’s been a wild run to lead this effor
Congrats to the whole Deep Think team from @GoogleDeepMind for this amazing milestone of #DeepThink V2 launch! Such a great a model that powers so many state-of-the-art results from reasoning (ARC-AGI2) to deep knowledge (Humanity's Last Exam), multimodality (MMMU-Pro), coding (Codeforces), the math research agent #Aletheia, and scientific discovery (that we shared just yesterday)! Blog: https://t.co/TB96WHPFed It has been a privilege witnessing the relentless progress 🔥: * ChatGPT -> Bard announcement (Mar 2023): 100 days * Announcement of IMO-gold achievement -> DeepThink v1 launch (Jul 2025): 10 days * Announcement of Aletheia agent & advancements in scientific research -> Deep Think v2 launch (Feb 2026): 1 day More to come! Stay tuned!
Yes, we provided 3 things for AI-assisted math: * Human-AI interaction (HAI) card (photo), inspired by model cards * Full transcripts https://t.co/NvO8p4Wiva * A label for novelty-autonomy, inspired by SAE Levels of autonomy, see #Aletheia paper https://t.co/8pLHmZZQO4 https://t.co/cFGZ6dbiWK
Really good question (note that DeepMind shared transcripts in their recent Aletheia paper, and I think this is clearly best practice). Hopefully OAI follows suit.
🤔Going to be in Vancouver tomorrow? 🧐Curious about whether or not scaling AI is a bad bet? 😎Catch TPN founder @zacharykarabell's chat with @GaryMarcus at @WebSummit at 10:20am. 🎟️Get your tickets here: https://t.co/ibej6zkToc https://t.co/g3q0AexRFx
In 2024 @sequoia said that the AI industry needed $600bn of revenue for the maths to make sense. Strangely they haven't updated it since then, so here's an attempt. $600bn is now more like $1.6trn.... @DavidCahn6 https://t.co/SxxKvkBhbQ
Michael Burry urged investors to scale back exposure to surging technology stocks, saying the current market environment has reached historically dangerous extremes reminiscent of prior speculative bubbles. The famed investor, best known for predicting the 2008 housing collapse, said investors should “reject greed” as enthusiasm around artificial intelligence and momentum-driven trades pushes valuations sharply higher. Read more: https://t.co/IEuUAbU3td
Which has better odds? Generative AI earning $1.6 trillion/year or the roulette wheel landing on zero? https://t.co/52cTL3iSHM
One estimate of how much annual revenue AI needs to “make sense”: 1.6 trillion. That’s four times what Google made in its best year. (total revenue so far is perhaps on order of 100 billion.)
Anthropic President Daniela Amodei: The risk no one says out loud is that this entire AI industry is betting on the future Compute is scarce and has to be bought years in advance, before companies know whether revenue will ever catch up "revenue has been historic, but that can change"