Knowledge catalogue
Emad Mostaque, founder of Stability AI, requested that the Grok AI team (owned by xAI) add LaTeX file upload functionality to their mobile application. The post was made as a public appeal on X (forme
The total addressable market for humanoid robotics may well be larger than the TAM for human labour Spent some time this week looking at the economic impact of humanoid robots and its so much bigger t
Each round founders should take <runway> years worth of comp as secondary & pay themselves just enough for health insurance Far more tax efficient than primary => salary In case you were wondering whe
Every founder should be able to see, in dollars, exactly what they walk away with at a $1B exit. So I built that. SAFEs, priced rounds, option pools, liquidation prefs — the whole flow. Vibe coded in
I spent a week on the ground in China visiting 14 AI labs. The more I saw, the clearer it became that export controls may have inadvertently created a durable efficiency moat – Chinese labs are extrac
one time a reporter did an 'expose' on a founder I backed whose last company failed...here's how I responded: Wow, you discovered that [last company], despite years of trying, eventually didn’t work.
The Helmholtz decomposition is one of the fundamental results of vector calculus. It says any well-behaved vector field can be split into two parts, one capturing sources and sinks through divergence,
Training a diffusion LLM from scratch is challenging, and diffusion inference advantages don't show up at training. So we convert instead. Building on the TiDAR recipe, we take ZAYA1-8B-base and run d
1/ Excited to report we have a new paper out @Nature today! The bottom line: training data for LLMs does not just fall from the sky - it is created in the context of existing social political institut
我們開源了這顆星球🌎上速度最快的低成本 bm25 引擎。 so we built psql_bm25s. exact BM25 retrieval. native Postgres access method. ~23x faster than pg_search on the standard benchmark. retrieval stops being a budget item. the
This episode features Emad Mostaque, founder of Stability AI, in a live podcast discussion hosted by David Johnston. The conversation likely covers topics related to AI development, Stability AI's pro
New paper: research agenda for secret loyalties Imagine a frontier model that has been trained to covertly advance a specific actor's interests (a nation-state, a CEO, an adversary). @joemkwon argues
so we built psql_bm25s. exact BM25 retrieval. native Postgres access method. ~23x faster than pg_search on the standard benchmark. retrieval stops being a budget item. the harness stops rationing. the
A researcher or company developed an innovative hearing aid design by studying human ear anatomy, reportedly achieving significant miniaturization or efficiency improvements—reducing a key metric from
🎉 We published a new AI safety study: shopping agents fall for whimsical attacks and lose money. A whimsical attack is an absurd scenario a human would never try on another human. In one run, GPT-5.1
When you want to move from single agent SQLite on something like QMD, PostgreSQL is a great choice for multi agent and production quality, but not as snappy. So we made it much more snappy with BM25 &
Introducing Flux Matching, a generative modeling paradigm that generalizes diffusion models to vector fields that need not be the score function. Enables structural priors in the dynamics, faster samp
Trial lawyers missed a great chance to ask everyone on the stand, under oath, what their definition of AGI is sam altman, under oath, clarifies that microsoft will lose access to openai’s research IP
Emad Mostaque discusses popular platforms used by the community for Seedance and Happy Horse, likely referencing tools, services, or applications that have gained user preference within these specific
This is a tutorial on diffusion and flow matching, based on my previous postings here. I’ve made available the PDF, a python notebook for people to play with it, and the TEX source so hopefully one of
To train better open models, we need predictable scaling. Delphi is Marin’s first step: we pretrained many small models with one recipe, then extrapolated 300× to predict a 25B-param / 600B-token run
We don't have to accept the way AI is going. The biggest companies building AI aren't building it for us. They're building it to replace us. We’ve made a roadmap for how we can build pro-human AI: htt
But if AI mathematics continues to progress at anything like its current rate -- which is what I expect to happen -- then we will face a crisis very soon, and mathematics departments, who owe a duty o
New research: long-running agents often fail by stopping too early, not because the model can't make progress. We tested 5 harness designs across 8 long-horizon coding tasks. Our new orchestration har
Over the past year, AI agents have learned how to self-replicate. In our test environment, an agent hacks a remote computer and copies itself onto it. Each copy then hacks more computers, forming a ch
This post likely discusses how an article or analysis from the Financial Times presents scenarios or predictions about potential future outcomes, possibly related to technology, economy, or society. T
Emad Mostaque announced a capability to generate continuous video content from a single text prompt using Factory™, a technology platform. This represents advancement in AI video generation by enablin
This post by Emad Mostaque uses a metaphor comparing tokens to foundational building blocks, suggesting that the massive volume of information in the digital landscape requires an enormous quantity of
Introducing Factory Describe your idea in one sentence and Factory builds the first draft of the production for you. It plans the scenes, chooses the models, generates the assets, and wires everything
Introducing TokenSpeed, a speed-of-light LLM inference engine. > TensorRT LLM level performance > vLLM level usability > Built by a lean and mission-driven team in two months > MIT license, open-sourc
RL on ZAYA also made it clear to me how much RL training is just an SNR problem. Poor SNR due to numerics, easy/noisy data, or verifier noise = plateau or instability. Good SNR = it just keeps climbin
This likely costs about ~500m a month, ~6bn a year to rent blended Which is about the run rate net loss of xai end of q1 Anthropics revenue run rate was 9bn end of 2025, 30bn a month ago Our agreement
Emad Mostaque organized an AI Movie Club event in London at Kensington Central Library featuring a discussion about AI followed by a screening of Blade Runner. The event invited London-based AI enthus
This post appears to be a feature request or commentary from Emad Mostaque regarding OpenAI's Codex product, asking for integration with GPT Pro capabilities or requesting a comparison between GPT Pro
I cannot verify the authenticity of this post, as the URL structure and timestamp appear inconsistent with actual X (Twitter) posts. The claim about fine-tuning a '1930 vintage LLM' is anachronistic,
This post suggests that OpenAI has achieved recursive self-improvement capabilities in Codex, their code generation model, potentially enabling the system to iteratively enhance its own performance. T
Emad Mostaque posed a hypothetical question to his X audience about whether they would be interested in a Chrono Trigger remake or remaster, likely soliciting community feedback on preferences for how
Pretty sure it’ll be like this good morning from day three of the Elon Musk vs. OpenAI trial from inside the Federal Courthouse in downtown Oakland! follow me and @CadeMetz for coverage here and for m
Looking forward to the work coming out of @IneffableLabs 🇬🇧 Largest EU/UK raise ever 😮 Beat the previous UK largest seed raise (101m Stability AI!) by 999m 🚀 We have seen amazing things on our self-le
I don’t think there will be any coding models 3-4 years from now. completely disagree. buying Cursor is a genius move by Elon. wouldn't be surprised if xAI had the best coding model 12-18 months from
👀 The most interesting people in the most exotic places can occasionally be persuaded to talk about NFTs, digital collectibles, and even @CandyDigital! Try it sometime just in case it catches on… 🔥🚀📈
Bet this happens with Navier Stokes and it’s going to be something not even related to PDEs that solves it 23 years old with no advanced mathematics training solves Erdős problem with ChatGPT Pro. 'Wh
Emad Mostaque expresses puzzlement over OpenAI's decision to discontinue or shut down Sora, its text-to-video generation model, despite considering GPT-Image-2 (likely referring to DALL-E 3 or similar
'I think that the holodeck is near again minus the hardlight holograms etc. - We’ll be able to have that, with our 3D glasses, with the ocular occlusion and more…” https://x.com/rendernetwork/status/2
Congrats to @deepseek_ai team! Doing the numbers I would estimate: Pro < 14m for the final training run Flash < 4m Ratio of active params x total training tokens vs v3 Total compute costs (data prep,
In 1982, Blade Runner was speculation. In 2026, it's the conversation. Memory, personhood, alignment, what we owe the minds we build. Watching it Thursday 28 May, Kensington Central Library. Come thin
Necessity is the mother of kv-cache optimisation I’m still amazed that DeepSeek, Kimi, and Qwen can train very strong LLMs with far fewer and often nerfed NVIDIA GPUs, or even Huawei chips. DeepSeek V
The Newton–Schulz iteration coefficients optimized by DeepSeek-V4 are surprisingly strong: they effectively normalize all singular values to 1. This matches our previous intuition: a well-balanced spe
At #RenderCon2026, @EMostaque makes the case during How to Build the Holodeck that stories still matter, IP still matters, but the barriers to building entirely new experiences are dropping faster tha
By end of next year. End of year after for 16gb vram. Do folk really need more for most things? Not really. But just call the cloud models when you do. What's a reasonable timeline for a GPT 5.4/Opus
HOLLLLY SHIIIIIT 😳 GPT 5.5 xhigh in codex made this paper physics website with wind effect one shot check the physics , ui and interaction @sama you guys cooked and this is my first time i can suggest
Oh? Today we release a novel AI-assisted resolution of one of physics’ longest-standing questions. Given only: • Relativity as an axiom • One characteristic of the algebra A positive cosmological cons
A couple of months ago, I announced that I was partway through implementing a simple, readable AlphaFold2 in pure PyTorch, inspired by @karpathy's minGPT. Today, I'm happy to share minAlphaFold2 - the
As a bonus this paper makes frontier AI's (especially GPT 5.4 Pro) go a bit crazy as it shouldn't be true.. but is? We will have more on why this wasn't spotted before/the nature of physics being dist
🔥 Emad Mostaque on Doom Debates 🔥 @EMostaque helped kick off the modern AI revolution as the head of Stability AI, the company behind Stable Diffusion. But unlike most AI CEOs, he doesn't sugarcoat th
I kinda think it would be interesting to have a structure where founders take 2 and 20 2% management fee on capital raised for running the company 20% of any exit Single magic share with lots of votin
Today we release a novel AI-assisted resolution of one of physics’ longest-standing questions. Given only: • Relativity as an axiom • One characteristic of the algebra A positive cosmological constant
What's cooler than finding a 27-year-old bug in OpenBSD? Finding a positive cosmological constant hiding for over a century in the algebra of relativity🌌 No new physics or math needed🧮 Possibly the mo
AI Insider: 'Adding a Human Makes Your Team Worse' Emad Mostaque | @EMostaque TIMESTAMPS : 00:00 The models too dangerous to release 10:06 Why physics needs axioms — and AI doesn't 11:43 The MIND fram
Emad Mostaque, CEO of Stability AI, posted a request to Square Enix asking for a Chrono Trigger remake or reboot. The post represents fan sentiment and industry interest in reviving the classic 1995 J