Sabitlenmiş Tweet
Foundry → Mithril (@mithrilcompute): The AI Omnicloud.
Now generally available!
We’re redefining GPU cloud economics, workload flexibility, and ease-of-use—for the compound AI & agentic era. 🧵 (1/8)
English
Jared Quincy Davis
392 posts
Jared Quincy Davis
@jaredq_
Founder and CEO, Mithril (@mithrilcompute). Orchestrating Compute. Fmr Research Scientist @GoogleDeepMind, Deep Learning Team. CS PhD @Stanford ML, Systems
Palo Alto, CA | London, UK Katılım Temmuz 2022
934 Takip Edilen1.7K Takipçiler
We're also partnering with Nebius so any current or future Nebius customer can opt into Flexible Reservations on their Nebius compute, opting it into Mithril's preemptible pools.
English
Haha welcome to Mithril, Suhail! B200s, H100s, H200s, a100s, etc -- always available in our preemptible pools.
We've also released Flexible Reservations where reserved compute customers can release resources back to the preemptible pool when they aren't using them, disincentivizing hoarding!
Suhail@Suhail
Hallelujah I found something but I won’t reveal due to selfish gpu hoarding.
English
Super exciting!! Congratulations, @amanrsanger and team!
Aman Sanger@amanrsanger
Composer 2 marks the one-year anniversary of our large model training efforts. Since then, we've built an exceptionally talent-dense team of ~40 people with some of the best researchers and engineers from the labs, academia, industry, and more heterogeneous backgrounds. And we are exclusively focused on coding. We don't care about models that can respond to emails, do your tax returns, or be your friend. Every FLOP, token, parameter, and researcher is entirely dedicated to software engineering.
English
I’m joining @SpaceX and @xai with @JasonBud.
X is the company realizing science fiction - reusable rockets, humanoid robots, data centers in space, and more. Almost 10 years ago, I joined SpaceX as an intern on Dragon 2 crew displays. This was in the era of the first rocket landings on barges, long before the Dragon 2 restored human spaceflight to America or Starlink delivered internet from space.
Every day since then, I’ve thought about the next steps to land on the Moon - and to build a city on Mars, data centers in space, the brains behind robots, and beyond. There is no better place to build teams and products from the ground up with planetary scale resources.
If you’re looking to work on the hardest problems that lay a foundation for humanity’s future to the Moon, Mars, and beyond - DM me.
English
Jared Quincy Davis retweetledi
Press coverage starting to pick up for ACM CAIS 2026. For anyone tracking the agentic systems space -- demo submissions close this Friday and workshop acceptances go out March 23.
caisconf.org
knowledgespeak.com/news/acm-launc…
English
Jared Quincy Davis retweetledi
The CAIS 2026 demo track is open.
Show us the agent that coordinates across tools in ways we haven't seen. The serving infrastructure that makes compound AI practical at scale. The debugging tool that finally makes production AI observable.
Submissions close March 13.
English
Harnesses, multi-agent topologies, and other compound AI systems are a major vector of progress! Exciting work and expect to see much more progress attributable to harnesses going forwards.
Michael Truell@mntruell
We believe Cursor discovered a novel solution to Problem Six of the First Proof challenge, a set of math research problems that approximate the work of Stanford, MIT, Berkeley academics. Cursor's solution yields stronger results than the official, human-written solution. Notably, we used the same harness that built a browser from scratch a few weeks ago. It ran fully autonomously, without nudging or hints, for four days. This suggests that our technique for scaling agent coordination might generalize beyond coding.
English
Sounds like the fast mode experiment is going well...
Jared Quincy Davis@jaredq_
Excited to see Anthropic's fast mode experiment. More options is always good — and counterintuitively, even 6x cost for 2.5x speed can make applications cheaper. We ran an experiment at @mithrilcompute that shows why flexible compute economics are so powerful 👇
English
Nice, Noam! A great new addition to the Pareto frontier.
Noam Shazeer@NoamShazeer
📢Introducing Gemini 3.1 Flash-Lite, our fastest and most efficient model, built for high-volume workloads. It outperforms 2.5 Flash in reasoning, reliability, and scalability at a lower cost. This model also introduces thinking levels. You can adjust compute by complexity of the task, burning zero thinking overhead on high-volume tasks, while reasoning through the complex edge cases. Maximum intelligence, minimal latency. Read more: blog.google/innovation-and…
English
We’re building an LLM chip that delivers much higher throughput than any other chip while also achieving the lowest latency. We call it the MatX One.
The MatX One chip is based on a splittable systolic array, which has the energy and area efficiency that large systolic arrays are famous for, while also getting high utilization on smaller matrices with flexible shapes. The chip combines the low latency of SRAM-first designs with the long-context support of HBM. These elements, plus a fresh take on numerics, deliver higher throughput on LLMs than any announced system, while simultaneously matching the latency of SRAM-first designs. Higher throughput and lower latency give you smarter and faster models for your subscription dollar.
We’ve raised a $500M Series B to wrap up development and quickly scale manufacturing, with tapeout in under a year. The round was led by Jane Street, one of the most tech-savvy Wall Street firms, and Situational Awareness LP, whose founder @leopoldasch wrote the definitive memo on AGI. Participants include @sparkcapital, @danielgross and @natfriedman’s fund, @patrickc and @collision, @TriatomicCap, @HarpoonVentures, @karpathy, @dwarkesh_sp, and others. We’re also welcoming investors across the supply chain, including Marvell and Alchip.
@MikeGunter_ and I started MatX because we felt that the best chip for LLMs should be designed from first principles with a deep understanding of what LLMs need and how they will evolve. We are willing to give up on small-model performance, low-volume workloads, and even ease of programming to deliver on such a chip.
We’re now a 100-person team with people who think about everything from learning rate schedules, to Swing Modulo Scheduling, to guard/round/sticky bits, to blind-mated connections—all in the same building. If you’d like to help us architect, design, and deploy many generations of chips in large volume, consider joining us.
English
Jared Quincy Davis retweetledi
tl;dr SWE-bench Verified is heavily contaminated for all frontier models, and many of the problems are also broken. Time to move on to harder, uncontaminated coding evals.
Olivia Grace Watkins@OliviaGWatkins2
In the past 6 months we’ve seen a divergence between the game-changing experience of coding w new models and tiny SWE-bench Verified gains. llm-stats.com/benchmarks/swe… New analysis finds most remaining unsolved problems have unfair tests, and many models are heavily contaminated.
English
Congrats, @adityagrover_, @StefanoErmon , and your team!
Aditya Grover@adityagrover_
Mercury 2 is now live! 🚀 The fastest reasoning LLM built for production speed. ~1000 tokens/sec vs <200 tokens/sec for comparable models. What this enables: 🤖 Fast agents: fast iteration loops, no compounding delays 🎙️ Voice and Search AI: tight turn-taking, natural conversations under strict latency budgets 💻 Interactive code completions, editing, and design workflows
English
Jared Quincy Davis retweetledi
Computer use models shouldn't learn from screenshots.
We built a new foundation model that learns from video like humans do. FDM-1 can construct a gear in Blender, find software bugs, and even drive a real car through San Francisco using arrow keys.
GIF
English
Every Mac sold is capex. It just might not be obvious yet.
staysaasy@staysaasy
Think different
English
@OriolVinyalsML Welcome back, Oriol! That's quite a "drastic" move. See you here.
English
Personal update: After an amazing 10 years in London, it's time for a major change. One-way ticket back to California 🌞! I'm incredibly excited to return to the Bay Area to continue building Gemini and pushing us toward the age of AGI 🚀
English
Jared Quincy Davis retweetledi
A physics textbook says certain particle interactions can't happen.
GPT-5.2 said "what if they can — under these specific conditions?"
Then it conjectured a formula. Then it proved it.
12 hours of reasoning. One new result in theoretical physics.
The preprint has IAS, Harvard, Cambridge, Vanderbilt authors alongside OpenAI. The AI wasn't just a tool — it's listed as having contributed the key conjecture.
This feels like a phase change.
OpenAI@OpenAI
GPT-5.2 derived a new result in theoretical physics. We’re releasing the result in a preprint with researchers from @the_IAS, @VanderbiltU, @Cambridge_Uni, and @Harvard. It shows that a gluon interaction many physicists expected would not occur can arise under specific conditions. openai.com/index/new-resu…
English
@joon_s_pk @karpathy @drfeifei @adamdangelo @rauchg @scottbelsky Congratulations, @joon_s_pk and team!
English
Introducing Simile.
Simulating human behavior is one of the most consequential and technically difficult problems of our time.
We raised $100M from Index, Hanabi, A* BCV, @karpathy @drfeifei @adamdangelo @rauchg @scottbelsky among others.
English
We've opened new pools for on-demand, SPOT, and self-serve reserve (arbitrary durations, from hours to weeks) NVIDIA B200 GPUs on Mithril.
In general, these chips are hard to get access to, so we hope this helps!
Spot floor at $0.01 for long-running and flexible jobs.
Blackwells are really nice to work with. Having all the extra HBM is super convenient.
English