Jake Taylor

Indeed, while the bones are there, CAISI needs three missing pieces (beyond congressional engagement): political support for the unpopular task of auditing, particular across executive branch lines; funding support for the challenge of both researching and doing evaluations; and ‘teeth’ for when audits come back with problems

@BlackHC AISI / CAISI does not set standards for firms, nor do they require audits. They also don't have a charter involving external participation from other parts of the ecosystem, including public interest auditors and labs. See our related proposal here: link.springer.com/article/10.100…

With #ICML closing out, the team has one last blog post covering our work on dynamic skill systems using memory. Take a look! axiomatic-ai.com/blog/skill-syn…

TIL that the Kafri & Taylor paper on classical forces is on the wikipedia article on quantum gravity #cite_note-12" target="_blank" rel="nofollow noopener">en.wikipedia.org/wiki/Quantum_g…

@g8ge @NIST @ORNLNeutrons @isisneutronmuon @psich_en @ILLGrenoble @neutron_org Long, tedious, and incredibly important job done well by the ncnr team

The @NIST Center for Neutron Research (NCNR) has resumed full-power reactor operations following an extended shutdown for maintenance and repair. Further information at #the-ncnr-resumed-reactor-operations-in-may-following-an-extensiv" target="_blank" rel="nofollow noopener">nist.gov/ncnr#the-ncnr-… @ORNLNeutrons @isisneutronmuon @psich_en @ILLGrenoble @neutron_org

@meggmcnulty Pls reach out to @LeopoldoSarra !

@quantum_jake would love to say hi!

Next week the team is at #ICML showcasing some of our recent successes in the space of formal methods. Two key pieces of infrastructure that we’re introducing: AxProverBase (a simple agent system that just works without bells or whistles) to baseline provers, and

NIST has announced an agreement with SRI International, a nonprofit research and development institution, to help advance U.S. quantum research, development and manufacturing capabilities. Learn more: nist.gov/news-events/ne…

…the formal intro of SorryDB for the lean enthusiast. Take a look: prover.axiomatic-ai.com

@S_Flammia Eg, my pizza cooks at twice room temp (in Kelvin). Rakine also works, but then I would have to rememorize meV to GHz to K

@S_Flammia On the Kelvin thing… it is pretty good for cooking tbh

1/5 I see among the AC debate that people are arguing about Fahrenheit vs. Celsius. My controversial take: Fahrenheit is actually better for day-to-day use. 🧵

That venerable work of the National Bureau of Standards has been significantly expanded and is free online as the @NIST Digital Library of Mathematical Functions dlmf.nist.gov

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…

Seeing really interesting capabilities on the local model side. Also Claude Fable/Mythos takedown is going to be the most important trial balloon of the year.

@yujitach 面白いことと想いますけど、leanがどんなぐらい出来たかわからないね

こちらは Harvard の知人の Xi Yin が準備中の場の量子論の教科書(?)のようです: github.com/xiyin137/QFT pdf はおいてないので、clone して manuscript/tex/main.tex を自分でコンパイルする必要はありますが。既に3000頁ぐらい(!)あります。すこし前から彼は、LLM つかって既存の場の量子論の知識を

@nick_farina But for irrops, no substitute exists:)

Blame Delta for this. Since they made domestic first class cheap in ~2015 and everyone followed suit, almost all first class flights sell out to paid customers at typically low prices. Upgrades are gone forever. Status is only good for IRROPS.

A new DT up: axiomatic-ai.com/go/digital-twi…

FDTD in the browser? Yes, Axiomatic Intelligence builds this!

So we’re pretty jazzed that marimo got GPU integration working, as of today (info here:  marimo.io/blog/reintrodu…). You hand someone a URL — they're driving a 3-D finite difference time domain simulator from a browser.

But simulation tools are often a bit old (GPU optional) They also are typically pretty hard to get going. With AI, we’re able to mitigate some of these disadvantages quickly, which is great. But it can be hard for the user to realize that without substantial back-and-forth. (3/4)

Making physics legible to AI systems today starts with digital twins – simulations of the system that are referenced back to empirical data, e.g., experiments. Our core tech today dramatically accelerates the engineer-in-the-loop build time for a digital twin. (1/4)