QuantumOnboard @IowaLearning

Science fiction

Turns out the future does have flying cars

Real life, baby

IBM Advances to Stage C of DARPA Quantum Benchmarking Initiative thequantuminsider.com/2026/10/07/ibm…

We just unveiled Steelfin at AUSA. It is a modular, low-cost mass-production missile defense interceptor that brings two-stage performance to dozens of existing platforms.

Anduril Industries @anduriltech

4 days ago

Today’s missile defense model does not scale. The modern fight demands an interceptor that can match cost and capability to threat. Introducing Steelfin.

88 519 7K 470K 783

this is now a new skill /in-your-own-words in pstack 0.15.23! extremely useful to ensure that you and your agent are aligned on the goal and that it understands your intent clearly

lauren @poteto

3 weeks ago

this has become one of my most used prompts recently: > restate in your own words what you think my goals are and what the problem i'm trying to solve is

343 598 14K 834K 12K

UCLA quantum physicist @PrinehaN says DARPA just advanced 4 quantum computing companies to its final stage, signaling that the science is solved and now they just have to build it: "The DARPA Quantum Benchmarking Initiative is DARPA's way of saying, 'We're skeptics. Show us that you have what it takes to actually build a fault-tolerant, useful quantum computer.' And that's this objective of a utility scale computer by 2033." "In advancing companies from Stage B to Stage C, DARPA is saying, 'Okay, the technical risk is gone. The science risk is gone.' Now you literally just have to build it." "Those are the companies that have an architecture that scales, algorithms that go with that architecture, and they know which problems they're gonna solve where there's actual value." @UCLA @DCVC

Point-form recap of the space discussing / brainstorming @Kardashev_AI - ICYMI: -Turning e/acc into a formal institution. The organizers want to move beyond memes, posting, and informal group chats into scientific papers, philosophy, policy proposals, and sustained influence. They said the nonprofit had been under discussion for over a year. -Leadership and initial setup. Beff/Gill would serve as chairman and help with fundraising while continuing to run Extropic. Maxwell, described as having worked with neuroscientist Karl Friston, would run the institute. A domain was purchased, but incorporation, legal structure, staffing, and budgets were still being worked out. -An unexpected financial windfall accelerated the announcement. Beff described unsolicited money arriving through X Money amid crypto/meme activity, reaching roughly $100,000 by late in the conversation. He said he had not launched or owned the associated token and intended to put the proceeds toward the nonprofit. Initial spending would cover legal fees, computers, basic AI usage, and a short period of operations before larger fundraising. -Creating a counterweight to effective altruism and AI-restriction advocacy. Participants argued that AI-risk organizations have far more funding and coordination than e/acc. They anticipated further funding from AI-company wealth and potential IPOs, and wanted an institution capable of responding with research and policy arguments. Those funding forecasts were speculative. -Researching intelligence across biological and artificial systems. Proposed work included thermodynamics, information theory, neuroscience, and alternative computing—especially understanding how intelligence emerges in different physical substrates, rather than focusing exclusively on conventional digital AI. -Exploring hybrid biological–silicon intelligence. One proposed experiment was to connect cultured-neuron computing systems from Cortical Labs with Extropic’s thermodynamic computing hardware. The aim would be to demonstrate a combined biological and artificial intelligence system. This was an exploratory idea, not an announced working prototype. -Developing a deeper intellectual foundation for e/acc. They discussed complex adaptive systems, active inference, economics, and how ideas spread and influence behavior. The ambition was to develop a substantial school of thought, with advancing civilization’s energy use and technological capabilities—the Kardashev-scale framing—as a guiding objective. -Protecting access to AI and compute. A recurring policy theme was the freedom to own computing hardware, train models, use open weights, and run inference independently. Speakers proposed a “First Amendment for synthetic thought” and a “Second Amendment for compute” as advocacy concepts. -Opposing concentrated control and regulatory capture. Participants worried that a few large AI labs could use regulation to exclude competitors and restrict independent AI use. They argued that centralized control, censorship, and authoritarian outcomes deserved greater attention in AI-risk discussions. -Studying markets as a mechanism for AI alignment. Speakers suggested that customer demand, competition, reputation, and liability could encourage useful, reliable models. They criticized safety certification that, in their view, could become procedural theater or shield companies from responsibility. -Connecting autonomous AI with crypto. They speculated about AI agents earning money and paying independent hosts for compute through cryptocurrency. Related ideas included community-funded model training, shared ownership of models, and distributing model revenue to contributors. -Debating nonprofit structure and independence. The group discussed research and lobbying entities, potentially using both 501(c)(3) and 501(c)(4) structures. A DAO was floated, but the conversation leaned toward a simpler nonprofit structure. They also emphasized avoiding dependence on one dominant donor. -Reaching people outside the tech/Twitter audience. Ana and Shaw stressed that research alone would not be enough. Suggestions included AI-assisted journalist research and outreach, accessible explanations, cultural content, events, and concrete stories about benefits such as better healthcare, more family time, and greater prosperity. -Making optimism credible and actionable. An extended discussion connected pessimistic narratives with “learned helplessness.” Participants argued for a specific, believable picture of a better future that motivates people to act, while acknowledging that broad prosperity and managing job disruption require actual solutions. -Choosing research with tangible public benefits. A participant with lobbying experience suggested focusing on problems that policymakers and ordinary people already care about, particularly healthcare. Maxwell also discussed distinguishing recoverable crises from self-reinforcing catastrophes, using demographic decline as an example. -Next steps and community participation. The organizers planned to formalize the organization, raise additional funds, publish materials explaining its direction, and invite researchers, policy specialists, communicators, and supporters to contribute. No detailed launch schedule or finalized research roadmap was presented. -Other conversation and closing banter. Side topics included Extropic’s progress, AI documentaries and podcasts, crypto-community dynamics, memories of early e/acc Spaces, San Francisco culture, and restaurant automation. The late discussion about a massive nonprofit data center was joking rather than a concrete commitment.

Beff (e/acc) @beffjezos

3 weeks ago

Kardashev Research is a proposed nonprofit research institute and think tank aimed at turning e/acc into a lasting scientific and policy institution. We discussed some of what we'll tackle in this X Space last night, summary of discussion in replies below

98 150 771 102K 148

Deploying DiffusionGemma-Jev (djev) just got a lot easier. You can now spin up a Jev API-compatible endpoint on Google Cloud Run using a single command. Performance is solid: ~35-60 ms for single step latency and batch@32 is ~100-123 requests/sec. It's a straightforward way to experiment without needing your own GPU. Runs at roughly $3/hr and drops to $0 when idle. Get the code and instructions here: github.com/taeold/djev-run

@mattpocockuk Let's go!!!

PsiQuantum has finalized a $100 million award from the U.S. Department of @CommerceGov to strengthen American leadership in quantum computing and semiconductor supply-chain security.​ With this funding, combined with our supply chain anchored in the United States, PsiQuantum will accelerate the R&D, manufacturability, and performance of our state-of-the-art critical components—right here in America. Learn more: psiquantum.com/news-import/ps…

@SawyerMerritt I use it to help continuing to improve my pipelines for grok'ing everything from my day to day life, school, and my own learning adventure. Also it helps fix/correct any local issues I have ever run into from this app or that app not working to diy issues around the house.

The Westinghouse eVinci™ microreactor has been selected by the U.S. Army (@USArmy) for the next phase of the Janus Program. The Janus Program is designed to accelerate advanced microreactor technologies capable of delivering secure, resilient and reliable energy for national defense installations. This milestone highlights the growing role that the eVinci microreactor can play in strengthening energy security, enhancing mission assurance and supporting operational readiness for critical defense missions. Learn more about this milestone and its impact on the future of defense energy: ow.ly/EVxa50ZFHxc

One of the more novel pieces of equipment we run is The Monster — conceived, designed, and built right here in Bastrop. It weighs ~140,000 lb and carries Prufrock sections that weigh between 465–480k lb. Its platform tilts ±10°, which lets the machine launch directly into the ground — no pit, no shaft, no concrete, and no crane. Because it eliminates the crane, we can launch and retrieve machines inside buildings and parking garages, which is key to maximizing tunnel usefulness in urban areas. When we complete a tunnel and retrieve Prufrock on The Monster, an SPMT can be driven directly under it. We then drive Prufrock to the next tunnel site, where The Monster tilts downward and launches it straight into the ground. This enables Prufrock to relaunch in under 24 hours. The video below is from us retrieving Prufrock 3 inside the Tesla Gigafactory.

Genuinely one of my favorite interviews I've done. Thanks to @tbpn for having me on and @johncoogan & @jordihays for asking great questions

Beff (e/acc) @beffjezos

3 months ago

Was just live on TBPN! (long time fan first time caller) Includes a restrospective on e/acc, discussion of the current and next market cycle, how thermo AI research is NewGame+, and more about our Gov LOI announced yesterday x.com/i/broadcasts/1…

8 5 93 82K 22

☑️ New from LangChain Academy: Become a LangChain Certified Agent Engineer The first certification for the full Agent Development Lifecycle. For the first 2 months, use code LCAE50 for 50% off. academy.langchain.com/pages/certific…

Thermodynamic computing is coming home 🇺🇸🫡

Extropic @extropic

3 months ago

We have signed a $75M LOI with the U.S. Department of Commerce to scale and onshore Thermodynamic Computing in the USA 🇺🇸 Read more: extropic.ai/writing/thermo…

45 56 647 149K 88

Today we're releasing Laguna S 2.1, our most capable model to date. It's a 118B total parameter Mixture-of-Experts model with 8B activated per token, a context window of up to 1M tokens, and thinking and no-thinking modes. Capable enough to hold its own against models many times its size. Small enough to run on a single @NVIDIAAI DGX Spark. Laguna S 2.1 is fully open under OpenMDW-1.1, with weights available today on @huggingface poolside.ai/blog/introduci…

@BrianRoemmele Blast from the past! Love it!

@Scobleizer Great work as always! Thx!