Being an AI first product means outages completely break your product. I was wondering why new user subscriptions were so low today but then saw that xAI uses AWS. If you're starting an AI-first business, please add backups. Thankfully, my systems are built to easily switch providers. For the future, @usescholarly will automatically handle outages by going down our backup list. Make sure you have all 3 large-scale providers as backup: OpenAI (Azure), Grok/Anthropic (AWS) & Gemini (GCP)

Gemini will be the truly next generation model. The model generations are: GPT 3.5, GPT 4.0 and o1/o3. o3 was an extension of o1, more RL. If gemini 3.0 is not a major cheap, ~20% increase in Arc AGI // HLE then we're at least ~5 years out from AGI.

climate friendly unfortunately rebranded to degrowth so it's completely done in the U.S. at least. Europe's climate policies have reduced it to a joke that lets 100k+ people die of heat rather than install ACs.

The Chinese way of promoting bureaucrats with tests might be the best. Mayors+ should have at least a college degree and some training in leading a big city. We should however NOT create a universal test or enforce single party rule. Something in between; reward talent, ensure our cities flourish.

i think @usescholarly will eventually merge all the existing features into a single website creator for users to use. I explored this a while back, but models were just not there yet. They required too much feedback, more tokens, more $$. Not profitable for a $20/mo plan.

Don't do this. One of my gripes of modern web programming is the lack of fundamentals. Use HTTP!! Don't be overly verbose, why success? Why error? Why waste that precious bandwidth and memory on unnecessary fluff.

AGI before GTA 6

I'm debating downgrading from my 20x max plan, or just stop using Claude entirely.

codex is really really good; much better than Sonnet. Still not as good as opus.

OpenAI will probably be the first to fall. They have amazing models, talent, and moat but they've just made fundamental mistakes. outside our AI bubble, people are starting to dislike AI for it's water & energy usage. OpenAI will be their scapegoat. They've also lost the mandate; no longer will they have the best models. Their only moat now is normal users which are starting to hate them.

Kimi K2 writing skills are SOTA by every metric. There's something special in there, just like Anthropic's "special sauce" for coding. bravo

Zohran Mamdani's win will be catastrophic for both sides of the isle. The country is extremely divided right now, now the far-left has enormous leverage over the party. The divide between the 2 parties will continue to grow until it just stops. Something will happen. 2028 will be the decider, we need someone to unite this fragmented country.

America is losing the real battle: truly open⎵AI. Everyone knows that in the fullness of time, open wins. Open is crucial to a thriving and competitive ecosystem. You see this with the proliferation of Chinese model finetunes among US startups. They’ve done a great job. This has been true for kernels and operating systems (Linux), compilers (GCC and Clang), browsers (Chromium and Blink), runtimes (v8 and Node.js), frameworks (React and Next.js), and it has to be true for AI. AI is too important to remain closed, because the future of civilization will be built on it.

🚨Sensational title alert: we may have cracked the code to true multimodal reasoning. Meet ThinkMorph — thinking in modalities, not just with them. And what we found was... unexpected. 👀 Emergent intelligence, strong gains, and …🫣 🧵 arxiv.org/abs/2510.27492 (1/16)

RL is just insane man, these models just want to learn.

Thank you, Mr. President @POTUS, for this opportunity. It will be an honor to serve my country under your leadership. I am also very grateful to @SecDuffy, who skillfully oversees @NASA alongside his many other responsibilities. The support from the space-loving community has been overwhelming. I am not sure how I earned the trust of so many, but I will do everything I can to live up to those expectations. To the innovators building the orbital economy, to the scientists pursuing breakthrough discoveries and to dreamers across the world eager for a return to the Moon and the grand journey beyond--these are the most exciting times since the dawn of the space age-- and I truly believe the future we have all been waiting for will soon become reality. And to the best and brightest at NASA, and to all the commercial and international partners, we have an extraordinary responsibility--but the clock is running. The journey is never easy, but it is time to inspire the world once again to achieve the near-impossible--to undertake and accomplish big, bold endeavors in space...and when we do, we will make life better here at home and challenge the next generation to go even further. NASA will never be a caretaker of history--but will forever make history. Godspeed, President Donald J. Trump, and Godspeed NASA, as America leads the greatest adventure in human history 🇺🇸

Wow.

It is unfortunate that NASA’s team and the broader space community have to endured distractions like this. There are extraordinary opportunities and some risks ahead and so the focus should be on the mission. With many reporters and other interested parties reaching out, I want to help bring some clarity to the discussion... unfortunately, that means another long post: I have met Secretary Duffy many times and even flew him in a fighter jet at EAA Oshkosh--probably one of the coolest things a cabinet secretary can do. I have also told many people I think he has great instincts and is an excellent communicator, which is so important in leadership. If there is any friction, I suspect it is more political operators causing the controversy. This isn't an election or campaign for the NASA Administrator job, the Secretary is the leader and I will root for his success across his many responsibilities. We both believe deeply in American leadership in the high ground of space--though we may differ on how to achieve that goal and whether NASA should remain an independent agency. It is true that Athena was a draft plan I worked on with a very small group from the time of my initial nomination through its withdrawal in May. Parts of it are now dated, and it was always intended to be a living document refined through data gathering post-confirmation. I would think it is better to have a plan going into a responsibility as great as the leadership of NASA than no plan at all. It is also true that only one 62-page version of the plan (with unique header/footer markings) was delivered in hard copy back in mid-August to a single party. I learned it was leaked to reporters and across industry last week. It seems some people are letting politics get in the way of the mission and the President’s goals for space. Personally, I think the “why” behind the timing of this document circulating--and the spin being given to reporters--is the real story. While the full plan exceeded 100 pages, it centered around five main priorities that I will summarize below, including some specifics on the topics attracting the most interest. There is the question--why not release the entire document? Well, one party is clearly circulating it, so I am sure it is only a matter of time before it becomes public--in which case, I will stand behind it. I think there are many elements of the plan that the space community and NASA would find exciting, and it would be disappointing if they never came to fruition. Mostly, I just don’t think the space community needs to debate line-by-line while NASA and the rest of the government are going through a shutdown. I will say everything in the report is consistent with my Senate testimony, my written responses to the Senate for the record, and all the podcasts and papers I have ever spoken to on the subject. - Reorganize and Empower Pivot from the drawn-out, multi-phase RIF “death by a thousand cuts” to a single, data-driven reorganization aimed at reducing layers of bureaucracy between leadership and the engineers, researchers, and technicians--basically all the “doers”. Align departments tightly to the mission so that information flows for quick decision-making. One example, which was mischaracterized by a reporter, was exploring relocating all aircraft to Armstrong so there could be a single hierarchy for aviation operations, maintenance, and safety. From there, aircraft like T-38s would operate on detachment at JSC. Other goals of the reorganization, would be to liberate the NASA budget from dated infrastructure that is in disrepair to free up resources to invest in what is needed for the mission of the day. And maybe most importantly, reenergize a culture of empowerment, ownership, and urgency--and recalibrate a framework that acknowledges some risks are worth taking. – American Leadership in the High Ground of Space Put more astronauts in space with greater frequency, including rebooting the Payload Specialist programs to give opportunities for the NASA workforce--especially on opportunities that could unlock the orbital economy--the chance to go to space. Fulfill the 35-year promise and President Trump’s Artemis plan to return American astronauts to the Moon and determine the scientific, economic, and national security reasons to support an enduring lunar presence. Eventually, transition to an affordable, repeatable lunar architecture that supports frequent missions. When that foundation is built, shift resources toward the near-impossible that no one else will work on like nuclear electric propulsion for efficient transport of mass, active cooling of cryogenic propellants, surface power, and even potential DoD applications. To be clear, the plan does not issue a directive to cancel Gateway or SLS, in fact, the word “Gateway” is used only three times in the entire document. It does explore the possibility of pivoting hardware and resources to a nuclear electric propulsion program after the objectives of the President’s budget are complete. On the same note, it also seeks to research the possibility that Orion could be launched on multiple platforms to support a variety of future mission applications. As an example of the report being dated, Sen. Cruz’s has subsequently incorporated additional funding in the OBBB for further Artemis missions--which brings clarity to the topic. - Solving the Orbital Economy Maximize the remaining life of the ISS. Streamline the process for high-potential science and research to reach orbit. Partner with industry (pharmaceuticals, mining, biotech, etc) to figure out how to extract more value from space than we put in--and critically attempt to solve the orbital economy. That is the only way commercial space station companies will have a fighting chance to succeed. I don’t think there is anything controversial here--we need to figure out how to pay for the exciting future we all want to see in space. – NASA as a Force Multiplier for Science Leverage NASA’s resources--financial (bulk buying launch and bus from numerous providers), technical, and operational expertise to increase the frequency of missions, reduce costs, and empower academic institutions to contribute to real discovery missions. The idea is to get some of that $1 trillion in university endowments into the fight, alongside NASA, to further science and discovery. Expand the CLPS-style approach across planetary science to accelerate discovery and reduce time-to-science... better to have 10 x $100 million missions and a few fail than a single overdue and costly $1B+ mission. I know the “science-as-a-service” concept got people fired up, but that was specifically called out in the plan for Earth observation, from companies that already have constellations like Planet, BlackSky, etc. Why build bespoke satellites at greater cost and delay when you could pay for the data as needed from existing providers and repurpose the funds for more planetary science missions (as an example)? With respect to JPL, it was a research request to look at overlaps between the work of the laboratory and what prime contractors were also doing on their behalf. The report never even remotely suggested that America could ever do without the Jet Propulsion Laboratory. Personally, I have publicly defended programs like the Chandra X-ray Observatory, offered to fund a Hubble reboost mission, and anything suggesting that I am anti-science or want to outsource that responsibility is simply untrue. – Investing in the Future The congressionally mandated “learning period” will eventually expire, and the government will inevitably play a greater role in certifying commercial missions (crewed and uncrewed) just like they do with aircraft, ships, trains, etc. NASA eventually should build a Starfleet Academy to train and prepare the commercial industry to operate safely and successfully in this future space economy, and consolidate and upgrade mission control into a single “NORAD of peaceful space,” allowing JSC to become the spaceflight center of excellence and oversee multiple government and commercial missions simultaneously. Other investments for the future included AI, replacing dated IT systems, and ways to alleviate the demand on the Deep Space Network. - Closing This plan never favored any one vendor, never recommended closing centers, or directed the cancellation of programs before objectives were achieved. The plan valued human exploration as much as scientific discovery. It was written as a starting place to give NASA, international partners, and the commercial sector the best chance for long-term success. The more I see the imperfections of politics and the lengths people will go, the more I want to serve and be part of the solution... because I love NASA and I love my country 🇺🇸🚀