The only generation that knows how computers work is also the most relaxed about AI:

does openai recognize these free credits as rev? (no particular reason as I guzzle credits)

You might think regulators would be highly adverse to this. But you have to take into account three inputs: A. This would be a new form of monopoly over intelligence which doesn’t have legal precedent. This buys the labs time as it works its way through courts. B. The stronger the rhetoric and political concern re safety the stronger the argument that this is the lesser of two evils C. Most importantly this is the optimal arrangement for the US govt. This guarantees that American intelligence is globally dominant. This lets the admin ensure that its US companies that are king made at the cost of their international counterparts. This ensures the economic rent generated by AI accrues to the US economy and can therefore be taxed and redistributed as job loss accelerates. This is game theory optimal for everyone with the exclusion of VCs, startups and legacy companies who haven’t managed to reach the number one spot in their respective vertical. Oh this is also terrible for anyone betting that AI diffusion will be slow and unwieldy due to human inertia. Long compute 🫡

Say I Have 1500 amps 480 ready to roll in pa. Is there a neo-cloud inference container I can install and run?

In 2010 I helped put on @NASAAmes's first workshop on what synthetic biology could do for space exploration. @lynnjrothschild, my adviser at Ames, and I spent years arguing that you could send a single cell to Mars and let it build from whatever it found there. The microbe was never the hard part. Earth already has bugs that make plastic, food, and fuel. The hard part was the dirt. Nobody knew what Martian regolith actually gives a cell to eat, or how badly the perchlorate, salts, and heavy metals would hurt it. You can't model your way out of that question. Someone has to grow things in it. That's exactly what @Pioneer__Labs has been doing, and this week they showed it works. First they turned real rover and lander data, from Phoenix's Wet Chemistry Lab and Curiosity's SAM instrument, into Defined Mars Media: a lab recipe for the nutrients and toxins in Mars dirt. Then they screened 16 Earth organisms against it and ran more than a trillion cumulative cell divisions of evolution. The result is sPL.001, a strain of Cupriavidus necator that pulls every nutrient it needs from Martian dirt, water, and air, and makes more than 3x the PHA bioplastic of its parent. It needs no fertilizer from Earth and no washing the perchlorate out first. Their reference design says one rocket's worth of equipment could build radiation-shielded habitats for a small city in about four years. It's also built to be safe. It only grows inside a bioreactor fed acetate made from Martian CO2. Even a catastrophic leak clears planetary protection limits by several hundred times. This is what data generation looks like for an interplanetary species. Rover measurements became a lab recipe, the recipe became a trillion cell divisions of experiments, and the experiments produced an organism. Pioneer is the only team I know of doing this end to end. Congrats to @erika_alden_d and co-founder @StorkDevon, and to @unattermann, Tom Pedersen, Nathan Hicks, Jordan Mancuso, Karin Isaev, Max Schubert, Fatima Martin, Jonathan Liu, Harley Greene, and Edward Sukarto. Thanks to @AsteraInstitute for backing it. @lynnjrothschild, Pete Worden, Stephanie Langhoff, Chad Paavola: this is where that 2010 workshop was pointing. Last year I joined Erika, Chris McKay, Pamela Silver, @edwinkite, Robin Wordsworth, @mason_lab, and others on a @NatureAstronomy paper making the case for Mars terraforming research. Organism one of five is already here. Erika has spoken at @SynBioBeta, and so has @NikoMcCarty, whose @AsimovPress interview with her is the best 51 minutes you can spend on this. It is a joy to see this research making headways and I am excited to see what the future holds.

an important life lesson for me is pursue the things your interested in. Full throttle. Of course, I can't control the direction, but it's good to go along for the ride with gusto.

NYC climate peeps! I have a backpack full of delicious mycelium chips. Find me for freebie ! @ me for a drop Save a potato. Eat a mushroom!

you can watch the entire market coalesce around two scapegoats (one equity, one credit) for the sins of the ai industry that once established, they’ll ceremonially behead, with the help of the state, to let the blood offer tithe to the voters while allowing progress to go on

NYC: MyChip giveaway Tomorrow in Chelsea at Mycostories event My backpack is full of My most favorite Mycelium product ever: Salty, crispy, delicious chips! 90 Cal 3 g protein 5g fiber Nourishment worth Craving. (no potatoes harmed in the making of this post)

Seems like these days you don't see cows on the road so much.

Red mars starts with a crew leaving earth Dec 2026; they had to make the microbes on the red planet. Spot on to send them ahead. Fantastic work!!

I just had a great idea for the next season of True Detective.

the button!

Our prototype is maintaining good yaw authority through yaw maneuvers! Manual control surface deflection is working for effectively blown lift yaw control in hover. This actually made the vehicle controllable, and our test pilot changed heading 5 times during this flight. Notice during this flight that the surfaces don't need to deflect far to maintain authority. Previously, with only canted (angled) motors and differential thrust, one set of 4 motors spinning in one direction had to spin up to near maximal rpms while the other 4 spun down to 10-20% throttle, and we we were still not getting the yaw authority we needed. The motor rpm being so high for 4 motors meant very limited pitch and roll authority especially in any winds. The blown lift surfaces allow for motor rpm differential between the clockwise and counterclockwise pairs to only need a median 5.5% differential, allowing a lot more thrust margin for roll and pitch. In comparison, a flight from yesterday required almost 19%. The 95% percentile motor rpm differentials were 29% vs 90%. That's a ~70% reduction in motor rpm difference. This means our pilot had much, much more control authority in flight today, even as winds started to pick up. Now this control implementation is only a make-it-exist version, we need to make it better. The controller is not actually responding to the control surface deflections but we will add a feedforward for the servo angle for the surfaces based on the yaw rate. To make it better, we'll work on a firmware update in the coming days. There's some SIL work and structure upgrades to be done in the meantime. We need to model control surfaces in SIL and the motor mount <> wing interface is shearing and cracking in a few places. It's not structural but it's worth rebonding. Once done, we should be ready to test the more robust controller through maneuvers in altitude hold mode, before expanding the envelope to transition soon after that! Proud of our small scrappy team for continuing to up the iterative cadence of learning.

All of us OG meme h/t credit to @venturetwins for Claude

Wild 24 hours for AI and lots of different proposals have been made. TLDR; the only *tangible* new fact is that OpenAI and Anthropic are going to have embedded 3rd party evaluators from unknown organizations with Dario floating METR as a possibility. Having 3rd party evaluators is smart as there is no Section 230 style liability shield for model outputs and showing a “duty of care” will be important in future litigation. Several internet companies might have gone bankrupt without Section 230 so limiting liability really matters. There are minimal investment implications from this single new fact, but I do think that for anyone who wants a “smoother for longer” cycle then most constraints are good: wafers, watts, real rates and spreads. Excessive regulation is a different matter but I don’t think we are anywhere close to this even if the vector changed over the last 24 hours. To summarize the events: Dario made the most maximalist proposal of the weekend: embedded 3rd party evaluators, a national regulatory regime for models beyond a certain capability/ingredient threshold, a broad international regulatory pact between democracies, stricter limits on compute/distillation for China and then a different international regulatory regime that encompasses China. Before there is a national regulatory regime, he wants a Sherman act waiver so that Anthropic can safely coordinate with OpenAI and other frontier labs without antitrust fears. TBF, this latest proposal is much less maximalist than some of his prior proposals like “Policy on the AI Exponential,” where he advocated for an FAA for AI. I believe he is sincere in his beliefs. And despite all the protestations, all of this would also probably be good for his business over the long-term. Sam agreed that embedded 3rd party evaluators were a good idea and stated they would implement them. Again, this is smart as should help limit future liability. Elon said “Dario is right” and later specified that “Dario is right that there should be some oversight. Peer review of AI by competitors is the right way to start this off.” This would be a MPAA like self-regulatory structure for AI with regular calls between the labs plus a process where each new model is evaluated for safety by competitors for a 1-2 week period before being released. That is *wildly* different from Dario’s proposal and in-line with what David Sacks has been proposing. Elon also stated that nothing was going to slow down open-weight models. Demis said that Dario’s essay was a “step in the right direction.” Dario also said that he was also open to Demis’ idea of a FINRA like self-regulatory structure as part of his proposal. David Sacks had a thoughtful post where he said that Dario and Sam should pace unilaterally, called the antitrust waiver a cartel request and denied that METR was truly independent given their ties to Anthropic. Sriram Krishnan, former White House AI advisor, noted that it would be important to have the 3rd party evaluators come from independent organizations that are not affiliated with any lab, which is basically an indirect statement about the relationship between METR and Anthropic which Sacks was explicit about. Clem from Hugging Face said they were open to being a neutral 3rd party evaluator, which is interesting especially if Jensen was consulted before that post. Alexander Wang from Meta noted that alignment would be an increasing focus going forward. An executive order seems likely after all this and the language in this EO is going to be really important. It is possible to democratize and distribute AI broadly and safely without centralizing it in the hands of a few corporations who might each become more powerful than any single government. I do not want a few humans in control of intelligence. I want us all to have our own intelligences that reflect our own values and human variation in all of its richness. Intelligence distribution over intelligence centralization FTW.

To rehash the timeline: - WSJ article goes up 18 minutes before this post - Arguably the most viral post from a IC researcher since the beginning of the supercycle - Kokotajlo on Rogan the next day - D senate leadership: Van Hollen (MD), Murphy (CT), Kelly (AZ), Gallego (AZ) post lockstep messaging about x-risk this morning - Appearances on Fox, CNN booked the following evening. - Sanders (VT) announces briefing next week with Tegmark and Cotra Its not at all conspiratorial to see the pattern here. Safety types have dominated this conversation since the beginning by remaining significantly more savvy than any other actor. They're locked out of the Trump admin by Sacks and Michael and are clearly making a play for a 2027 legislative push within D controlled congress.

Can you predict next week's weather yourself? We just wrote a blog post that explains how! ⛅️🌍🌧️ Running weather forecasting required having a supercomputer and CS/physics Phd. Now, AI-based models need fewer resources and are available to anyone on @huggingface. So, prediction the weather should be accessible to all. But only a few labs can actually run these models because there is no standardized documentation guiding you from the model weights to an actual forecast. Together with @EarthmoverHQ, we wrote a blog post and a demo walking you through that: huggingface.co/blog/hugging-… The blog covers the code for fetching initial conditions from the Earthmover Marketplace, building a model's input batch, running inference and comparing the forecast against historical climate data. You can also run Aurora (Microsoft), WeatherNext2 (Google) and AIFS (ECMWF) on our demo space. This is a first step to make AI-based weather forecasting models more accessible, and we'd love to build more of that. Tell us what's missing or your feedback in the blog comments!