@creon
Account based inUnited States
About this account
- Account based in
- United States
- Connected via
- United States App Store
Account-level information from X, not a live location or the device used for a specific post.
NASA for 32 years Dir. of R&D @Planet Labs since 2015 -- Science is poetry written about nature in the language of mathematics. -- Tweets are PERSONAL views
N. California (often)
Joined May 2008
- Tweets7.1K
- Following1.2K
- Followers6K
- Likes16.4K
Pinned Tweet
some favorite books: docs.google.com/document/d/1…
my substack: creon.substack.com
regenerative agriculture: docs.google.com/document/d/1…
health & life extension: docs.google.com/document/d/1…
Summary of the progress in particle physics since I was born:
By 1974 the Standard Model (SM) existed as a completed theory. Its predicted particles are all found, ending with the Higgs in 2012
The only established laboratory phenomenon beyond the SM is neutrino oscillation (1998-2002), which requires nonzero neutrino mass
Gravity lies outside it. Dark matter & dark energy aren't established laboratory phenomena.
Possible simple solutions to some AI safety concerns:
1. Prevent (and/or prohibit) employing massive coordinated long-running agent swarms for offensive cyber hacking tasks.
2. Allow (1) for defensive cyber security tasks.
3. Elon-style competitive safety checking between frontier labs, where they monitor each other's models for (1) and other dangerous behaviors. Including Chinese models
4. All Frontier Labs adopt the Microsoft Humanist AI (HAI) constitution. Pressure China to do the same.
5. Remember that the vast majority of users don't need to solve the Riemann hypothesis (and don't want to pay for that capability), and that solving it not a safety issue anyway.
No general pause required. No kill switch required. No government takeover or frontier lab regulatory capture required.
So, when do y'all think Optimus or any humanoid robot will be able to do things like this?
This is a MAJOR breakthrough.
Tackling a 60-year-old challenge in quantum chemistry: making density functional theory scale nearly linearly with system size.
This has huge implications for opening the door to realistic systems that have traditionally been too expensive to simulate.
AI has attempted to accelerate these calculations, but models generally struggle to extrapolate, particularly to systems larger than those seen during training. Unlike text or images, quantum-mechanical training data is extremely expensive to generate.
We present a single unified AI model that performs quantum-mechanical simulations of both molecules and materials in quasi-linear time.
Using a novel Fourier neural operator variant, we learn the underlying Kohn–Sham equation map to produce physics-informed, self-consistent answers.
Unlike prior AI approaches that directly predict chemical properties, our model works through intermediate steps to improve difficult predictions, resembling inference-time reasoning in large language models.
To demonstrate its scalability, we run a self-consistent calculation of a magnesium dislocation with about 80k electrons on a single GPU, which previously required about 7k GPUs.
tensorlab.cms.caltech.edu/us…
@DanishK42 @Caltech
This does not illustrate "why x-ray lasers?" It illustrates "why high numerical aperture?"
And did you know it is just the uncertainty principle?
I can show you in three lines.
Frankenstein is relevant to the AI and longevity communities.
The original book Frankenstein; or, The Modern Prometheus (Mary Shelly, 1818) is arguably the first science fiction novel, and arguably the most important.
Written over one stormy weekend at the dawn of the scientific age. Galvanism had just been discovered - the (earliest) connection between electricity and life.
Shelly saw the future and wrote it. Some examples:
Grafting. Regeneration. Life extension. Artificial life.
Artificial beings. Artificial consciousness. The question of "what is the soul?”. The dual nature of technological progress. The rights and responsibilities of humans v. artificial beings.
In the novel, the creature (never called the "monster") is named Adam: The first man.
Her novel was also the deepest “frame story” (a story inside a story inside a story, inside a story, …) until very recent times. It should be required reading.
Approximately 500 film adaptations have been produced. I think Del Toro's (2026) is the best by far. It is complex, authentic, creative, and definitely worth seeing.
The Last Man is one of Shelley's later novels. Written exactly 200 years ago, it is about the last man to exist. It is set in her far future, which is now. The last man was the only survivor of a global pandemic,
Do I understand this correctly? @DarioAmodei is concerned the CCP will use open-weights models to repress their people, thus US users shouldn't be allowed to use or modify open weights models?
And implementing KYC to prevent Chinese distillation of @AnthropicAI's models (trained on stolen data) will slow Antropic's growth? The company with the highest growth in history? So the answer is... ban use of Chinese models in the US?
And their (Anthropic's) models are so dangerous and Anthropic so singularly wise that only Anthropic can decide what the elected US government (and anyone else) can use their models for?
Transparent greed, anti-competitive behavior, and arrogance. Or so it seems to me.
Dario Amodei has responded to the controversy swirling around Anthropic's refusal to sign the open letter. He says he agrees with much of it, but does not agree that open-weights models necessarily make it easier to develop safeguards, or that broad access to capabilities necessarily helps defenders more than attackers. His main concern is the attacker-defender asymmetry in biological attacks.
He also reiterates his position on authoritarian governments. I will quote;
'My primary concern is the risk that authoritarian governments—not solely the Chinese Communist Party (CCP), although the CCP is clearly the most capable threat—build AI models that are more powerful than those built by the US, and use them to achieve permanent military superiority or perpetrate incredibly deep repression of their own people.'
I will also quote his closing paragraph in full;
'To summarize my and Anthropic’s position, we have not and are not advocating for a ban on open-weights models as a category. We should instead focus on keeping powerful chips out of authoritarian hands, stopping industrial-scale distillation, and requiring safety testing of all sufficiently capable models, open and closed.'
This Bell Labs lecture from 1959 explains the basics of semiconductor electronics in a very simple way. It is the foundation of all VLSI that runs the world today.
One of the men who invented the transistor stood at a Bell Labs blackboard and explained how it works. AT&T kept the film. Almost nobody watches it.
Walter Brattain shared the 1956 Nobel Prize in physics for the transistor. Every chip training every model in 2026 descends from what he built with Bardeen and Shockley in 1947.
This is 28 minutes of the inventor teaching semiconductor physics himself. Properties, simple model, history, impact, new phenomena, written on the board in his own hand.
Watch the section where he draws the crystal lattice and shows what an electron and a hole actually do inside it. The abstraction most engineers carry around, drawn from scratch by the man who found it.
An embedded engineer I know watched it twice and said 28 minutes of chalk beat a semester of textbook diagrams.
Free on YouTube from the AT&T archives, 10 chapters, black and white.
Everyone uses his invention. Almost nobody hears him explain it.
creon levit retweeted
This letter is one of the best responses I’ve seen..
Dear Mayor Mamdani,
I'm writing from Canada, which means I have no vote, no dog in your reelection, and no reason to care about your press strategy. Except that I do, because it's hard not to notice a mayor treat foreign policy like it's the only portfolio he actually wants.
I'll be honest about how this started. When the video first crossed my feed, I assumed it was fake. Not in the "bold move" way, in the actual "someone ran this through a deepfake generator to make a point" way. A sitting mayor, on camera, gravely announcing he'd exhausted every legal avenue to arrest a foreign head of state, only to concede he never had the authority in the first place. I sat there looking for anything that would let me file this under a bad prank instead of a real statement from a real sitting mayor. Then I found the CNN piece, the NPR piece, the France 24 piece, all reporting it straight, and had to sit with the fact that this actually happened, on purpose, from the mayor of New York City.
So now I'm just curious, genuinely, no axe to grind, from way up here with no stake in your reelection. Whose mandate is this? You ran for mayor of New York. Housing, transit, safety, the budget, childcare, groceries people can actually afford, that's the job description. Nowhere in there does it say "unofficial spokesperson for the International Criminal Court" or "designated antagonist of the Israeli government." Nobody voted on that. There was no ballot line for it. So who exactly gave you the right to carry this out on the city's behalf, and in whose name are you actually doing it? From where I'm standing, this mandate doesn't look like it belongs to the people of New York at all. It looks like it belongs to Hamas, Hezbollah, and the IRGC.
Here's the part that makes it feel less like conviction and more like theatre. Ask yourself the question your video conveniently skipped. If the head of the IRGC landed at JFK, would there be a video? If a Hezbollah commander wanted a night out in Manhattan at some swanky club, would you tell him, with that same solemn cadence, that he's not welcome? If someone from Hamas's political bureau, the men who actually planned October 7th, decided to visit, would you even flinch? Somehow I doubt it. The cameras only seem to come out for the one target that plays well to a specific room, and I notice, from all the way up here, that the room never asks about the others. It makes a person wonder who this performance is actually for.
You don't owe me an answer. I can't vote, I can't donate, I can't even yell at you at a town hall. But the question doesn't need an American passport to be obvious. It's obvious from Canada. It's probably obvious from the moon.
With all the respect you've earned, which is none.
Francine Goldberg
Amtrak is a wonderful way to see the country. But why doesn’t it have Starlink @elonmusk
People first programmed in assembly language. Then, with the first compilers, people wrote in a high-level language (HLL) but checked the assembly code output.
Now nobody does checks the assembly language.
Prompting is the new high-level language. Python (or another HLL) is the new "assembly language". At first, we checked it. Now almost nobody does.
This means your average programmer generates far more "lines of code" - arguably a mixed blessing. Good coding agents also make it much easier to progress on multiple projects at once. That probably is a blessing.
But I now see it also truly means someone who is "not a programmer" can write software. If they're very good at specifying what they want (in English), monitoring the results, and correcting mistakes.
Of course, they also should work on a reasonable problem. And have taste.
I just met a guy at a party who'd created some useful apps (w/Claude code). They looked good, they felt good, they performed valuable functions (some of them). He was not writing code. He didn't know how. For instance, he didn't know what a library was.
Interesting.
Disappointed in Steven Pinker:
- Promoting fake food
- Claims MAHA = "quackery"
- His cited article is clear: unprocessed food wins. Did he read it?
- "Unprocessed" includes oats, grains, nuts, etc: bad for many (note @bryan_johnson's recent troubles)
- And he blocked replies!
I've always thought that the campaign against "ultra-processed" foods was mindless. What the heck does that even mean: the more things you do to food, the less healthy it is? New RCTs and analyses confirm that the stigma is meaningless (it's calories, fiber, & other causes, not "ultraprocessing") and even harmful: it encourages RFKj's MAHA quackery. I suspect that food researchers are subject to class bias: food that poor people like and can afford must be unhealthy. vox.com/future-perfect/49404…
