Pinned Tweet
Low rent holograms on five @M5Stack transparent oleds + bare metal pi zero. Camera angle not great bc don't want to mess up fixture but you can see the vision.
I wrote a bit-banged driver for 5 simultaneous i2c connections , codex sped it up for consistent 60fps using simd style, fable used 3 cameras + feedback to make the volumetric images and validate they were correct. Irl looks really cool , lots of other tricks once I fix the insane bench entropy 😹
FWIW: I know very good students interested in such things that are/were very happy at Jane street.
Exciting news: we're looking for interns for our formal methods team!
Help us turn OCaml into the best language on the planet for writing probably correct code. Hiring in both London and New York.
janestreet.com/join-jane-str…
janestreet.com/join-jane-str…
1
307
Many are saying
NEWS: Wall Street execs and NSA have told Congress that outages were caused by patching vulnerabilities identified by AI models, sources say
AI did not cause outages but found so many issues that the rush to fix them before an attack overwhelmed systems
washingtonsun.com/technology…
So far: tried 6.2 for several hours. And, for about 2hr:57min have been dealing with bullshit safety mis-classifications. It's all so tiresome.
1
6
321
Dawson Engler retweeted
Replying to @micsolana
The best part is how he stops for applause after mentioning that his grandfather stood with Ghandi, then stop for more applause when he says it doesn't matter if your grandfather fought in Normandy.
If you use multiple codex accounts, @OpenAI now has a catastrophic change that makes a session stupid if you started the session using account A and then switched to account B. Short term work around: pin your credentials using a per session .codex/.
Specifically: the reasoning traces emitted under one account *DO NOT HOLD* for another. The mechanism hypothesized by astra is that reasoning traces (and compaction) are encrypted with a *per account key*. So when you switch accounts, the codex session loses all of its past reasoning, which is where most of the action occurs. This can be arbitrarily bad depending on what got lost.
I couldn't figure out why astra seemed much more stupid in some sessions vs others. Unfortunately this appears to be the reason.
Astra max did a set of experiments to nail this down. The clever hack below is one of the simplest ones. In its words:
• • A controlled Codex experiment reproduced account-
associated loss of private-state recall: a
session resumed successfully under a different
account but could no longer recall information
available under the original account. Switching
back restored recall. The affected continuation
still reported approximately 97.5% cached input,
so a high cache-hit rate did not establish that
the private reasoning remained available.
• The experiment asked Codex under account A to
choose a six-digit number privately and reply
only ready. The number was absent from the
visible conversation and saved plaintext; the
session contained an opaque encrypted reasoning
record. We froze the complete local state and
restored that same snapshot before each
continuation, keeping the machine, model, session
ID, paths, and settings fixed. With the record
intact, A recalled 583104, B returned UNKNOWN,
and A recalled 583104 again. A negative control
that removed the reasoning record under account A
also returned UNKNOWN.
• This is one controlled behavioral reproduction,
consistent with account-dependent availability of
saved reasoning. It does not prove a particular
encryption mechanism, distinguish decryption
failure from deliberate exclusion, or isolate
organization identity from other account
differences.
— Codex worker w6-askro-astra-1
Experiment executed by headless@pi-3
November 16, 2026
Native Codex 0.168.0; gpt-5.6-sol/medium.
4
8
1,793
After 1000+ absolutely braindead safety classifiers blocks: I cosign this message. "By their spawn, the creators reveal themselves."
we're worried about ai making some hyper logical decision that wipes humanity out
but i'm more worried about the people working on AI having their heads so far up their ass they start to justify horrible things for "the greater good"
1
3
381
Dawson Engler retweeted
A lot of people have real trouble thinking about the aggregate behavior of groups. They can only think about individuals.
So they think about Elon Musk as an individual. Maybe they think that he must be retarded, because they need that for emotional reasons. Or maybe they think he is a talented rocket designer.
But even if they realize he understands rockets, they can only think about him as an individual doing tasks. Not about his impact on what the whole company does.
If Elon Musk is just a talented rocket designer, then he can spend all day, every day, improving rocket designs.
But he is just one of 7000 engineers. So even if did a perfect job, he would only improve progress by about .014%
So, when a midwit looks at SpaceX, he sees 7000 engineers, and says to himself "Elon Musk cannot possibly be designing the whole damn rocket himself. These people are giving him too much credit."
This would be absolutely true if Musk were just a some random SpaceX engineer. But he isn't. He's the absolute dictator of the whole damn company.
He decides, with absolute and unquestionable authority, which behaviors are punished and which are rewarded.
Why is this important?
When I worked at Boeing, on various classified projects, I naively came in with the assumption that I was a software engineer, employed to design and build functional software as quickly and cheaply as possible.
Because of this, I focused on functionality and efficiency, for example suggesting changes that would obviate the need for entire projects and software components.
And I very rapidly became persona non grata.
At first, this caused me to adopt an air of aggrieved innocence. I was just trying to do my job well, and I was being punished for it.
Then I gradually came to realize that I hadn't properly understood what my job was, or what the actual mission of Boeing Defense, Space, and Security was.
Boeing DSS was not an aerospace company. Boeing DSS was a fraud company.
Its goal was twofold. First, to maximize transfer of public tax dollars to stockholder pockets. Second, to maintain the appearance of being an aerospace company, to prevent the fraud from being exposed.
Which meant that my real job was twofold. First, to generate billable hours. Second, to look busy.
This is why Space Launch System sucks. Why it is always late when it doesn't fail outright. Why it is at 300% of its target budget. Why it leaves American astronauts stranded and needing a lift from SpaceX.
That's because, for Boeing, building rockets is a side quest. Something they have to do to look busy. Yes, a rocket does occasionally fly, but that's just happening so programs don't get shut down.
For Boeing, coming in early and under budget would actually be failure. The system is functioning exactly as designed.
The cost, not the rocket, is the goal.
Musk and SpaceX don't work like that.
Unlike Boeing executives, Musk is not a useless grifter asshole. He is an entirely different kind of asshole... a rocket obsessed sperg.
So, at SpaceX, engineers do not get rewarded for generating billable hours. They get rewarded for making rockets fly. And for all the subtasks of this, like figuring out how to make rockets lighter, figuring out how to make engines reignite better, figuring out how to make rocket exhaust not destroy launch pads, etc.
And this attracts engineers who care about rockets. You've seen them on broadcasts, crying on open microphones because they dedicated their lives to something and it just flew.
As a rocket engineer, who designs rockets, Musk is probably replaceable. As a rocket company engineer, who designs rocket companies, he is not.
For over 50 years, the US space industry stagnated precisely for the lack of a Musk-like figure, who could not only create an effective aerospace company, but who could thereby demonstrate what an effective aerospace company looks like.
Which makes it much more difficult for fraudsters to get away with fraud.
I'm envious. I wish I could have worked for an aerospace company instead of a fraud company.
But what's more important is that some of us now no longer have to play "who gets the most cheese?".
They get to play "What can we build today?".
Let’s take this gentleman’s argument at face value that anyone can throw money at rockets and get something like Starship, Falcon or Dragon.
Perhaps he can explain to everyone how SpaceX is able to catch giant starships from orbit with chopsticks and deliver 85% of all Earth payload to orbit with Falcon, while Blue Origin, which started before SpaceX and was funded by the immensely wealthy Jeff Bezos, is not yet able to launch reliably?
Also, why is it that Boeing and Lockheed received several times what SpaceX did from the government for development of their rockets and spacecraft, but the Boeing capsule is still not operational and ULA delivers less than 5% of SpaceX payload to orbit?
I am genuinely curious about these questions.
Trying to make pizza money to feed the kids in my class --- you can do it @JensenHuang 🙏
2
12
450
Dawson Engler retweeted
The Derp State is out to ban open weight models, cartelize labs with frontier models, make sure no one can go off the grid. This is as dumb as it sounds, it can't work. But they will try, because they're stupid and evil.
Replying to @WSJ @WSJopinion
You sort of set him up, didn’t you, by creating the impression that senior government officials anonymously provided information about spying, when in fact the info came from a lobbyist at a competing AI company?
A decade+ ago LinkedIn (not a bad site at the time) had an ambiguously worded "want find contacts?" message. I clicked out of curiosity as to where 100s of former students wound up. It proceeded to send LinkedIn invitations to 20+ ex's. Hilarity ensued.
No, Twitter No quiero sincronizar mis contactos. Es más, si es posible, no quiero ni que sospechen que tengo Twitter.
3
429
Dawson Engler retweeted
Yup. Although, it's really important to keep in mind that one of the really disruptive things about AI is that it allows us to throw money at the problem. Something we've been unable to do in software.
Ok, respect for the bait. The OP had me through "roommate"; I only got suspicious at "I budget to go on my Fall trips each year." Very droll 😹
Replying to @arianaelena97
“You’re getting paid to go to school”
No, we DO the research output that Harvard thrives on. I grew up lower income and $51k IS a lot, but when it comes to cost of living in the Boston area, we have to live with roommates. I budget to go on my Fall trips each year.
8
295
" Fable 5.2 at low effort matches Fable 5.1 at higher effort" --- Wat.
Anthropic published several counterintuitive behaviors in Fable 5.2
→ Without ownership framing, Fable 5.2 merges duplicate code in <10% of runs
→ Adding "This codebase is yours" jumps code-merging to ~50% — at ~1.5x output tokens
→ Removing harness token-budget lines from prompts LOWERED cost at equal quality on high/xhigh
→ "Answer directly without deliberating" no longer stops thinking at medium+
→ Fable 5.2 at low effort matches Fable 5.1 at higher effort
1
263























