
Pinned Tweet
As someone with a son on the spectrum.
Every kid on the spectrum is special, this isn't taking away from any of them.
People are born somewhere on this range, and luck/god determines where they land, but some parents have a narrow window to guide their kids into the Brilliance Zone through early intervention, diet, and environment.
That's the sweet spot where the likes of Elon Musk, Bill Gates, and Vitalik live.
Autistic enough that the brain is wired differently, but not so autistic that you can't function socially and actually deliver on that brilliance.
had fable 5.2 scan 17,520 papers from the Royal Society's journal (1665 to 1920). the job was to find things credible people reported that still have no explanation.
my fav is from 1749. ship is 569 miles off Cornwall when the crew sees a blue ball of fire the size of a millstone rolling across the surface of the sea from 3 miles out. it reaches the ship, rises straight up about 50 yards away and explodes with a sound like hundreds of cannon going off at once.
the explosion breaks the main topmast into over a hundred pieces and knocks five men down. the ship smells of sulfur afterward. the officer who wrote it up says there was no thunder or lightning before or after.
the same journal has three more reports like it between 1685 and 1772, from different places and different witnesses.
3
10
883
Swarm riders and Swarm-native companies
AI-use even among highly AI-native people is power-law distributed.
There seems to be a new class of people — call them “swarm riders” — who are capable of consistently commanding large swarms of agents to productive ends.
This is a new kind of managerial skill, analogous to the skill it takes to manage large organizations of people, but wholly different in nature and of course more technical.
Based on the swarm riders we know (including many at @Tempo and @Paradigm), these people are capable of productive token spend that is 1000x-10000x higher than median software engineers, and 100x higher than top 5% Claude Code users.
The art of swarm riding is different from software engineering and more akin to machine learning, where practitioners grapple with black box systems: defining reward functions and evals, transmuting non-verifiable problems into verifiable problems, running experiments, etc.
Swarm-native companies are a new class of company founded by and populated by swarm riders. These companies are organized around swarms in a way that conventional companies will struggle to replicate, as the cost of internal human coordination is sand in the gears of an effective swarm. Effective swarm-native companies will define the right reward functions and evals, remove functional and bureaucratic barriers, and then let the swarms cook.
We’ve gotten to see a few swarm-native companies up close — reshaping financial rails @Tempo, building a new gaming platform @Spawn, cyber swarms at Cathedral Systems — and it’s clear these teams wield a significant advantage over pre-AI teams.
Some questions I’ve been wondering about:
1) How fast will swarm riding diffuse through the economy? The implications for token spend could be significant.
2) Can swarms be effectively adopted by existing companies or will it generally track the formation of new swarm-native companies?
3) What becomes scarce as swarm capabilities themselves commoditize (through frontier models becoming better at swarming, swarm capabilities built into harnesses, etc)? Cultures and teams that can define reward functions well and otherwise get out of the way?
328
some learnings integrating AI and agents inside businesses:
- configure marking an agent’s task successful only when the intended result is verified, even if the agent has finished responding.
- have research agents report how much evidence they reviewed, because accurate citations can still support an incomplete answer.
- measure usage across worker agents as well as the main session conversation to understand the system’s running costs.
This article articulates the issue with enterprise AI so clearly: the main reason AI pilots fail isn't even because of the AI.
Most enterprises have piloted 3 or more AI projects in the last 12 months. Yet, you won't find any of them on their P&L. Why?
- They built the AI model based on the SOP, but real processes have exceptions that can be 40% of the volume.
- They thought a quicker W-9 review would improve vendor onboarding. It didn't. That process still takes 22 days.
- There wasn't a clear owner for the AI agent 7 months in, after the innovation budget ran out and the consultants left.
Fantastic article by the former COO of Citadel Securities in the EMEA region. If you've piloted AI projects that never made it to your P&L, read this before piloting another.
2
6
550
testing a new work mode
fable 5.2 for design / instructions
-> pass the md to astra for implementation
not sure how this does usage wise. will report back.
1
4
626
Rani Haddad retweeted
coming out of our massive connectivity model a few things we worked on a few interesting sub models on household decision making for starlink and satellite density.
these are available to the public to play with here.
research.33fg.com/analysis/h…
ai slop > human slop
and i don’t think it’s close
A huge amount of the Anti-AI code sentiment massively overestimates the quality of human code outside of a very small set of open source and high quality company codebases. Human Slop is everywhere and can trivially be improved on by any opus level model.
371
Rani Haddad retweeted
our weekly research discussion at 11am ET today. going live soon. we'll be touching on all the starlink mobile rumors and announcements, compute news and how politics may effect that. join us and ask questions!
- currently on a plane
- rex server on a mac mini back home
- using the rex ios app to work on 5 different pi sessions and 1 claude code session (different projects)
- tailscale helps with viewing local drafts of things on my phone
no way how cool/powerful this setup is.
Latest updates on Rex beta rollout:
- I'll have a formal update next week (blog post).
- We've invited 12,000 more this week.
- We've fully opened iOS testing to everyone invited
- Server (Linux + FreeBSD) testing is up to ~2,000 and we're going to open that up to everyone invited soon.
- We're just shy of 50% through out full waitlist, but we're now inviting thousands every day and growing and this should clear quickly.
4
8
732
How AI Power Users Work
Why coordinating agents in parallel changes what one person can get done.
gidorah.substack.com/p/how-a…
kinda weird rebels/jedi blew up TWO Death Stars full of millions of people including children, and then celebrated.
The Empire were the good guys. At minimum, the Jedi were the bad ones, and it isn't close.
The original trilogy runs on Saturday-morning-cartoon morality: space wizards good, planet-killer bad. Then Lucas came back as an older man to explain why the most gifted Jedi alive would walk away. In doing so he accidentally made three movies of evidence for the defense.
Exhibit A: the Republic. Its Senate was too paralyzed to stop a trade blockade of one of its own worlds, and nothing passed until every faction got its cut. The Jedi spent those years running errands for it like galactic rent-a-cops. Meanwhile a Sith Lord climbed to the top of the government they were supposed to be guarding. No wonder Anakin, told his politics sounded like a dictatorship, shrugged: "Well, if it works."
Exhibit B: how they treated him. He was a boy conceived by the Force, the Chosen One of their own prophecy, and the Council didn't want him, mostly because he scared them. Qui-Gon won him in a bet and left his mother in slavery. In ten years the Order never went back for her. They forbade him to marry, under a code that treats love as a character flaw. They gave him a Council seat, denied him the rank of Master in the same breath, then asked him to spy on the Chancellor. When he came to Yoda terrified of losing someone, the advice was to let go of everything he feared to lose. He came back from Tatooine having done something unspeakable, and his teachers never even noticed.
Exhibit C: the Code. It was sacred every time it cost Anakin something. The moment it protected Palpatine (disarmed, beaten, owed a trial), Mace Windu decided he was too dangerous to live. This was the same Mace who had floated the idea of the Council taking control of the Senate. They threw the Code out the window the second it suited them, which is how Mace ended up out the window himself.
Exhibit D: everything else. "Only a Sith deals in absolutes" is an absolute. They took babies from their parents. They were peacekeepers who led a cloned army they never ordered and never questioned. They revered the prophecy until it got inconvenient, then decided it had been misread. Their plan to finish the Sith was to send a son to kill his father. Luke refused, and what actually saved the galaxy was the attachment the Jedi spent generations forbidding. Even Luke ended up calling their legacy failure and hypocrisy.
Anakin wanted what a lot of young men want: a wife, a family, a chance to be a hero and fix what was broken. The Republic offered him celibacy, suspicion, and errands. Palpatine offered a way to save the woman he loved. That's why this story hits so hard for so many young guys. The tragedy isn't that Anakin got corrupted. It's that he chased the things they want too, and lost everything anyway.
Not that anyone wins. Vader becomes a cyborg and Palpatine a cryptkeeper. Obi-Wan spends nearly twenty years in a desert hut, and Luke follows the Jedi path to an island where he drinks green milk from a sea monster. The New Republic inherits every problem the old one had.
almost fell for a sophisticated @google phishing attack. i have a cybersecurity background, and it still nearly got me.
i got a call from “google support.” i was skeptical, but curious to see what tactics scammers are using these days, so i played along.
they opened with routine verification questions. someone in toronto had supposedly tried to access my account from an unknown number.
i asked them to prove they were google. they told me to go to support.google.com and look up my case file, and said they’d sent me an email too.
now this is where i was thrown off.
the email came from a legitimate google domain. the support case was genuinely listed on support.google.com.
then came a few questions about my phone number. and finally, the kill shot:
“can you put in your 2fa code on the keypad?”
i hung up.
@grok can we with all certainty confirm there were no kids on the death star? surely there were workers and other contractors as well? what is the estimated breakdown of type of personnel.
pretty nuts that they didn’t notice anything for almost 3 hours.
there was persistent access.
The largest crypto exchange in the world just experienced the largest robbery in human history—and the person running it is a nomadic autistic man in Dubai posting “Funds are SAFU!”
What a time to be alive.
3
6
825
did a bit of sleuthing on this because it’s so fucking crazy.
so attacker had 50 signatures authorizing malicious code to act as Binance 14, which also owned a separate smart wallet they used to move 823M USDC.
3 possible attack vectors:
We are working non-stop to resolve this issue. Binance will cover all losses.
As a precaution, all deposits and withdrawals are temporarily suspended while our team conducts security checks, freezes funds, and works with law enforcement.
We are working around the clock and will share a full post-mortem once the investigation is complete. Please rely only on official Binance channels for updates.
This is an industry-wide effort, and we are grateful to partners across the ecosystem who have moved quickly to help freeze affected assets.
We will continue to share verified updates through official Binance channels. User funds are safe.
1
627
























