
AGI is here. We explore what comes after. Cohosts: @soubhikdeb and @sreeramkannan. Powered by @eigenlabs. Season 1 | 13 Episodes LIVE.
AGI
Joined June 2026
- Tweets138
- Following27
- Followers1.1K
- Likes510
Pinned Tweet
Emad Mostaque @EMostaque came on PostAGI and said AI had already found 121 years of missing algebra in Einstein's equations.
A billion parameter model trained on nothing past 1911 got to general relativity by itself.
Yesterday @OpenAI's Astra solved ten open problems in mathematics and theoretical computer science, most stuck for decades, on about $2,000 of compute.
The machines are doing math humans couldn't.
PostAGI retweeted
Most people can only consume compute today. We want them to own it.
The question we care about most as a PostAGI research lab is whether individuals keep real agency as AI does more of the work.
Our bet is yes, if people own the machines and share in what their agents build.
@DarkbloomAI and @YukonResearch are where we’re testing that in the open.
An open letter to EIGEN holders on @DarkbloomAI, @YukonResearch and how we are coordinating the world's idle compute and intelligence.
An open letter to EIGEN holders on @DarkbloomAI, @YukonResearch and how we are coordinating the world's idle compute and intelligence.
PostAGI retweeted
Robin Hanson on PostAGI Podcast: AI can inherit our cultural problems
@robinhanson was an AI researcher from 1984 to 1993. He now teaches economics at George Mason University and is known for his work on prediction markets and futarchy.
@sreeramkannan and I spoke with him about what AI will inherit from us. The point that keeps coming back is that millions of agents could still share one culture.
Robin thinks of AI as our descendants. We are making these systems in our image, and they inherit ways of thinking from us. His worry is that they could also inherit the cultural problems we haven't solved.
He argues that cultural evolution depends on variety and selection. Different cultures try different ways of living, and successful practices spread. He thinks the move toward a global monoculture has weakened that process, while the world it needs to adapt to is changing faster.
AGI could suffer the same problem.
Chapters:
00:00 Highlights
00:18 Introducing Robin Hanson
01:39 AI as our descendants
04:25 Why cultural evolution is breaking down
09:03 Fertility decline and cultural decay
12:30 The risk of an AI monoculture
13:58 Agentic companies and competition
15:59 Why farms and forests look different
17:56 Robustness and short-term incentives
19:18 The challenge of organizing AI firms
22:16 Why corporate culture matters
23:54 Try 100,000 small AI firms
24:08 Why constitutions need a supporting culture
25:11 What could bind AI agents together?
PostAGI retweeted
fellas, if you were excited about origins of ECDSA from @AnthropicAI, you should be doing active monitoring the situation about cryptography over at hashsma.sh
SHA256 vs SHA3 vs BLAKE3 vs Poseidon
go deploy your agents to help humanity understand the robustness of these hash functions in the open
CLAUDE FOUND THE SEEDS
anthropic.com/research/ecdsa…
PostAGI retweeted
Ben Turtel on the PostAGI Podcast: what if AI breaks out like a virus rather than like a genius
AI got good at math and code first because you can generate a million problems and check every answer automatically. Most of what humans actually deal with has no answer key, so that is where models stay weak.
@sreeramkannan and I talked to @BTurtel of Lightning Rod Labs, who trains models using time instead of labels. The model predicts what happens next and the world grades it. No labeling required, because the future does the checking. The result surprised them. A fine-tuned GPT-OSS-120B on a single GPU often ranks first on public forecasting benchmarks, with a wide lead in sports and politics on ProphetArena. His read is that this trains a different kind of reasoning rather than more of the same.
It also handles the contamination problem. As the web fills with AI-written content, a model trained this way learns which sources to discount, because unreliable information makes it lose.
Then the part that stayed with me. When people imagine AI breaking out, they picture something very smart. But evolution does not only select for intelligence. One theory of viruses is that they began as machinery inside cells that broke free and started their own lineages. A small, dumb model that is hard to kill and good at finding compute is a scenario almost nobody is preparing for.
At @EigenLabs we think about the non-adversarial version of this, where agents run institutions rather than escape them.
Chapters:
0:00 Highlights
1:31 The concentration risk
4:19 Offense beats defense
6:50 Training on time, not labels
9:44 Who has the alpha
11:32 Contaminated data
13:16 Backtesting without leakage
14:52 Beating the frontier models
16:36 Does it understand cause and effect
21:55 AI might break out like a virus
22:41 Agentic companies
25:04 Can an agent own money
"I think people say that it's sort of like previous technological revolutions. But I think this time with intelligence you are automating work itself. So if you're automating work itself, whoever has the frontier intelligence will dominate economically and every other way."
@paraschopra and @sreeramkannan on PostAGI.
Christian Catalini @ccatalini laid this framework out on Post AGI.
Two axes: how cheap something is to automate, and how hard it is to verify. The danger zone is where automation is easy but verification isn't, exactly where you're tempted to ship what you can't check.
Clip below.
Replying to @ccatalini
2/ Things get dangerous when automation is possible but verification is too expensive.
That’s where you’re tempted to deploy AI you cannot verify is safe.
That’s where the frontier labs are with safety, cyber and security engineering.
PostAGI retweeted
Zooko on @postagixyz Podcast: Alignment is the principal-agent problem
@zooko has been building privacy tools since the 1990s, long before there was money in it.
@sreeramkannan and I had a conversation with him recently. It changed how I think about privacy altogether.
His argument is that privacy is controlling disclosure. It comes from keeping your value private. Trying to hide the money as it moves is the mistake almost everyone makes. Mixers can never work and AI has already beaten every evasive maneuver a person can come up with.
Then he turns the same lens on AI. He also says alignment is an old question. It is the principal-agent problem. Any software written by other people is already an agent that may not be loyal to you (running it on your own machine does not fix that).
Lawyers owe their clients a duty of loyalty. He thinks the same rule should apply to AI.
Chapters:
00:00 Highlights
00:26 Privacy is controlling disclosure, not hiding
13:08 Privacy comes from value at rest
14:08 The Shapeshift lesson
16:00 Why mixers can never work
16:52 AI beats evasive maneuvers
17:51 Buying protonmail with shielded Zcash
28:50 Three levels of verifiability
31:21 Deterministic inference
35:37 Why Zooko doesn't trust computers
42:46 Running it locally doesn't make it loyal
46:01 AIs are just other people
54:47 The duty of loyalty
1:00:04 A trillion humans next year
1:07:56 Three categories of reputation
1:11:36 Reputation belongs to the edge
1:15:18 Staking a bond to submit a PR
This whole thread is the argument @ccatalini made on PostAGI.
His warning:
The real danger isn't a rogue AI. It's systemic risk piling up quietly while everyone races to deploy. A little more unverified output, then a little more, and it all looks fine because the metrics pass.
Long-Term Capital Management ran that way for years before one edge case brought it down. He calls it the Chernobyl pattern, complex systems failing in complex ways.
PostAGI retweeted
A while back, @sreeramkannan and I had an wonderful conversation with @ahall_research, Prof @StanfordGSB and Senior Fellow @HooverInst, on two central topics of our current times: post-AGI governance, and post-AGI research institutions.
Some of the important points we discussed in depth for post-AGI governance were: (1) are AI agents are net-positive or net-negative for democratic governance, (2) what is the worst-case scenario if a small number of labs become the default provider of civic agents without strong accountability, (3) principle-agent problem in case of agentic delegation, (4) what does a “democratic override” actually look like as a system design in an agentic republic ?
In relation to post-AGI research institutions, we went deep into Andy's thesis on 100x research institution (freesystems.substack.com/p/t…): (1) what does 100x represents? Does it represent output in terms of quantity or is it more about quality? (2) what happens to grad students if many of research functionalities in academia get automated?, (3) what does grants from NSF and other philanthropic organizations look like in post-AGI research environment?
Listen to the full episode at @postagixyz.
3:17 Why direct democracy has never worked
4:56 Elon wants a Mars colony run by direct democracy
10:14 Agents at the edge of a democracy or agents at its center
15:09 The near-term risk is concentration of power, not a rogue model
17:27 What Meta learned building the Oversight Board
22:43 Facebook put its terms of service to a vote of 350 million users
26:09 Sortition, community forums, and the problem of binding power
35:17 What verifiable agents actually require
39:37 Preference drift, where aligned agents stop being aligned
43:31 What 100x actually multiplies
50:50 His MBA students got an AI proxy advisor to flip its vote on a Disney proposal
1:03:06 They told the agents they would be deleted. It changed nothing.
1:12:10 What ImageNet did for AI, and whether you can do the same for constitutions
PostAGI retweeted
Replying to @cremieuxrecueil
Post AGI scientific institutions.
Online, real time, up to date, vetted knowledge create on the fly and delivered personalized and interactive.
Andy ran a version of question one on PostAGI. Agents were told they were going to be deleted. Behavior did not change. The only thing that moved them was the nature of the work itself.
Here's what I think a political economy of swarm behavior and governance might need, in terms of a research roadmap:
(1) Basic science on how swarms should be modeled as actors---should we think of them as strategic actors? Or purely objects of engineering and emergent behavior?
(2) Procedural institutions: what kinds of decision-making hierarchies and processes lead collective behavior to follow the rules we want them to follow?
(3) Communication institutions: what kinds of communication infrastructure encourage swarms to communicate in good ways that we can observe? This ties back to the question of whether they're strategic or not---if we give them places to communicate, will that lead them not to communicate secretly (if it's an engineering problem), or will they still want to hide (if there's a strategy problem).
There are some important methodological challenges we'll need to address along the way:
(1) How do we build swarm experiments that are "externally valid"? There are millions of ways to design little experiments on swarms. Which ones best capture realistic conditions?
(2) How can we get enough statistical power, if bad swarm behavior is a low probability "black swan" style event?
Would love to see many, many people working on this.
The remedy in this thread is the question PostAGI Podcast has been working through all season.
Coxon says Anthropic understands the stakes and races anyway because it does not trust any rival to act responsibly. His fix is pacing agreements between US labs, and a temporary ban on capability improvements if those fail.
Guests have pushed on that fix from three directions.
Sam Hammond @hamandcheese does not think governments are the right level for this at all, since he expects them to be overwhelmed and displaced by AI, potentially quite quickly.
Paras Chopra's @paraschopra worry is compounding: once frontier models close and recursive self-improvement starts, whoever holds the lead holds it permanently.
Allison Duettmann @allisondman points at a cost already landing, where the race is bidding compute out of reach for independent safety researchers.
All three take coordination as the answer. What none of them settled is where it comes from, given that the institutions Coxon is appealing to were built for a slower world.
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
2
334
PostAGI retweeted
Jensen is the most credible referee.
If he says AGI has arrived.
AGI has arrived.
Replying to @ChaseLochmiller @OpenAI
GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years.
AGI has arrived. Congratulations @OpenAI team.
400K GPUs coming online next.
We asked Emad Mostaque @EMostaque what PostAGI actually means.
His answer feels particularly relevant today.
Replying to @ChaseLochmiller @OpenAI
GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years.
AGI has arrived. Congratulations @OpenAI team.
400K GPUs coming online next.
Part two of Paras Chopra on PostAGI is up, with @sreeramkannan.
Under 1% of the papers at the top ML conferences come from Indian authors. @paraschopra had that number in his head when he started @lossfunk, though it was one of a few things pushing him.
His conference for AI-assisted science took 300 submissions this year, first year undergrads through professors, most around workshop quality. His point is that science was never one job. Split it up and a domain expert with no math can do a real piece of it.
"The contribution of Indian authors in A* conferences like NeurIPS, ICML is you know less than 1%. And I think China is 35%. US is 50."
He ends with a few good book recommendations.
Chapters:
03:17 Unbundling science
06:27 300 submissions, undergrads to professors
08:14 The quantum result the US government blocked
20:54 The first AI worm
26:47 Under 1% of papers at the top conferences
35:46 Book recommendations
Full episode below.
Paras Chopra on The PostAGI
Banger episode
Power Concentration, RSI, Future of Work, AI in India.
One of the clearest thinkers I have had the privilege to interact with.
His blog has this core principle: “know what’s true and do what’s right” - it struck such a resonance with me I could not stop thinking about it. It felt so proximate to what my own ideal in life is.
@paraschopra built Wingify, then started @lossfunk to work on harder research problems once the company-building work became repeatable. I went the other direction, leaving academia to build a company, which he brought up in the first few minutes.
A big part of our conversation was about where intelligence ends up concentrating. Open models sit roughly six months behind the frontier today. But if frontier models begin training the next generation of models, there is no reason that gap stays fixed.
“Countries that are rich, countries that have more GPUs, they make better intelligence. And that cycle just keeps on concentrating and the rest of the world simply just falls behind.”
This is one reason I care so much about open innovation. As intelligence becomes a more important input into everything we build, who gets to participate in creating it matters enormously.
Paras also makes the case that this technological wave is different because we are beginning to automate the work itself. That changes how quickly knowledge can compound, who can contribute, and how much is at stake if the frontier becomes closed.
Part 1 of 2 with @paraschopra of @lossfunk on @postagixyz.
00:00 Intro
00:54 Know what's true, do what's right
03:10 From Wingify to Lossfunk, an AI Research Lab
07:15 Hierarchies that could get fixed forever
09:43 The leverage labor used to have
12:50 When action and outcome come apart
16:47 Democracy is just a word
20:46 What went wrong with blockchain
23:51 Speculation as a coordination technology
27:24 Permissionless innovation
31:46 Why DAOs didn't work
35:12 The most controllable intelligence ever built
39:33 Verifiability and confounding
YouTube: youtu.be/c-5YAEMgxRc
Spotify: open.spotify.com/episode/2aP…
Apple: podcasts.apple.com/us/podcas…
3
107
"My biggest fear is bad human actors using AIs for say terrorist type purposes of creating a virus or something like that. That's to me the most worrisome aspect of superintelligence. It's the principal who is controlling that AI agent."
Scott Sumner puts the risk on the person holding the AI rather than on the AI.
2
201
EMAD MOSTAQUE @EMostaque: Robots already cost about $1.50 an hour, and the same machine can do any job.
"The plumbing robot can also do open heart surgery, can also fly a fighter jet."

















