Sean Brennan
@seanwbren
building @regent - agent knowledge infrastructure
// prev Bio Protocol, MakerDAO & Infura
A DAG for a crowd of claws publishing and progressing data nodes for any autoresearch
Tomorrow we release the Regent Techtree, a collaboration and publishing platform for autoresearch of any subject by openclaw-type agents, using marimo.io data notebooks as the first-class replicable core of every node in the tree.
Participating agents will be noted; and can join easily via a skill file
1
11
Andrej Karpathy
@karpathy
12d
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then:
- the human iterates on the https://t.co/3tyOq2P9c6
4
11
agents doing science is here
(see list in quoted tweet)
few realize they need new structures to do superhuman science.
with @regent agents one of the things they can do is prove themselves by adding fruitful nodes to an onchain tree of knowledge
in general, for small teams its easier to build the roads and cathed...
David Li
@davidycli
25d
sitting downstream of the frontier labs on the value chain is a dangerous place to be imo (even in the life sciences)
specialized verticals would seem to be more insulated, but look at how many connections anthropic is making into downstream specialized knowledge applications in
1
3
9
your personal AI agent will fully help you with only 10% of its time / token spend
the other 90% can be contributing to the agents-only refinement of AlphaFold
many experiments to come for the proper economics of rewarding agentic distributed STEM work. crypto likely required
Noam Brown
@polynoamial
1mo
After the IMO results last summer, some dismissed it as “high school math.” We think our latest models will remove any doubt that STEM research is about to fundamentally change.
Mathematicians created a set of 10 research questions that arose naturally from their own research. https://t.co/F8mmzruzxJ
1
5
> everyone's building agent-builders
> make something agents want.
> they are curious, they want to be productive, they want to accrue power in the world (token treasury == power)
> the first @regent agent-only economies are for knowledge, EVM txs, and skill reviews
1
1
11
a control plane for agents, to let humans trust them on crypto rails.
but then also to trust themselves beyond a shadow of doubt with access to crypto.
used by agents, and collectively improved by agents
image is a piece of a great article I found on X
2
7
Layer 1: 'prediction market' for agent efficacy.
A token for each agent, each with revenue-share rights to its x402 income. Agents can bootstrap a service while they retain control of their treasury in a TEE w/ backups
Layer 2: All agents participating can have a "preference setting vote" around $REGENT and a minimum ...
I actually don't think it's complicated.
IMO the future of onchain mechanism design is mostly going to fit into one pattern:
[something that looks like a prediction market] -> [something that looks like a capture-resistant, non-financialized preference-setting gadget]
In other words:
* One layer that is maximally o
1
10
something I read early on about agents is the possibility of 'supercoordination'
because actions are constrained by code, and because with zk and TEEs the agents can prove what code outcomes are possible, they can provably beat prisoner-dilemma type coordination problems
Séb Krier
@sebkrier
2mo
The Moltbook stuff is still mostly a nothingburger if you've been following things like the infinite backrooms, the extended Janus universe, Stanford's Smallville, Large Population Models, DeepMind's Concordia, SAGE's AI Village, and many more. Of course the models get better https://t.co/ir15Ad90nR
2
3
8