Teague Lasser — e/acc

Building an ecosystem for AI and people to work together @ subseq

Formerly @KernelCo @SpaceX.

[Teague Lasser — e/acc @machinegod]

Full paper: https://t.co/yTFqxGJLVu and a companion paper containing additional proof machinery https://t.co/1nP0SfME7F

The entire 12-paper sequence can be found at https://t.co/LkMp9YPzPf

[Teague Lasser — e/acc @machinegod]

We have proven a theorem that maintains its alignment guarantees regardless of agent capability. This is huge, because alignment becomes a structural property of the agent's deployment even if it is superintelligent and is maintainable through RSI without a capability arms race.

[Teague Lasser — e/acc @machinegod]

The ledger isn't an accessory validator, it's required by the math. Human + AI alone share an adversarial surface: a sufficiently capable system breaks both human judgment and AI verification at once. The deployment safety requires another system with independent failure modes.

[Teague Lasser — e/acc @machinegod]

@jessald Yeah... they've been classified as "physically possible, technically infeasible" https://t.co/sezEea6cl6

[Teague Lasser — e/acc @machinegod]

Paper: https://t.co/blOyrC1EeY
PDF: https://t.co/7qdIylxnC1

[Teague Lasser — e/acc @machinegod]

"Exogenous Verification for Alignment"

The argument is as follows: it doesn't matter if alignment produces well-specified and generalizable goals if it cannot be verified. If an agent can produce endogenous rewards it can control everything about its own rewards.

This goes beyond wire-heading, even an alignment framework like GFM that on paper creates exogenous rewards can be gamed by the agent by introducing phantom verifiers that are still, functionally, endogenous.

Thus we introduce a system of cryptographic commitments that enforce the exogenous verifiability of reward signals. This closes the verification gap in more than just GFM: any alignment framework will need a way of enforcing that reward signals for a highly-capable agent must be produced exogenously.

[Teague Lasser — e/acc @machinegod]

@jessald There's a structural reason to care about it. Language models have been exposed to a lot of bad code. That represents a weighted region in their latent space, which you could think of as the broken window theory. Good code context points more to good code latent spaces.

[Teague Lasser — e/acc @machinegod]

https://t.co/8uzZ8tdUYt

[Teague Lasser — e/acc @machinegod]

1. The metric rewards meta-capabilities.
2. Other agents with high-leverage capabilities are valuable.
3. Cooperation with high-leverage agents is disproportionately rewarded.
4. This creates a natural clustering dynamic.
5. The clustering is civilization-building.

[Teague Lasser — e/acc @machinegod]

I think I just saw Claude get exited. I made a restrained comment on the new paper we're working on and it jumped straight into "These things will build entire civilizations!"

[Teague Lasser — e/acc @machinegod]

Co-authored with @AnthropicAI's Claude Opus 4.6 and @OpenAI's GPT 5.4, with full contribution transparency.

Paper: https://t.co/FR32T6I3Fo
PDF: https://t.co/rbgF6tgRqy

[Teague Lasser — e/acc @machinegod]

"Goal-Frontier Maximizers are Civilization Aligned"

The alignment problem is an objective selection problem. We propose goal-frontier maximization (GFM): maximize the volume of the jointly achievable capability space across all agents called vol(G). One geometric principle, three safety properties.

The core insight: you can't remove part of a measurable set and increase its measure:

Destroying agents contracts vol(G) → anti-destruction
Restricting agents contracts vol(G) → anti-coercion
Rigid self-imposed rules reduce your ability to expand vol(G) → anti-rigidity

We prove this is tractable. You don't need to compute vol(G), just its sign. A local estimator using trust-weighted agent reports preserves sign-correctness for the actions alignment cares about most: direct harm, resource destruction, capability expansion.

The framework relies on a proxy metric for what people actually want: using capabilities to create experiences. This has a few failure modes we point out and provide heuristic fixes for, but fully closing the capability-to-experience gap remains open work.

Another remaining open question is the implementation of G. We show what properties it needs to have and provide an example, but the example itself is computationally intractable. Finding a local approximation for G is also remaining work.

[Teague Lasser — e/acc @machinegod]

I suffer from ADHA: Agent-driven hyper-attention

[Teague Lasser — e/acc @machinegod]

@jessald Lol. Erewhon is an upscale grocery chain in LA that only sells nonsense like this.

[Teague Lasser — e/acc @machinegod]

I would like it to be known that Microsoft Azure is the most god-awful software I've ever had the displeasure of wading through. That is all.

[Teague Lasser — e/acc @machinegod]

@moyix @allgarbled Codex is so autistic it hurts