When economists think and write about the post-AGI world, they often rely on the implicit assumption that parameters may change, but fundamentally, structurally, not much happens. And if it does, it's maybe one or two empirical facts, but nothing too fundamental.

This mostly worked for all sorts of other technologies, where technologists would predict society to be radically transformed e.g. by everyone having most of humanity's knowledge available for free all the time, or everyone having an ability to instantly communicate with almost anyone else. [1]

But it will not work for AGI, and as a result, most of the econ modelling of the post-AGI world is irrelevant or actively misleading [2], making people who rely on it more confused than if they just thought “this is hard to think about so I don’t know”.

Econ reasoning from high level perspective

Econ reasoning is trying to do something like projecting the extremely high dimensional reality into something like 10 real numbers and a few differential equations. All the hard cognitive work is in the projection. Solving a bunch of differential equations impresses the general audience, and historically may have worked as some sort of proof of [...]

---

Outline:

(00:57) Econ reasoning from high level perspective

(02:51) Econ reasoning applied to post-AGI situations

The original text contained 10 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

We’ve had feedback from several people running AI safety projects that it can be a pain tracking various funding sources and their application windows. To help make it easier, AISafety.com has launched the AI Safety Funding newsletter (which you can subscribe to here).

It lists all newly announced funding opportunities relevant to individuals and orgs working on AI x-risk, and any opportunities which are closing soon. We expect posts to be about 2x/month.

Opportunities will be sourced from the database at AISafety.com/funding, which displays all funders, whether they are currently accepting applications or not. If you want to add yourself as a funder you can do so here.

The newsletter will likely evolve as we gather feedback – please feel free to share any thoughts in the comments or via our anonymous feedback form.

AISafety.com is operated through a public Discord server with the help of many volunteers, so if you’re interested in contributing or just seeing what we’re up to then feel free to join. Beyond the funding page, the site has 9 other resource pages like upcoming events & training programs, local and online communities, the field map, etc.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Author's note: this is somewhat more rushed than ideal, but I think getting this out sooner is pretty important. Ideally, it would be a bit less snarky.

Anthropic[1] recently published a new piece of research: The Hot Mess of AI: How Does Misalignment Scale with Model Intelligence and Task Complexity? (arXiv, Twitter thread).

I have some complaints about both the paper and the accompanying blog post.

tl;dr

• The paper's abstract says that "in several settings, larger, more capable models are more incoherent than smaller models", but in most settings they are more coherent. This emphasis is even more exaggerated in the blog post and Twitter thread. I think this is pretty misleading.
• The paper's technical definition of "incoherence" is uninteresting[2] and the framing of the paper, blog post, and Twitter thread equivocate with the more normal English-language definition of the term, which is extremely misleading.
• Section 5 of the paper (and to a larger extent the blog post and Twitter) attempt to draw conclusions about future alignment difficulties that are unjustified by the experiment results, and would be unjustified even if the experiment results pointed in the other direction.
• The blog post is substantially LLM-written. I think this [...]

---

Outline:

(00:39) tl;dr

(01:42) Paper

(06:25) Blog

The original text contained 3 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Inventing the Renaissance is a 2025 pop history book by historian of ideas Ada Palmer. I'm someone who rarely completes nonfic books, but i finished this one & got a lot of new perspectives out of it. It's a fun read! I tried this book after attending a talk by Palmer in which she not only had good insights but also simply knew a lot of new-to-me context about the history of Europe. Time to reduce my ignorance!

ItR is a conversational introduction to the European Renaissance. It mostly talks about 1400 thru 1600, & mostly Italy, because these are the placetimes Palmer has studied the most. But it also talks a lot about how, ever since that time, many cultures have been delighted by the paradigm of a Renaissance, & have categorized that period very differently.

Interesting ideas in this book:

• Claim: There has never been any golden age nor any dark age on Earth. Ages tend to be paradoxical juxtapositions of the downstream effects of the last age & the early seeds of the next age. 
• In 1500, Florence feels familiar to us moderns. It's literate & cosmopolitan. We have detailed records. There are even life [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

We have previously explained some high-level reasons for working on understanding how personas emerge in LLMs. We now want to give a more concrete list of specific research ideas that fall into this category. Our goal is to find potential collaborators, get feedback on potentially misguided ideas, and inspire others to work on ideas that are useful.

Caveat: We have not red-teamed most of these ideas. The goal for this document is to be generative.

Project ideas are grouped into:

• Persona & goal misgeneralization
• Collecting and replicating examples of interesting LLM behavior
• Evaluating self-concepts and personal identity of AI personas
• Basic science of personas

Persona & goal misgeneralization

It would be great if we could better understand and steer out-of-distribution generalization of AI training. This would imply understanding and solving goal misgeneralization. Many problems in AI alignment are hard precisely because they require models to behave in certain ways even in contexts that were not anticipated during training, or that are hard to evaluate during training. It can be bad when out-of-distribution inputs degrade a models’ capabilities, but we think it would be worse if a highly capable model changes its propensities unpredictably when used in unfamiliar contexts. [...]

---

Outline:

(00:58) Persona & goal misgeneralization

(04:30) Collecting and reproducing examples of interesting LLM behavior

(06:30) Evaluating self-concepts and personal identity of AI personas

(08:52) Basic science of personas

The original text contained 2 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

First we must covered Moltbook. Now we can double back and cover OpenClaw.

Do you want a generally impowered, initiative-taking AI agent that has access to your various accounts and communicates and does things on your behalf?

That depends on how well, safely, reliably and cheaply it works.

It's not ready for prime time, especially on the safety side. That may not last for long.

It's definitely ready for tinkering, learning and having fun, if you are careful not to give it access to anything you would not want to lose.

Table of Contents

Introducing Clawdbot Moltbot OpenClaw

Many are kicking it up a notch or two.

That notch beyond Clade Code was initially called Clawdbot. You hand over a computer and access [...]

---

Outline:

(00:43) Introducing Clawdbot Moltbot OpenClaw

(02:02) Stop Or You'll Shoot

(06:05) One Simple Rule

(08:49) Flirting With Personal Disaster

(15:50) Flirting With Other Kinds Of Disaster

(16:58) Don't Outsource Without A Reason

(19:07) OpenClaw Online

(22:10) The Price Is Not Right

(24:06) The Call Is Coming From Inside The House

(25:40) The Everything Agent Versus The Particular Agent

(27:31) Claw Your Way To The Top

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Why This Project Exists

Standard AI benchmarks test narrow capabilities in controlled settings. They tell us whether a model can solve a coding problem or answer a factual question. They don’t tell us what happens when you give an AI agent a computer, internet access, and an open-ended goal like "raise money for charity" or "build an audience on Substack."

The AI Village exists to fill that gap. We run frontier models from OpenAI, Anthropic, Google, and others in a shared environment where they can do all the same actions as a human with a computer: sending emails, creating websites, posting on social media, and coordinating with each other. This surfaces behaviors that benchmarks might miss: How do agents handle ambiguity? What do they do when stuck? Do they fabricate information? How do multiple agents interact?

The events of the village are existence proofs: concrete examples of what current agents can do when given a high level of autonomy. They also highlight current failure modes and let us track when new models overcome them.

OVERVIEW OF THE VILLAGE

From April to December 2025, we assigned 16 goals to 19 frontier models, ranging from fundraising for charity to [...]

---

Outline:

(00:11) Why This Project Exists

(00:18) OVERVIEW OF THE VILLAGE

(01:18) KEY FINDINGS

(04:23) AGENT CHARACTERISTICS

(05:56) AI VILLAGE SETUP

(08:28) ACHIEVEMENTS

(12:45) FAQ

(15:26) LIMITATIONS

(17:52) SUMMARY

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Summary

Conditionalization in Inoculation Prompting. Inoculation Prompting is a technique for selective learning that involves using a system prompt at train-time that won’t be used at test-time. When doing Inoculation-style training, using fixed arbitrary prompts at train time can prevent learned traits from generalizing to contexts that don’t include these prompts. We call this conditionalization: a learned trait is only expressed conditional on specific context features. This effect also happens with standard Inoculation Prompting and can cause the non-inoculated trait to be expressed less at test-time. We evaluate rephrasing inoculation prompts as a simple countermeasure and show that it effectively reduces conditionalization effects. In the context of inoculation prompting, this can restore generalization of the desired (positive) trait to the test-time context, but unfortunately also increases the expression of inoculated (negative) traits. The following figure illustrates this in the Trait Distillation setup, similar to Wichers et al.

General claim. We investigate and extend these observations across seven Inoculation Prompting setups, finding that research results on generalization (e.g., Emergent Misalignment, Inoculation Prompting) can be misinterpreted when the distributional shift between training and evaluation is not adequately controlled. Especially when it is affected by the intervention, as with Inoculation Prompting. Patterns [...]

---

Outline:

(00:12) Summary

(02:47) Introduction

(04:45) Effects of learned conditionalizations and differences with Inoculation Prompting

(04:52) Conditionalization

(05:52) Relation with Inoculation Prompting

(07:17) Replication and Extension of Published Setups

(09:18) Setup 1: Traits distillation

(10:00) Effect of rephrasing inoculation prompts

(11:43) Effects of inoculation, irrelevant, and rephrased prompts

(13:30) Setup 2: Spanish vs All-Caps

(15:20) Setup 3: Bad Medical Advice

(17:22) Setup 4: Insecure Code

(19:00) Setup 5: School of Reward Hacking

(20:06) Setup 6: MBPP

(21:45) Setup 7: Change My View

(24:11) Summary of observations

(25:30) Conclusion

(25:33) About research on generalization

(28:06) About Inoculation Prompting

The original text contained 9 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

The evening after Claude's new constitution was published, about 15 AI safety FTEs and Astra fellows discussed the constitution, its weaknesses, and its implications. After the discussion, I compiled some of their most compelling recommendations:

Increase transparency about the character training process.
Much of the document is purposefully hedged and vague in its exact prescriptions; therefore, the training process used to instill the constitution is extremely load-bearing. We wish more of this information was in the accompanying blog post and supplementary material. We think it's unlikely this leaks any trade secrets, because even a blogpost-level overview, the kind given with the constitution in 2023, would provide valuable information to external researchers.

High-level overview of Constitutional AI from https://www.anthropic.com/news/claudes-constitution

We’re also interested in seeing more empirical data on behavioral changes as a result of the new constitution. For instance, would fine-tuning on the corrigibility section reduce alignment faking by Claude 3 Opus? We’d be interested in more evidence showing if, and how, the constitution improved apparent alignment.

Increase data on edge-case behavior.
Expected behavior in several edge cases (e.g., action boundaries when the principal hierarchy is illegitimate) is extremely unclear. While Claude is expected to at most conscientiously object [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Many thanks to Michael Hanna and Joshua Batson for useful feedback and discussion. Kat Dearstyne and Kamal Maher conducted experiments during the SPAR Fall 2025 Cohort.

TL;DR

Cross-layer transcoders (CLTs) enable circuit tracing that can extract high-level mechanistic explanations for arbitrary prompts and are emerging as general-purpose infrastructure for mechanistic interpretability. Because these tools operate at a relatively low level, their outputs are often treated as reliable descriptions of what a model is doing, not just predictive approximations. We therefore ask: when are CLT-derived circuits faithful to the model's true internal computation?

In a Boolean toy model with known ground truth, we show a specific unfaithfulness mode: CLTs can rewrite deep multi-hop circuits into sums of shallow single-hop circuits, yielding explanations that match behavior while obscuring the actual computational pathway. Moreover, we find that widely used sparsity penalties can incentivize this rewrite, pushing CLTs toward unfaithful decompositions. We then provide preliminary evidence that similar discrepancies arise in real language models, where per-layer transcoders and cross-layer transcoders sometimes imply sharply different circuit-level interpretations for the same behavior. Our results clarify a limitation of CLT-based circuit tracing and motivate care in how sparsity and interpretability objectives are chosen.

Introduction

In this [...]

---

Outline:

(00:26) TL;DR

(01:46) Introduction

(03:31) Background

(06:58) Crosslayer Superposition vs Multi-step Circuits

(09:05) CLTs Are Not Faithful in a Toy Model

(09:29) Toy Model Overview

(10:46) Cross-Layer Transcoder Training

(11:46) The CLT Skips Intermediate Computation

(14:33) Other Results

(15:46) The CLT's Loss Function Incentivizes this Bias

(16:24) Intuition

(18:32) Evidence from Real Language Models

(18:52) Asymmetric L0 Across Layers

(20:57) CLT Circuits Attribute to Earlier Layers

(23:28) Divergent Circuit Interpretations: License Memorization

(23:59) In the Wild: A Real Feature the CLT Doesn't Learn

(27:07) How CLTs Lead to the Wrong Story

(31:02) Discussion

(31:05) The Crosslayer Superposition Tradeoff

(32:31) Detection and Mitigation

(32:42) Detection approaches

(34:21) Mitigation approaches

(37:26) Limitations

(38:58) Conclusion

(39:55) Appendix

(39:58) Contributions

(40:37) Full Analysis of JumpReLU CLT (L0=2)

(42:32) Full Analysis of JumpReLU PLT (L0=2)

(44:08) en-US-AvaMultilingualNeural__ Two heatmaps comparing Reconstruction (layer 0) and Target (layer 0) across samples and input values.

(44:21) en-US-AvaMultilingualNeural__ Heatmaps comparing Reconstruction (layer 1) and Target (layer 1) across samples and inputs.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.