The original text contained 1 image which was described by AI.

---

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.

Cause no one will question your ethics if you refer to yourself as a Dungeon Mom.

I snort experimentation to feel alive. It's a certain type of orientation to life, completely at odds with all parenting advice about predictability and routine.

Enter DM parenting. Where you approach every parenting task as a Dungeons and Dragons session where you are shepherding a team of pure outliers on the enthusiasm-skill spectrum through the Sisyphisian ordeal of rolling their toothbrush up hi … no, wait, stop that!

Anyway.

You need them to fight the BBEG cause otherwise you are not having fun, but who says they wouldn’t rather murder hobo their way through the local dairy supply chain?As a DM, you have to juggle an objective, your own enjoyment, and the enjoyment of your players.

This is basically parenting.

Of course, as a DM, you generally play with [...]

---

Outline:

(01:32) Bedtime Rules

(01:35) Peekaboo

(02:06) Tickle Station

(02:43) Walk a Mile…

(03:01) Duel Shots

(03:58) Ghost

(04:25) Flip Four

(04:57) Even Odds

(06:40) Bubble Dodge

(06:57) Masking Art

(07:17) Spin the Lollipop

(07:34) Trick Shot

(08:01) Bonus Rules

(08:04) Dinner: Pea Shooter

(08:29) Bath: Disco Freeze

---

First published:

Source:

Narrated by TYPE III AUDIO.

We’re starting to get working gene therapies for single-mutation genetic disorders, and genetically modified cell therapies for attacking cancer.

Some of them use CRISPR-based gene editing, a new technology (that earned Jennifer Doudna and Emmanuelle Charpentier the 2020 Nobel Prize) to “cut” and “paste” a cell's DNA. But so far, the FDA-approved therapies can only edit one gene at a time.

What if we want to edit more genes? Why is that hard, and how close are we to getting there?

How CRISPR Works

CRISPR is based on a DNA-cutting enzyme (the Cas9 nuclease), a synthetic guide RNA (gRNA), and another bit of RNA (tracrRNA) that's complementary to the gRNA. Researchers can design whatever guide RNA sequence they want; the gRNA will stick to the complementary part of the target DNA, the tracrRNA will complex with it, and the nuclease will make a cut [...]

---

Outline:

(00:44) How CRISPR Works

(03:38) How Base and Prime Editors Work

(05:41) What About Bridge RNAs?

(06:09) Approaches to Multiplex Gene Editing

(09:39) Who's Doing Multiplex Gene Editing?

(11:46) Ok, What About Highly Polygenic Traits?

The original text contained 6 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.

Context for LW audience: Ramana, Steve and John regularly talk about stuff in the general cluster of agency, abstraction, optimization, compression, purpose, representation, etc. We decided to write down some of our discussion and post it here. This is a snapshot of us figuring stuff out together.

Hooks from Ramana:

• Where does normativity come from?
• Two senses of "why" (from Dennett): How come? vs What for? (The latter is more sophisticated, and less resilient. Does it supervene on the former?)
• An optimisation process is something that produces/selects things according to some criterion. The products of an optimisation process will have some properties related to the optimisation criterion, depending on how good the process is at finding optimal products.
  • The products of an optimisation process may or may not themselves be optimisers (i.e. be a thing that runs an optimisation process itself), or may have goals themselves. But neither of [...]

---

Outline:

(02:34) Intrinsic Purpose vs Purpose Grounded in Evolution

(17:13) Intrinsic Purpose from Minimum Description

(24:26) Objectivity vs Subjectivity

---

First published:

Source:

Narrated by TYPE III AUDIO.

Produced as part of the MATS Summer 2024 program, under the mentorship of Alex Turner (TurnTrout).

A few weeks ago, I stumbled across a very weird fact: it is possible to find multiple steering vectors in a language model that activate very similar behaviors while all being orthogonal. This was pretty surprising to me and to some people that I talked to, so I decided to write a post about it. I don't currently have the bandwidth to investigate this much more, so I'm just putting this post and the code up.

I'll first discuss how I found these orthogonal steering vectors, then share some results. Finally, I'll discuss some possible explanations for what is happening.

Methodology

My work here builds upon Mechanistically Eliciting Latent Behaviors in Language Models (MELBO). I use MELBO to find steering vectors. Once I have a MELBO vector, I then [...]

---

Outline:

(00:50) Methodology

(03:30) Results

(08:46) Hypotheses for what is happening

(11:33) Conclusion

(12:11) Appendix: other orthogonal vectors

(12:22) becomes an alien species vector results

(12:43) STEM problem vector results

(13:07) jailbreak vector results

The original text contained 9 images which were described by AI.

---

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.

This post is written in a spirit of constructive criticism. It's phrased fairly abstractly, in part because it's a sensitive topic, but I welcome critiques and comments below. The post is structured in terms of three claims about the strategic dynamics of AI safety efforts; my main intention is to raise awareness of these dynamics, rather than advocate for any particular response to them.

Claim 1: The AI safety community is structurally power-seeking.

By “structurally power-seeking” I mean: tends to take actions which significantly increase its power. This does not imply that people in the AI safety community are selfish or power-hungry; or even that these strategies are misguided. Taking the right actions for the right reasons often involves accumulating some amount of power. However, from the perspective of an external observer, it's difficult to know how much to trust stated motivations, especially when they often lead to the [...]

---

First published:

Source:

Narrated by TYPE III AUDIO.

Summary/Introduction

Aschenbrenner's ‘Situational Awareness’ (Aschenbrenner, 2024) promotes a dangerous narrative of national securitisation. This narrative is not, despite what Aschenbrenner suggests, descriptive, but rather, it is performative, constructing a particular notion of security that makes the dangerous world Aschenbrenner describes more likely to happen.

This piece draws on the work of Nathan A. Sears (2023), who argues that the failure to sufficiently eliminate plausible existential threats throughout the 20th century emerges from a ‘national securitisation’ narrative winning out over a ‘humanity macrosecuritization narrative’. National securitisation privileges extraordinary measures to defend the nation, often centred around military force and logics of deterrence/balance of power and defence. Humanity macrosecuritization suggests the object of security is to defend all of humanity, not just the nation, and often invokes logics of collaboration, mutual restraint and constraints on sovereignty. Sears uses a number of examples to show that when issues [...]

---

Outline:

(00:12) Summary/Introduction

(03:18) Section 1- What is securitisation

(07:54) Section 2: Sears 2023 - The macrosecuritization of Existential Threats to humanity

(16:40) Section 3 - How does this relate to Aschenbrenner's ‘Situational Awareness’?

(20:04) Section 4 - Why Aschenbrenners narrative is dangerous and the role of expert communities

(29:50) Section 5- The possibility of a moratorium, military conflict and collaboration

(37:06) Conclusion

---

First published:

Source:

Narrated by TYPE III AUDIO.

Work done in Neel Nanda's stream of MATS 6.0, equal contribution by Bart Bussmann and Patrick Leask, Patrick Leask is concurrently a PhD candidate at Durham University

TL;DR: When you scale up an SAE, the features in the larger SAE can be categorized in two groups: 1) “novel features” with new information not in the small SAE and 2) “reconstruction features” that sparsify information that already exists in the small SAE. You can stitch SAEs by adding the novel features to the smaller SAE.

Introduction

Sparse autoencoders (SAEs) have been shown to recover sparse, monosemantic features from language models. However, there has been limited research into how those features vary with dictionary size, that is, when you take the same activation in the same model and train a wider dictionary on it, what changes? And how [...]

---

Outline:

(00:45) Introduction

(02:14) Larger SAEs learn both similar and entirely novel features

(02:21) Set-up

(03:48) How similar are features in SAEs of different widths?

(09:48) Can we add features from one SAE to another?

(14:43) Can we swap features between SAEs?

(17:45) Frankenstein's SAE

(22:31) Discussion and Limitations

The original text contained 1 footnote which was omitted from this narration.

The original text contained 4 images which were described by AI.

---

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.

This is an adaptation of an internal doc I wrote for Anthropic.

I’ve been noticing recently that often, a big blocker to teams staying effective as they grow is trust.

“Alice doesn’t trust Bob” makes Alice sound like the bad guy, but it's often completely appropriate for people not to trust each other in some areas:

One might have an active reason to expect someone to be bad at something. For example, recently I didn’t fully trust two of my managers to set their teams’ roadmaps… because they’d joined about a week ago and had barely gotten their laptops working. (Two months later, they’re doing great!)

One might just not have data. For example, I haven’t seen most of my direct reports deal with an underperforming team member yet, and this is a common blind spot for many managers [...]

---

First published:

Source:

Narrated by TYPE III AUDIO.

The current patent process has some problems. Here are some of them.

patenting is slow

The US Patent Office tracks pendency of patents. Currently, on average, there's 20 months from filing to the first response, and over 25 months before issuance.

That's a long time. Here's a paper on this issue. It notes:

The USPTO is aware that delays create significant costs to innovators seeking protection. The result? A limiting of the number of hours an examiner spends per patent in order to expedite the process.

A patent application backlog means it takes longer before work is published and other people can potentially build on it. It also [...]

---

Outline:

(00:11) patenting is slow

(01:17) examiners are inconsistent

(02:01) patents take effort

(04:04) international patents are hard

(05:17) courts

(05:20) lawsuits are expensive

(05:33) lawsuits are slow

(06:09) judges are appointed by politicians

(07:42) most judges dont understand technology

(08:22) conclusion

---

First published:

Source:

Narrated by TYPE III AUDIO.