---

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.

A quick post on a probably-real inadequate equilibrium mostly inspired by trying to think through what happened to Chance the Rapper.

Potentially ironic artifact if it accrues karma.

1. The sculptor's garden

A sculptor worked in solitude for years, carving strange figures in his remote garden. Most of his statues failed: some cracked in winter, others looked wrong against the landscape. But occasionally, very rarely, one seemed to work.

The first visitors stumbled upon the garden by accident. They found themselves stopped by his angels—figures that somehow held both sorrow and joy, wings that seemed about to flitter.

Word traveled slowly. More visitors came, drawn by something they couldn't quite name.

The sculptor felt recognized for the first time. Not famous—but understood. His private work had somehow become communicable. He carved more angels, trying to understand what made these particular statues resonate.

As crowds grew, their attention [...]

---

Outline:

(00:19) 1. The sculptors garden

(02:57) 2. The mirror trap

(06:01) 3. Resisting the trap

(06:31) Brief practical ideas for creators

(07:19) Brief practical ideas for audiences

---

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.

Epistemic status: One week empirical project from a theoretical computer scientist. My analysis and presentation were both a little rushed; some information that would be interesting is missing from plots because I simply did not have time to include it. All known "breaking" issues are discussed and should not effect the conclusions. I may refine this post in the future.

[This work was performed as my final project for ARENA 5.0.]

Background

I have seen several claims[1] in the literature that base LLM in-context learning (ICL) can be understood as approximating Solomonoff induction. I lean on this intuition a bit myself (and I am in fact a co-author of one of those papers). However, I have not seen any convincing empirical evidence for this model.

From a theoretical standpoint, it is a somewhat appealing idea. LLMs and Solomonoff induction both face the so-called "prequential problem," predicting a sequence [...]

---

Outline:

(00:40) Background

(03:35) Methodology

(04:56) Results

(07:21) Conclusions

The original text contained 4 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.

As you know, histograms are decent visualizations for PDFs with lots of samples...

...but if there are only a few samples, the histogram-binning choices can matter a lot:

The binning (a) discards information, and worse, (b) is mathematically un-aesthetic.

But a CDF doesn't have this problem!

If you make a bunch of predictions, and you want to know how well they're calibrated, classically you make a graph like this:

But, as with a histogram, this depends on how you bin your predictions.

Is there some CDF-like equivalent here? Some visualization with no free parameters?

I asked that question to several people at Arbor Summer Camp. I got three answers:

1. "You get from a PDF to [...]

---

First published:

Source:

Linkpost URL:
https://optimizationprocess.com/calibration-cdf/

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.

We know what linear functions are. A function f is linear iff it satisfies additivity _f(x + y) = f(x) + f(y)_ and homogeneity _f(ax) = af(x)_.

We know what continuity is. A function f is continuous iff for all ε there exists a δ such that if _|x - x_0|_ < δ, then _|f(x) - f(x_0)|_ < ε.

An equivalent way to think about continuity is the sequence criterion: f is continuous iff a sequence _(x_k)_ converging to _x_ implies that _(f(x_k))_ converges to _f(x)_. That is to say, if for all ε there exists an N such that if k ≥ N, then _|x_k - x|_ < ε, then for all ε, there also exists an M such that if k ≥ M, then _|f(x_k) - f(x)|_ < ε.

Sometimes people talk about discontinuous linear functions. You might think: that's [...]

---

First published:

Source:

Linkpost URL:
http://zackmdavis.net/blog/2025/06/discontinuous-linear-functions/

Narrated by TYPE III AUDIO.

AISI is being rebranded highly non-confusingly as CAISI. Is it the end of AISI and a huge disaster, or a tactical renaming to calm certain people down? Hard to tell. It could go either way. Sometimes you need to target the people who call things ‘big beautiful bill,’ and hey, the bill in question is indeed big. It contains multitudes.

The AI world also contains multitudes. We got Cursor 1.0, time to get coding.

On a personal note, this was the week of LessOnline, which was predictably great. I am sad that I could not stay longer, but as we all know, duty calls. Back to the grind.

Table of Contents

---

Outline:

(00:54) Language Models Offer Mundane Utility

(05:00) Language Models Don't Offer Mundane Utility

(08:04) Language Models Could Offer More Mundane Utility

(12:13) Huh, Upgrades

(16:20) Fun With Media Generation

(22:07) Choose Your Fighter

(30:14) Deepfaketown and Botpocalypse Soon

(31:41) Get My Agent On The Line

(35:51) They Took Our Jobs

(48:41) The Art of the Jailbreak

(51:15) Unprompted Attention

(59:14) Get Involved

(59:39) Introducing

(01:01:18) In Other AI News

(01:04:04) Show Me the Money

(01:07:57) Quiet Speculations

(01:14:10) Taking Off

(01:15:47) Goodbye AISI?

(01:21:27) The Quest for Sane Regulations

(01:29:50) Copyright Confrontation

(01:32:56) Differential Access

(01:35:37) The Week in Audio

(01:36:17) When David Sacks Says 'Win the AI Race' He Literally Means Market Share

(01:37:50) Rhetorical Innovation

(01:41:44) Aligning a Smarter Than Human Intelligence is Difficult

(01:45:49) Misaligned!

(01:52:41) People Are Worried About AI Killing Everyone

(01:54:16) The Lighter Side

The original text contained 2 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.

Epistemic Status: Attempt to codify new term.

I want to define a new term that will be useful in the discourse.

Imagine you landed in a foreign country with dogs, but with no word for dogs. You would see these things, and want to talk about them. You would like to be able to tell people whether they can bring dogs into your house, but you can’t, because there isn’t any word for “dog.” You’d like to ask if you can pet someone's dog, but you can’t do that either, because when you ask “can I pet your dog?” they say “pet my what?” It would get pretty frustrating.

This is how I feel about a certain phenomenon I’ve noticed over the last year or so. I notice it all the time—It's as ubiquitous in society as dogs are, and I see it just about as often—but there [...]

---

First published:

Source:

Narrated by TYPE III AUDIO.

Previously: #1, #2, #3, #4, #5

Dating Roundup #4 covered dating apps. Roundup #5 covered opening without them.

Dating Roundup #6 covers everything else.

Table of Contents

---

Outline:

(00:26) You're Single Because You Can't Handle Basic Logistics

(01:39) You're Single Because You Don't Ask Questions

(04:18) You're Single Because of Your Terrible Dating Tactics

(07:49) You're Single Because You Refuse to Play Your Role

(13:24) You're Single Because People Are Crazy About Age Gaps

(13:54) You're Single and You Need Professional Help

(16:20) You're Single Because You Never Close

(21:49) You're Single Because You're Bad at Sex And Everyone Knows

(26:21) You're Single Because You Are Only a Fan

(31:10) You're Single Because of Preference Falsification

(36:51) You're Single Because You Have Insufficient Visual Aids

(40:00) You're Single Because You Told Your Partner You Didn't Want Them

(46:25) You're Single Because of Your Terrible Dating Strategy

(49:05) You're Single Because You Don't Enjoy the Process

(50:33) You're Single Because You Don't Escalate Quickly

(52:58) You're Single Because Your Standards Are Too High

(56:36) You're Single Because You Read the Wrong Books

(01:00:05) You're Single Because You're Short, Sorry, That's All There Is To It

(01:01:45) You're Single Because of Bad Government Incentives

(01:04:02) You're Single Because You Don't Realize Cheating is Wrong

(01:05:15) You're Single Because You're Doing Polyamory Wrong

(01:08:00) You're Single Because You Don't Beware Cheaters

(01:09:48) You're Single Because Your Ex Spilled the Tea

(01:13:42) You're Single Because You're Assigning People Numbers

(01:14:54) You're Single Because You Are The Wrong Amount of Kinky

(01:17:06) You're Single Because You're Not Good Enough at Sex

(01:21:00) You're Single But Not Because of Your Bodycount

(01:23:59) You're Single Because They Divorced You

(01:28:11) You're Single Because No One Tells You Anything

(01:31:51) You're Single And You're Not Alone

(01:32:18) You're Single Because Things Are Steadily Getting Worse

(01:36:29) You're Single Because You Didn't Go to College

(01:37:41) You're Single But This Isn't About You

(01:39:33) You're Single so Let's Go to the Videotape

(01:39:43) You're Single Because You Don't Seek Out Good Advice

(01:40:46) You're Single So Here's Some Hope

---

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.

Has someone you know ever had a “breakthrough” from coaching, meditation, or psychedelics — only to later have it fade?

Show tweet

For example, many people experience ego deaths that can last days or sometimes months. But as it turns out, having a sense of self can serve important functions (try navigating a world that expects you to have opinions, goals, and boundaries when you genuinely feel you have none) and finding a better cognitive strategy without downsides is non-trivial. Because the “breakthrough” wasn’t integrated with the conflicts of everyday life, it fades. I call these instances “flaky breakthroughs.”

It's well-known that flaky breakthroughs are common with psychedelics and meditation, but apparently it's not well-known that flaky breakthroughs are pervasive in coaching and retreats.

For example, it is common for someone to do some coaching, feel a “breakthrough”, think, “Wow, everything is going to be different from [...]

---

Outline:

(03:01) Almost no practitioners track whether breakthroughs last.

(04:55) What happens during flaky breakthroughs?

(08:02) Reduce flaky breakthroughs with accountability

(08:30) Flaky breakthroughs don't mean rapid growth is impossible

(08:55) Conclusion

---

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.

Imagine each of us has an AI representative, aligned to us, personally. Is gradual disempowerment solved?[1] In my view, no; at the same time having AI representatives helps at the margin.

I have two deep reasons for skepticism.[2] Here is the first one.

Humans are Not Alone

We, as individuals, are not the only agents or “agencies” in this world. Other goal-oriented strategic entities include states, corporations, and to some extent egregores.

For the sake of argument, imagine the current distribution of agency in the world as approximately 60% humans, 20% corporations, 20% states (numbers are illustrative)

Now, suppose we introduce individually aligned AI assistants for each person. At first glance, this seems empowering—each individual gains a cognitive extension, potentially leveling the playing field.

But: do corporations or states also get their representatives?

If your answer is No, you'd need to explain how individuals gain AI power [...]

---

Outline:

(00:31) Humans are Not Alone

(02:15) The Substrate Shift

(04:17) Coordination and Collective Action

(05:10) The Role of the Leviathan

The original text contained 2 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.