The Revolution of Rising Requirements has many elements. The most onerous are the supervisory requirements on children. They have become, as Kelsey Piper recently documented, completely, utterly insane, to the point where:

Whereas I think that if you don’t allow your 10-year-old to play alone in a park, that is a much better (although still quite bad) potential reason for a CPS investigation.

This is not an idle threat, per the common statistic that around 35% of American families get investigated by CPS. Even if you are confident that will ultimately turn out fine, and given the vagaries and insanities one can never fully be sure, the process is already the punishment.

As Kelsey Piper says [...]

---

Outline:

(01:47) We Don't Let Kids Be Kids

(03:42) Let Your Children Play

(05:36) The Algorithm of Fear

(07:33) The Only Thing We Have To Fear Is Fear Itself

(11:48) No Means No

(14:20) Child Abduction Is Either Custody Disputes Or Government Intervention

(17:29) Stranger Danger Danger

(20:39) The Before Times

(21:33) Could It Be... Murder?

(22:30) The Long Road Home

(22:47) Either Bored Or Happy

(23:53) All Eyes on Me

(27:03) Screens From The Past

(28:50) Medicalization

(31:02) Television

(32:14) Suspiciously Large Effect Size

---

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 key risk factor for scheming (and misalignment more generally) is opaque reasoning ability.

Important caveat: To get human completion times, I ask Opus 4.5 (with thinking) to estimate how long it would take the median AIME participant to complete a given problem.

Here are the 50% reliability time horizon results:

I find that Opus 4.5 has a no-CoT 50% reliability time horizon of 3.5 minutes and that time horizon has been doubling every 9 months.

In an earlier post (Recent LLMs can leverage filler tokens or repeated problems to improve (no-CoT) math performance), I found that repeating the problem substantially boosts performance.

---

Outline:

(02:33) Some details about the time horizon fit and data

(04:17) Analysis

(06:59) Appendix: scores for Gemini 3 Pro

(11:41) Appendix: full result tables

(11:46) Time Horizon - Repetitions

(12:07) Time Horizon - Filler

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.

[This is an entry for lsusr's write-like-lsusr competition.]

"I solved the alignment problem," said Qianyi.

"You what?" said postdoc Timothy.

It was late at the university computer laboratory and Timothy' skepticism was outvoted by his eagerness to think about anything other than his dissertation.

"You heard me," said Qianyi.

"You do realize that solving the alignment has lots of different components, right?" said Timothy, "First you need to figure out how to build a general superintelligence and world optimizer."

"I did that," said Qianyi.

"Then you'd need to align it with the Coherent Extrapolated Volition (CEV) of humanity," said Timothy.

"I did that too," said Quanyi.

"Except CEV is barely even a coherent concept. This isn't even a technical problem. It's a socio-ontological one. If people disagree with each other, then CEV is undefined. And every large group of people on Planet Earth has holdouts who disagree about everything you can imagine. There are subcultures who believe in underground lizardpeople," said Timothy.

"Solved it."

"There's also the problem of falsified preferences. What people say they want and what people actually want. What people say they believe differs from [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Merry Christmas! Today I turn an earlier LW shortform into a full post and discuss "unknown knowns" "obvious" ideas that are actually hard to discuss because they're invisible when you don't have them, and then almost impossible to unsee when you do.

Hopefully this is a fun article for like the twenty people who check LW on Christmas!

__

There are a number of implicit concepts I have in my head that seem so obvious that I don’t even bother verbalizing them. At least, until it's brought to my attention other people don’t share these concepts.

It didn’t feel like a big revelation at the time I learned the concept, just a formalization of something that's extremely obvious. And yet other people don’t have those intuitions, so perhaps this is pretty non-obvious in reality.

Here's a short, non-exhaustive list:

• Intermediate Value Theorem
• Net Present Value
• Differentiable functions are locally linear
• Grice's maxims
• Theory of Mind

If you have not heard any of these ideas before, I highly recommend you look them up! Most *likely*, they will seem obvious to you. You might already know those concepts by a different name, or they’re already integrated enough into your worldview [...]

---

Outline:

(02:04) The Intermediate Value Theorem

(03:54) Net Present Value

(05:24) Differentiable functions are locally linear

(07:01) Grice's maxims

(08:17) The Pig Hates It

(08:20) Theory of Mind

(09:58) Why this all matters

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.

[...]

---

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.

…if it's anything like mine[1]. If it isn't, then maybe treat this as an opportunity to fight the typical mind fallacy by getting a glimpse at what's going on in other minds.

I've made many predictions and have gone through a lot of calibration training over the years, and have generally become pretty well calibrated in several domains. Yet I noticed that in some areas, I seem surprisingly immune to calibration. Where my brain just doesn't want to learn that it's systematically and repeatedly wrong about something. This post is about the three four such areas.

(1) Future Ability to Remember Things

Sometimes, I feel incredibly confident that some piece of information will be easy for me to recall in the future, even when this is entirely false. Some examples:

• When I meet a new person, and they mention their name, then while hearing it, I typically feel like "yeah, easy, it's trivial for me to repeat that name in my head so I will remember it forever, no problem", and then 5 seconds later it's gone
• When I open some food/drink product and then put it into the fridge, I usually take a note on the product [...]

---

Outline:

(00:43) (1) Future Ability to Remember Things

(02:32) (2) Local Optima of Comfort

(04:13) (3) Interpersonal Conflict

(05:26) (4) Bonus: Recognizing Im in a Dream

(06:19) Takeaway

The original text contained 1 footnote which was 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: This is a quick analysis that might have major mistakes. I currently think there is something real and important here. I’m sharing to elicit feedback and update others insofar as an update is in order, and to learn that I am wrong insofar as that's the case.

Summary

The canonical paper about Algorithmic Progress is by Ho et al. (2024) who find that, historically, the pre-training compute used to reach a particular level of AI capabilities decreases by about 3× each year. Their data covers 2012-2023 and is focused on pre-training.

In this post I look at AI models from 2023-2025 and find that, based on what I think is the most intuitive analysis, catch-up algorithmic progress (including post-training) over this period is something like 16×–60× each year.

This intuitive analysis involves drawing the best-fit line through models that are on the frontier of training-compute efficiency over time, i.e., those that use the least training compute of any model yet to reach or exceed some capability level. I combine Epoch AI's estimates of training compute with model capability scores from Artificial Analysis's Intelligence Index. Each capability level thus yields a slope from its fit line, and these [...]

---

Outline:

(00:29) Summary

(02:37) What do I mean by 'algorithmic progress'?

(06:02) Methods and Results

(08:16) Sanity check: Qwen2.5-72B vs. Qwen3-30B-A3B

(10:09) Discussion

(10:12) How does this compare to the recent analysis in A Rosetta Stone for AI Benchmarks?

(14:47) How does this compare to other previous estimates of algorithmic progress

(17:44) How should we update on this analysis?

(20:13) Appendices

(20:17) Appendix: Filtering by different confidence levels of compute estimates

(20:24) All models

(20:45) Confident compute estimates

(21:07) Appendix: How fast is the cost of AI inference falling?

(23:56) Appendix: Histogram of 1 point buckets

(24:29) Appendix: Qwen2.5 and Qwen3 benchmark performance

(25:31) Appendix Leave-One-Out analysis

(27:08) Appendix: Limitations

(27:13) Outlier models

(29:41) Lack of early, weak models

(30:35) Post-training compute excluded

(31:17) Inference-time compute excluded

(32:16) Some AAII scores are estimates

(32:55) Comparing old and new models on the same benchmark

The original text contained 11 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.

Zvi said:

Unconditional Grants to Worthy Individuals Are Great

The process of applying for grants, raising money, and justifying your existence sucks.

A lot.

It especially sucks for many of the creatives and nerds that do a lot of the best work.

If you have to periodically go through this process, and are forced to continuously worry about making your work legible and how others will judge it, that will substantially hurt your true productivity. At best it is a constant distraction. By default, it is a severe warping effect. A version of this phenomenon is doing huge damage to academic science.

Compelled by this, I'm considering funding ~three people for two years each to work on whatever they see fit, much like the the Thiel Fellowship, but with an AI Alignment angle.

I want to find people who are excited to work on existential risk, but are currently spending much of their time working on something else due to financial reasons.

Instead of delegating the choice to some set of grant makers, I think that aggregating the opinion of the crowd could work better (at least as good at finding talent, but with less overall time spent)

[...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

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 entry for lsusr's write-like-lsusr competition.]

A new study out of the University of Michigan has confirmed what many people have long suspected: the benefits of meditation do not come from retraining your nervous system to be less reactive, but instead entirely derive from the act of telling people that you meditate.

The lead investigator of the study, Dr. Susan Connor, tested this by strapping brain wave helmets to participants while they meditated. She found no statistical significance between stress levels before and after meditating. What the participants didn’t know, however, was that there was a second part of the study and it began once they finished their meditation session: a “random” passerby (Dr. Connor's assistant) would enter the room and ask for directions to a nearby building on campus, and Dr. Connor would measure with a stopwatch how long it took for the participant to gently inform the passerby that they just meditated.

“We knew that it's virtually impossible for people to refrain from telling others that they meditate,” Dr. Connor admitted. “It's similar to how vegans and cyclists always find a way to mention their identity as early as possible in conversation. So we figured [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.