Dataflow matrix machines (by Anhinga anhinga)

Сборник разнообразных артистических манифестов (2001)

2026-02-20T05:08:16Z

Есть на сети стараниями бразильских профессоров: www.professores.uff.br/ricardobasbaum/wp-content/uploads/sites/164/2017/11/mary-ann-caws-manifesto-a-century-of-isms-1.pdf

Редактор до сих пор иногда пишет в своём блоге, несмотря на возраст: en.wikipedia.org/wiki/Mary_Ann_Caws

comments

some Oct-Nov posts I wrote

2025-12-18T04:53:56Z

LessWrong organized a blog writing program encouraging people to create 30 relatively long blog posts in Oct-Nov as an exercise (relatively long meant being 500+ words).

I have actually done that, here are the posts (mostly AI-related, although there were some diary-like posts too; 30 posts + 2 auxiliary ones): mishka-discord.dreamwidth.org/

(They are open for commenting, but I don't expect to continue posting new posts there.)

comments

Helion details

2025-09-02T05:41:07Z

Their CEO wrote a more detailed post a few weeks ago on how their fusion scheme is expected to work:

www.helionenergy.com/articles/how-to-make-fusion-electricity-without-ignition/

comments

Tao on coordinate vs coordinate-free math reasoning

2025-05-12T14:05:41Z

> On when to use coordinates and other concrete constructions in mathematics, and when to use coordinate-free formulations and abstractions:

> 1. If your priority is to perform computations in mathematics, use coordinates and concrete constructions.
> 2. If your priority is to generalize to as broad a range of use cases as possible, use coordinate-free formulations and abstractions.
> 3. If your priority is to actually understand what is going on behind the mathematical formalism, learn how the coordinate-based and coordinate-free approaches are equivalent.

mathstodon.xyz/@tao/114456756661540097

comments

"Aging as a loss of goal-directedness"

2025-05-07T06:15:47Z

osf.io/preprints/osf/m5bnx_v1 (Michael Levin's collaboration, Feb 2025)

Aging as a loss of goal-directedness: an evolutionary simulation and analysis unifying regeneration with anatomical rejuvenation

comments

"Narrow AGI" this year?

2025-01-07T02:38:29Z

"Narrow AGI" is mostly an AGI-level artificial software engineer, an AGI-level artificial mathematician, an AGI-level artificial AI researcher (and probably a single entity combining these three application areas, because a strong AI researcher has to be a decent software engineer and a decent mathematician).

It seems that at least OpenAI (and, perhaps, other entities) should have this by the middle of 2025, if not earlier, at least for their internal use (assuming no major disasters, that is, assuming that San Fransisco Bay Area is intact, and AI companies continue functioning normally).

What do we know about the technical aspects? We see o1 performance (and can experience it directly), we see the claimed (and partially confirmed) numbers for the demo versions of o3 and o3-mini, in math and in software engineering. We know that the jump from o1 to o3 took about 3 months. Two more jumps like that would probably be sufficient (and one can add "scaffolding" on top of that).

Another thing we know is that Sam Altman sounds much more confident recently. I've come to these conclusions a number of days ago, but now it turns out that Sam's mood has also shifted in a similar fashion. I'll put some links in the comments.

Jan 19 update: Sam Altman will allegedly do a closed-door government briefing on Jan 30 (that's apparently is not a very big secret and has been leaked; the main topic is presumably as follows: many people in the leading AI labs have approximately the same degree of techno-optimism as I have myself, and so their timelines are tentatively quite short). www.axios.com/2025/01/19/ai-superagent-openai-meta

Dec 31 update: no, we did not get "Narrow AGI" this year. But we had 2 revolutions, one centered in June plus/minus 2 months (o3---GPT5, "mature reasoning"), and one centered on the border of October/November plus/minus 6 weeks (GPT5-Codex---GPT5.2-Codex/Opus 4.5/Gemini 3, "competent programming software agents"). The distance between subsequent revolutions keeps shrinking: 7.5 years - 3 years - 18 months - about 9 months - about 4-5 months, mishka-discord.dreamwidth.org/4806.html (we don't know if this trend of distance between subsequent revolution shrinking continues, but we are watching what happens in January in this sense, in particular, whether we can progress to "trustworthy autonomy" in January; the current generation of agents is competent in software engineering, but there is no expectation of trust, so one has to work really hard to carefully organize security, quality control, and such, the whole lifecycle of software engineering can't be outsourced yet).

comments

Automating the Search for Artificial Life with Foundation Models

2024-12-24T07:19:58Z

pub.sakana.ai/asal/

comments

C to safe Rust automatic translation using LLMs and dynamic analysis

2024-12-19T05:29:59Z

This is a non-trivial angle: x.com/StringChaos/status/1869571815548596308 (UC Berkeley)

comments

"Anatomy of a Formal Proof"

2024-11-20T13:47:32Z

A really nice Lean tutorial: arxiv.org/abs/2411.11885

comments

GonzoML

2024-11-17T04:51:55Z

For some reason, I keep losing this remarkable blog by Grigory Sapunov and finding it again, instead of just reading it regularly:

gonzoml.substack.com/

comments

Transformers as a Computational Model (workshop)

2024-11-12T23:55:15Z

End of September 2024; video recordings are available:

https://simons.berkeley.edu/workshops/transformers-computational-model/schedule

comments

Geometric Langlands correspondence

2024-10-06T16:26:28Z

Говорят, что Денис Гайцгори и его коллеги доказали достаточно общий вариант геометрической гипотезы Ленглендса.

Это вполне эпохальное событие, и надо собрать вместе всякие линки, относящиеся к этому делу. Вместе с тем, это для меня слишком сложно (может быть, ИИ (современный или будущий) сможет мне, со временем объяснить детали всего этого так, чтобы у меня возникло уверенное понимание).

comments

Massachusetts - "Yes" on Question 4

2024-10-05T15:49:07Z

Voting by mail has started in Massachusetts.

I am urging "Yes" on Question 4 (legalization of benign psychedelics)

We passed Question 2 in 2008, Question 3 in 2012, and Question 4 in 2016 and our quality of life is better because we did that.

Let's do this again!

Links are in the comments

comments

A new book by Dima Kaledin

2024-09-27T17:17:16Z

"Enhancement for categories and homotopical algebra", arxiv.org/abs/2409.17489

600 pages

"We develop foundations for abstract homotopy theory based on Grothendieck's idea of a "derivator". The theory is model-independent, and does not depend on model categories, nor on simplicial sets. It is designed to accomodate all the usual potential applications, such as e.g. enhancements for derived categories of coherent sheaves, in a way that is as close as possible to usual category theory."

He also released references [K3] and [K4]:

arxiv.org/abs/2409.18380 and arxiv.org/abs/2409.18378

comments

On exploding pagers and walkie-talkies

2024-09-26T00:59:43Z

With all discussions on how this has been technically possible, I've seen only one person to offer a version which makes any sense.

"Причем, надо отдать должное устроителям, взорвались только те модели, которые были снабжены функцией самоуничтожения, при попадании девайса к врагам."

This is the only thing which makes sense. They themselves equipped their own devices with the ability to explode (and with the ability to trigger those explosions remotely).

After that, all it took was a bit of successful hacking by adversaries...

comments

The next AI revolution

2024-09-12T18:03:47Z

OpenAI is finally releasing their next set of models. Those models take time to ponder and reason internally before talking. This is what has been known as mysterious "Q*" and "Strawberry", but is now released as "o1 series of models".

They promise a preview version availability today for ChatGPT+ users.

Links in the comments.

comments

A possible anti-aging breakthrough

2024-07-08T13:00:26Z

"Gene Therapy-Mediated Partial Reprogramming Extends Lifespan and Reverses Age-Related Changes in Aged Mice"

A Feb 24 paper seems to claim an impressive result in "wild-type" mice via a partial cellular reprogramming protocol, which includes cyclic administration of doxycycline for temporal control of reprogramming, and which does not seem to be particularly cancer-inducing (that's usually a big problem with cellular reprogramming).. 124-week male mice, Doxycycline-treated controls average survival till 133 weeks, treatment group average survival till 142.5 weeks (for a human something like this would roughly double remaining life expectancy of a 75-year old male from about 10 more years to about 20 more years if equally efficient).

It would be nice if this turns out to be a real breakthrough (we have learned from experience to be very skeptical about claims in this field of study).

comments

NumPy 2.0

2024-06-19T14:14:10Z

There are breaking changing (this includes breaking binary compatibility):

blog.scientific-python.org/numpy/numpy2/

numpy.org/devdocs/numpy_2_0_migration_guide.html

comments

Golden Gate Claude

2024-05-24T14:29:55Z

I think humans tend to have these idée fixes:

www.anthropic.com/news/golden-gate-claude

A pretty interesting example of conversation by a very experienced interlocutor:

www.lesswrong.com/posts/cxuzALcmucCndYv4a/daniel-kokotajlo-s-shortform#hyXAiafwbwfEPiX5Q

comments

Two AI events today

2024-05-13T14:30:40Z

The author of Kolmogorov-Arnold Networks will present at 11am Eastern time.

OpenAI "Spring Updates" livestream at 1pm Eastern time.

Both events should be recorded.

comments

A study of untrained, random-weight networks

2024-05-01T14:30:00Z

"Neural Redshift: Random Networks are not Random Functions", arxiv.org/abs/2403.02241

Interesting geometry to explore.

comments

Grayscale

2024-04-28T01:59:05Z

Most of the devices (mobile phones, computers) can be switched to grayscale mode (and back).

This is usually under "Accessibility/Ease of Access" via turning "Color Filters" on (and off).

I like the effect :-)

comments

WorldCoder (Kevin Ellis group), understanding LLM refusals (Neel Nanda collaboration)

2024-04-27T13:51:06Z

"WorldCoder, a Model-Based LLM Agent: BuildingWorld Models by Writing Code and Interacting with the Environment", arxiv.org/abs/2402.12275

Not a widely known paper (the authors don't promote it), but pretty spectacular (a friend of mine said, "Is it AGI already?").

I think I mostly understand how this works and I made some notes yesterday.

A meta-note here: GPT-4-level models mostly understand what they are doing, but are unreliable; so the question is, can one organize a process which reliably produces needed results based on that. There are plenty of papers trying to push in this direction, but this one is very elegant, and the results are quite good.

******

www.lesswrong.com/posts/jGuXSZgv6qfdhMCuJ/refusal-in-llms-is-mediated-by-a-single-direction - very elegant and simple

******

May 9, 2024 update: Since this is access-list-only at the moment (although this post is likely to become public eventually), it's a good place for my notes on switching to Twitter "X Premium" experience (in comments).

May 13: let's move this post to being public.

comments

Llama 3 models are pretty spectacular

2024-04-19T14:51:46Z

Both 70B and 8B versions are very impressive on initial blind comparisons: chat.lmsys.org/?leaderboard

If Llama 3-70B-Instruct turns out to be indeed more or less equivalent to early GPT-4, this would have a lot of wide-ranging implications.

One can use Llama 3-70B-Instruct at www.meta.ai/ for free.

comments

Solar eclipse

2024-04-09T18:08:12Z

Впечатления от полного солнечного затмения.

Много разных впечатлений, но самое яркое вот какое.

Камеры делают фотографии, которые никуда не годятся; ну, или, по крайней мере, принято публиковать именно такие фотографии. На самом деле, новолунная луна, полностью закрывающая солнце, конечно, не чёрная, как на всех фотографиях, а такого цвета, как закатное небо вокруг, - очень красивого умеренно-тёмно-синего, и даже не особо тёмного, цвета, вполне волшебного. Почему-то, камеры это не способны взять (из-за контраста с короной, что ли), но хоть бы уж фотошопом обрабатывали бы, чем публиковать настолько не похожее на то, что на самом деле...

comments