Richard_Kennaway

Karma: 8,616

Computer scientist, applied mathematician. Based in the eastern part of England.

Fan of control theory in general and Perceptual Control Theory in particular. Everyone should know about these, whatever subsequent attitude to them they might reach. These, plus consciousness of abstraction dissolve a great many confusions.

I wrote the Insanity Wolf Sanity Test. There it is, work out for yourself what it means.

Change ringer since 2022. It teaches learning and grasping abstract patterns, memory, thinking with your body, thinking on your feet, fixing problems and moving on, always looking to the future and letting both the errors and successes of the past go.

I first found an LLM useful (other than for answering the question “let’s see how well the dog can walk on its hind legs”) in September 2025. As yet they do not form a regular part of anything I do.

Richard_Kennaway 19 Jun 2026 21:45 UTC
6 points
4
on: Cosmological Odyssey
Did not follow links.

It’s like when I receive an email telling me that Evri has received my parcel from Aliexpress and click here for tracking info, as happened this morning. I deleted it un-clicked-on, because I know that I am not expecting a parcel from Aliexpress for delivery by Evri.

Or when I see a video thumbnail with a clickbait title, I don’t click, because I already know the video behind it will be fake AI stuff and every ad on the page will be a scam designed for people stupid enough to click the thumbnail.

Or when I see a link to yet another video in which someone forges a knife from paperclips, or shows how ball bearings are made, or seven things you never knew about whatever, I mean, each of these things can be genuinely interesting the first time but no more than that.

Even so, an account of a stupendous AI-assisted discovery that unlocks “insight after insight” does not reach the bar. As you say, “There’s no way in hell this is how it will be discovered.” Truer than you think, there really isn’t.

If you can’t judge it yourself, unaided, don’t expect anyone else to.

Richard_Kennaway 15 Jun 2026 17:28 UTC
13 points
0
on: A frontier AI company should shut down
I have no business experience, but I am wondering about the practicalities of shutting down such a company. If it has already gone public, its directors may not even own it. The shareholders would have to agree, which they’ll hardly be willing to do or they wouldn’t have invested. Even if the directors technically own it, what stops the employees banding together to take it over? Or other companies mounting a takeover, or poaching the employees and IP? Or the current regime or the next nationalising it, which there has been mention of from both political sides?

Are these companies already too big to shut down?

Richard_Kennaway 14 Jun 2026 21:35 UTC
2 points
−2
on: A simple argument for trying less hard

It seems deeply hard to distinguish between what we want and what we believe.

Er … speak for yourself?

My beliefs concern what is already true. My wants concern what is not already true. The difference could hardly be more stark.

The things that I want but are not, it is up to me to bring about. I cannot merely open my mouth and believe a roast duck will fly in.

Richard_Kennaway 14 Jun 2026 21:02 UTC
2 points
0
on: Tequila Sunset at the Hog’s Head (A Scene)
More!

ETA: Unless it was AI generated, in which case: Less!

Richard_Kennaway 12 Jun 2026 16:11 UTC
4 points
1
on: If you, a human, can imagine red and green being swapped, you are probably conscious
I recall an anecdote from two or three centuries ago ^[1] , in which someone told of a man blind from birth, who opined that although he had never seen red, he imagined it must be something like the sound of a trumpet.

The teller of this tale intended it as illustrating the foolishness of those who speak of what they do not know, but I think that the sound of a trumpet is an excellent projection of the visual experience into the audio gamut. A bright red, the colour used on flags, and pillar boxes, and fire engines. It is just like the sound of a trumpet.
1. ↩︎
  But I flatter myself that I don’t look a day over 120.

Richard_Kennaway 8 Jun 2026 13:48 UTC
3 points
0
in reply to: lilkim2025’s comment on: Three Stories about not Thinking too Hard

If this is possible, then how did the splitting method that guarantees that the oldest gets X - (N-1)//2 coins every time become a tradition? Seems like the first guy to propose it would’ve met the same fate as Larry.

The puzzle is a traditional way to weed out the Larrys.

Richard_Kennaway 7 Jun 2026 6:05 UTC
2 points
0
in reply to: RedMan’s comment on: leogao’s Shortform
I’ve only seen “doomer” in the context of x-risk from AI. If you don’t thnk it’s an x-risk, what is the doom you’re a doomer about?

Richard_Kennaway 4 Jun 2026 6:58 UTC
4 points
2
in reply to: G Wood’s comment on: G Wood’s Shortform

You still own the words you write?

Writing is an exercise in developing and crafting your thoughts. Outsourcing the job is missing the point.

You don’t own the LLM’s words. You can put your name to them, but if you didn’t write them yourself, it’s all too easy to just nod along and think, yes, that’s exactly what I meant!, even if they make no sense. It’s like reading the answers to the questions in a textbook and thinking, yes, of course that’s the answer, except that LLMs don’t give any assurance that what they wrote means anything.

“In glorious future, LLM words own you.”

Richard_Kennaway 3 Jun 2026 8:52 UTC
−5 points
0
on: China won’t win the AI race but would it be much worse if it did?

I do wish people would simply abandon using the race to justify anything

I think you just did.

Tl;dr: To stop rogue AI we must be willing to destroy everything else.

Richard_Kennaway 2 Jun 2026 7:40 UTC
4 points
2
in reply to: Julian Bradshaw’s comment on: NYT: Senator Sanders Proposes Gov’t Take 50% Ownership of AI Labs

the argument is that the US federal government is controlled by the US general public (“us”)

I find this breathtakingly naive, especially given the regime most recently voted in by the US general public. I certainly wouldn’t expect better of any government than the current leaders of the major AI companies, and in fact, considerably worse.

Politics warning:

The Trumpists will bleed the goose to line their own pockets, while the Dems (if they ever get back in) will wring its neck for having the effrontery to be laying golden eggs without their permission.

Richard_Kennaway 1 Jun 2026 16:52 UTC
12 points
1
on: NYT: Senator Sanders Proposes Gov’t Take 50% Ownership of AI Labs
What is the mechanism whereby this nationalisation of the AI industry will “guarantee that the trillions of dollars potentially generated by A.I. are used to improve the lives of all of us”?

Richard_Kennaway 1 Jun 2026 15:27 UTC
11 points
10
in reply to: Oliver Sourbut’s comment on: leogao’s Shortform
It seems to me that you are greatly broadening (that is, redefining) the word “analogy” to mean any sort of approximation. For example, survey results are not an analogy to the ground truth but an approximation to it.

Richard_Kennaway 30 May 2026 16:12 UTC
2 points
0
on: A Formula for Fun

count every joy, discount repeats

I saw this line and immediately heard it to the song “Climb every mountain”. Fitting the rest in is trickier, but “keep the glass expanding, let no life decline” is perfect. A project for the rationalist songbook?

ETA:

Count every rapture, keep them unique
Upwards, ever upwards, both the strong and weak
Keep the glass expanding, let no life decline
Use well all negentropy, while the stars still shine.

Richard_Kennaway 29 May 2026 19:34 UTC
4 points
0
on: Infinite ethics and UDASSA

(I sometimes phrase this as the Reverse Pascal’s Wager: the gods almost certainly exist, but if they don’t, the stakes are much higher, so you should act as if they didn’t exist.)

There’s a Jewish story related to that, variously attributed in various versions. As told by Martin Buber of an unnamed rabbi, who said:

“When an atheist performs an act of charity, visits someone who is sick, helps someone in need, and cares for the world, he is not doing so because of some religious teaching. He does not believe that God commanded him to perform this act. In fact, he does not believe in God at all, so his actions are based on his sense of morality. Look at the kindness he bestows on others simply because he feels it to be right. When someone reaches out to you for help, you should never say ‘I’ll pray that God will help you.’ Instead, for that moment, you should become an atheist – imagine there is no God who could help, and say ‘I will help you’.”

Richard_Kennaway 24 May 2026 4:47 UTC
13 points
0
in reply to: Seth Herd’s comment on: Basic principles for dressing better.
There’s plenty of general principles there, but it all seems very culture-specific, the culture being something I can’t identify but fairly narrow.

Richard_Kennaway 23 May 2026 21:46 UTC
49 points
37
on: Basic principles for dressing better.
It is fascinating to receive such communiqués of strange news from another star. I would not know, without your saying so, that the set of 10 pictures are intended to be good examples and not bad examples. I am not even sure that these people exist. They look like the AI-generated product images I see on Temu.

Generic tank tops and t-shirts aside, there is not a single item of clothing there that I have, or would buy, or would wear. The same applies to what I saw on the web sites of a sample of the menswear suppliers you listed.

I don’t mean this as a criticism. It’s just a completely different world from any I have inhabited.

Richard_Kennaway 23 May 2026 21:11 UTC
2 points
0
in reply to: Karl Krueger’s comment on: Will we really put data centers in space?

nobody can stop you without a military strike (and the risk of Kessler syndrome).

Would it work to send a robot to rendezvous with it and turn it so the solar cells aren’t facing the sun? Or cut the cables collecting the electricity?

One can think up countermeasures. And countermeasures to those countermeasures.

Richard_Kennaway 20 May 2026 23:14 UTC
2 points
0
in reply to: Alex_Altair’s comment on: Work with me on agent foundations: independent fellowship
I thought Google Translate must be an LLM these days, but of course there’s more than one. Claude did what looks at first glance to be a good job, but I’ve not given it more than a casual glance. I’ve put online the Russian original, the Latex and PDF files made by Claude, and a screenshot of Claude’s comments on the translation here. Hope this is useful.

Richard_Kennaway 19 May 2026 14:58 UTC
2 points
0
on: Work with me on agent foundations: independent fellowship

A. A. Brudno, Entropy and the complexity of the trajectories of a dynamical system (1983)

I’m not applying (OP was 18 months ago anyway) but where can I find this paper in English? I have found a 1982 paper by Brudno in Russian whose title translates to that, and a 1978 paper in English, “The complexity of the trajectories of a dynamical system”, which is much shorter, less than two pages, but I can’t find online a translation of the 1982 paper, although I have found references to one.

I tried getting Google Translate to translate the Russian PDF but it did not do a useful job.

Richard_Kennaway 19 May 2026 10:26 UTC
2 points
0
on: The anthropic argument against the existence of God.
Here is an argument for God creating all possible universes that are each better to exist than not. One of them is the maximally stupendously good universe. Ours just isn’t that one.

Not that I’m arguing which of these or other ideas is right, I just thought you might like to see that one.