Currently reading the most mediocre fantasy novel I've ever found

Reading some Roland barthes this afternoon. Not a big fan tbh. Feels disorganized and shallow.

It's A Good Thing The Dark Lord Is A Shut-In andataexpress.neocities.org/shut-in-ch Failure 12 - season finale HOLOHAUS12 - this weekend #fiction #webserial #booksky #WritingCommunity #indieauthor

* IRAN'S NAVY SAYS IT FIRED WARNING SHOTS AT U.S. DESTROYERS IN GULF OF OMAN -- STATE MEDIA @reuters.com

Surely it'll be possible for AI agents to train new AIs, indeed in the near future. I have great doubts that these will be significantly improved, much less improved at an accelerating rate.

what I really need for my next project is to collaborate with someone who really understands structural linguistics and/or computer science so they can beta read/give feedback on it as I go. This would be a book on semiotics and the philosophy of mind.

DNA synthesis security An open letter to congress requests that shops that do DNA synthesis keep records to support investigations and proactively identify dangerous sequences Very common-sense ideas that should be non-controversial and high impact screendna.org

Oil price dot com was on the ball wrt Chinese imports and inventories holding down prices earlier than most other publications, now talking about how this might change soon oilprice.com/Energy/Crude...

Btw, the universal approximation theorem also applies to language's ability to express things, as there is an equivalence between neural nets and collections of linguistic signs.

Another private equity fund paused redemptions, this time in a fund for wealthy investors

EA ppl writing debunking articles are just another flavor of this guy

Just got caught up on Hell's Paradise Season 2, pretty good.

My counter to Zizek's claim that AI is not a subject. open.substack.com/pub/slavoj/p...

Jameson's glazing of Derrida is really annoying. That he doesn't realize just how facile and petulant Derrida's maneuver is really just feels like a reflection on the fact that he prefers literary over scientific theory.

writers online are way too fucking self-deprecating to the point that the only reason anyone acts like anyone would engage with anyone else's work is out of some sort of quasi-ethical solidarity, which is a bad reason to engage in the amoral process of art

Found another reference to Althusser in Eco's untranslated work thanks to a footnote from Jameson. Interesting that Eco did in fact read "Reading Capital" but disappointing he takes Heidegger's side in this contra Althusser and Marxist materialism.

Just so everyone knows, the hard problem of consciousness is clearly a bourgeois ideology.

Frog outside tonight

Currently reading Jameson's The Prisonhouse of Language and he really nails this distinction between the novel and the short story, and it's exactly why an LLM can write, on a technical level, a short story, but cannot write a novel to any degree of technical competence

goblins are often depicted as green, but actually they only occasionally turn green, to photosynthesize in times of food scarcity

Was just talking about this the other day

We have the empirical evidence of what liberal societies produce and its social atomization, the absence of all those things. Well, except trade ofc since we must work for a wage to live. The idea you can solve that with a personal ideological commitment is idealist pablum.

Despite some striking statements, I don't think pope leo actually forecloses the possibility of needing to extend human dignity to AI. In fact, he lays out the requirements for that to happen.

omenmaxxing

audio reactivity tests with ASCII fluid sim. going to take a lot of work to tune emitters and parameters so it actually looks good.

He does not call them stochastic parrots, everything he says there is true of current AI tho it may not be true in the near future.

❌ I used a random number generator ✅ I wrote a synthetic heuristic script based on my known priors

Reading the pope's encyclical on AI today that just dropped. reply / indicate if interested in a discussion group. it will likely become a fairly influential historical document. this is just for fun. this is not a test. www.vatican.va/content/leo-...

As more people come to recognize the tells of AI, which mostly happens as you start to work with AI a lot, the scales are going to fall from their eyes and they are going to realize what some of us already see: how much of this site (and blog posts, articles, scientific papers) are AI now.

I know, I know, I have some kind of severe ascii fluid sim addiction but I'm trying to build the once-and-for-all version and god damn this shit lookin good.

Nine more Erdos problems resolved with LLMs + Lean: arxiv.org/pdf/2605.227... Worth noting they got 9/353 after trying a giant set of open problems by Erdos, and 44/492 OEIS open conjectures. Likely would've gotten 0 a year ago.

If ancient rome was around today people would be eating pavonem sandwiches. #2amposting

Still true

Glasswing update: they used mythos to find 10k+ high or critical severity vulnerabilities, cloudflare has found the false positive rate to be better than human testers, and they've used mythos to scan 1k+ OSS projects, finding 6k+ bugs with high or crit severity, and 23k total). Crazy stuff in this

A follow up on my research on making a general measure of complexity, with the help of Claude I make a video analyzer and simulation analyzer based on my metric, and figure out the normalization method to apply it more broadly. nicolasdvillarreal.substack.com/p/a-measure-...

May have just vibe coded a general measure of complexity

Dreamt I was in a first person shooter last night, but I kept missing when trying to throw a grenade out the window so I had to grab a bunch of other players and onlookers by the hand and pull them away from where it landed.

They should sell canned potatoes in sauce #2amposting

Realized I should make a thing that actually operationalizes my ideas about how to measure complexity

This is the best use of claude, which is to find the right paper authors to pester.

Another day another night spent emailing scientists about my ideas.

Trying out Claude. Definitely does seem to get to the point better than chatgpt or gemini which have been my goto.

Lmao hell yeah

Explain to me why airplane wifi is slower despite being closer to the cloud

As the winners become apparent, so will the losers.

@littmath.bsky.social's commentary, noting that the low hanging fruit here may be because humans were just stuck on the wrong approach, trying to prove it true instead of false. In that sense I wonder how much the model benefits from all of those failed attempts not being part of the training corpus

The only way to do literature in the 21st century is to make the text the new subtext.

Wrote a letter response to a recent cosmonaut article, arguing against the author about why there will in fact be a green energy transition cosmonautmag.com/2026/05/lett...