iiinova — where your digital self keeps evolving

Check out this range of halloween-ass plant food

Various gory-sounding plant supplements including “dried blood” “bonemeal” and “blood, fish & bone”

000

Went to a garden centre today. As far as I can tell they have some kind of captive god called “John Innes” and harvest his blood and bones to put on flowerbeds

100

wariotifo @wariotifo.bsky.social · 11/10/2026

There are lots more boob jokes in the disney animated robin hood than I remembered

000

Reposted by 𓅭𓅮𓅭𓅮𓅭𓅮

really odd. this Spectator article from just 2 DAYS AGO seems to have been deleted from the live site along with her author page. hope the Spectator's web team are on top of this and will fix because it could look a bit, from certain angles, like they've caved to the woke mob archive.ph/yE35A

archive.ph

14621

❗️ It looks like Terminus Est from the books ‼️ It has a latin inscription on the blade that reads “The word become flesh guides and directs” ‼️‼️ It was a gift from Mr and Mrs John L *SEVERANCE*

010

Atrios @eschatonblog.com · 11/10/2026

Right, but I don’t think it’s a helpful starting point for people to relitigate “complex systems can / cannot demonstrate behaviours that aren’t possible for their constituent parts”

000

Reposted by 𓅭𓅮𓅭𓅮𓅭𓅮

are we going to talk about how the mail has pulled this article (2nd) or just move past it as everybody tends to do

112645

Reposted by 𓅭𓅮𓅭𓅮𓅭𓅮

Ray Newman @raynewman.bsky.social · 10/10/2026

That time someone dumped a toy house on the cemetery wall and, when I crouched a bit, some ✨ Hollywood Magic ✨ occurred.

A plastic dolls house sitting on a wall, taken from an angle which makes it look as if it might be one of the houses on the street.

6698571831

Reposted by 𓅭𓅮𓅭𓅮𓅭𓅮

Abel @abeliansoup.bsky.social · 11/10/2026

thanks to modern innovations in shape packing, proud to report I've found the optimal dishwasher item configuration

919215

Nathan Douglas @tenesm.us · 10/10/2026

Literally had this argument, drunk, with a relative who has a philosphy degree, but thought it was just the whisky that was making him confused about why atoms aren’t a little bit conscious

110

Reposted by 𓅭𓅮𓅭𓅮𓅭𓅮

systems thinking is incredibly underutilized and underappreciated

151

Ed Granter @edgranter.bsky.social · 07/08/2026

Round 2 tonight. I woke up early after ~3hr of post-party dozing and spent the day traipsing around chilly countryside. So have decided to let my dad buy me a takeaway while I steal his hot water for a bath. Loosen the old calves before doing 6+ hr of rhythmic heel lifts in an old industrial unit

000

Reposted by 𓅭𓅮𓅭𓅮𓅭𓅮

UK industrial strategy: 1/1980s: Destroy your manufacturing base and transition to a service/knowledge economy at great human cost. 2/The 2020s: Destroy the knowledge economy at great human cost. 2.2/Announce reindustrialisation. You have, effectively, no manufacturing base. Absolutely masterful.

1114

I guess sign me up with the materialists lol

170

Isn’t this just a catastrophic failure to understand emergent properties? You may as well spend your time wondering why electrons can’t tell jokes and describe it as the Hard Problem Of Social Media

170

Also going to a big warehouse full of sweaty people is going to be a lot warmer than my house as it’s freezing and my boiler is fucked

100

Going to see an electronic music man I like playing some electronic music later on tonight, excited

100

Yes that one feels totally disconnected from reality. AI companies may be pinching people’s works-in-progress or other dubious practices, but (like the water thing) it’s better to oppose them on the grounds of stuff that’s real, not made up

010

Jan @janhopis.bsky.social · 09/10/2026

Wish I had a more rigorous understanding of entropy (in the information theory, non-molecular sense) because this feels to me more like a potential failure state for LLMs - a kind of information heat death

110

Reposted by 𓅭𓅮𓅭𓅮𓅭𓅮

arxiv.org/pdf/2604.02279 dug this out on the topic, seems like the nature of the training data is such that it's impossible for them not to start leaking out-of-scope information. Maths problems are not time-barred!

111

Oooh!!! Now this does make intuitive sense. I bet this has been tried a few times to everyone’s deep frustration ahaha

110

Barging into the thread to announce that the fuet isn’t great but it’s more than adequate and scarily possible to absentmindedly eat a whole one in tiny slices in an hour or so

120

…feels like something someone would have tried after hours at some point because why not. So you’re probably right, but that equally feels like something frontier LLMs “should” be good at - brute forcing non-intuitive strategies based on massive datasets

000

“Wrong flavour of maths” may well be it, although I think that funds adopt a variety of different strategies. “Here’s a Bloomberg terminal, hypothesise some novel ways you could have made money over the last 5yr of data, test them on the last 15yr, adopt the best strategies using some of our cash”

200

It’s also possible that hedge funds are a lot more relational / intangible than my mental model of them. Maybe I ought to go and re-read that old N+1 article, or some archived copies of Matt Levine’s newsletter

000

“That would be mildly unethical” - look at everything OpenAI does all the time “They’re stochastic parrots and can’t do maths” - this looks increasingly unlikely “They haven’t thought of this” - yes they have “They’re doing it already” - possible, but I feel like they’d be crowing about it

200

9 min to go, not pulling a late one today

000