iiinova — where your digital self keeps evolving

Axe Ghost. On Steam! @axeghostgame.bsky.social · 4h

straight-facedly pronouncing Influencer as Influenza

Reposted by Axe Ghost. On Steam!

So the Dutch government is now developing its own Linux Distro based on NixOS but hosts it on Codeberg? 1st: nice to see Linux getting more traction 2nd: considering Codeberg's vehement anti-AI policy I think the Dutch government will eventually regret hosting it there

2141

Reposted by Axe Ghost. On Steam!

Matt Kane @mk.gg · 20h

Hello everyone please don't use this model that's almost as good as ours at 1/10th the cost and without all the guardrails that won't even let you fix an auth bug

319431

Axe Ghost. On Steam! @axeghostgame.bsky.social · 11h

GLM-5.3 is very good, and relatively cheap

020

Reposted by Axe Ghost. On Steam!

Andy Masley @andymasley.bsky.social · 12h

I really do not like the recent push against anthropomorphic language

813819

Axe Ghost. On Steam! @axeghostgame.bsky.social · 29/09/2026

sidequest. have my agent deliver a different-voice-per-character/persona audiobook of blindsight, using elevenlabs

230

Reposted by Axe Ghost. On Steam!

Connor Lynch @connorlynch.bsky.social · 29/09/2026

This just strikes me as basic insincerity. If you took 10 people who casually accepted the “stochastic parrot” framing and 10 people who accepted the “reasoning” framing and asked them to predict what frontier models can do, virtual guaranty the latter group would outperform the former group.

3293

Reposted by Axe Ghost. On Steam!

Connor Lynch @connorlynch.bsky.social · 28/09/2026

her work has a problem: it either produces no testable hypothesis and is consistent with all observations (in which case, yes, these things are reasoning and that’s fine) or it has produced testable hypotheses that are wrong (because it asserts, incorrectly, they cannot reason)

1293

Axe Ghost. On Steam! @axeghostgame.bsky.social · 29/09/2026

'sent' is imo ambiguous. 'to annoy' is clearly false. the post solicited counter-arguments.

0120

Reposted by Axe Ghost. On Steam!

SpaceTessContinuum @eigenvalue.bsky.social · 29/09/2026

We are rapidly entering into an era where the old joke "Everything is open source if you can read assembly" is not a joke.

721321

Reposted by Axe Ghost. On Steam!

Kate Sills @katelynsills.com · 28/09/2026

We don't even need to get into theory. There are direct and easy to understand counterexamples that disprove what you say here. For example, Levant Alpöge used Claude Fable (an LLM) to show the Jacobian Conjecture was false, and this was neither brute force nor a targeted machine learning model.

I'm baffled by these PR stunts. Targeted machine learning models and brute force algorithms can sometimes solve difficult problems that have eluded human beings. But using an LLM doesn't add anything (and makes success less likely). They are not "AI." They are stochastic parrots.
1/3

2362

Axe Ghost. On Steam! @axeghostgame.bsky.social · 28/09/2026

dreamed there was a Sesame Street reboot but instead of Big Bird there was Big Gun, which was also a person in a costume and leggings.

010

Reposted by Axe Ghost. On Steam!

ponder @ponder.ooo · 28/09/2026

sorry it's just. i can't deal with woodchips rn. no you're good your fine it's just, i have a lot going on and i just can't handle woodchips, on top of all that? like. just not a day for woodchips

0252

Reposted by Axe Ghost. On Steam!

Jaz @jaz.sh · 27/09/2026

Okay so I've built a new Atlas, this time it lets you explore the conversations on Bluesky over the past 7 days graphically. It should update every ~6 hours and help you navigate the information environment that exists on here. I've learned a LOT about what people do here today... atlas.jazco.dev

atlas.jazco.dev

Atlas

A living map of what Bluesky is talking about: the past week's conversations, grouped into topics and regions and rebuilt every six hours.

1161388406

Axe Ghost. On Steam! @axeghostgame.bsky.social · 27/09/2026

The grass will get longer, but there's no point at which it's meaningful to say, 'Yesterday we didn't have long grass. Today we do'. That's why it's meaningless to say 'What will happen one day if we have long grass?' I don't even understand the question. 🧠

161

Reposted by Axe Ghost. On Steam!

post malone ergo propter malone @proptermalone.bsky.social · 27/09/2026

putting together a kind of quickstart kit for ADHD users and LLMs, i'm excited about this, should have it up in the next couple of days

273848

Axe Ghost. On Steam! @axeghostgame.bsky.social · 27/09/2026

Sleepytime decaf implies the existen...

110

Axe Ghost. On Steam! @axeghostgame.bsky.social · 27/09/2026

per the godot community rules, AI authored PRs are closed within hours - and subsequently ignored. *even when they relate to memory safety bugs in exported games* ie. live attack primitives. but at least they don't have to deal with the icky machine output!

4430

Reposted by Axe Ghost. On Steam!

sōm ⁂ @hyl.st · 26/09/2026

God I swear to god I feel like a crazy person talking to "non deterministic means unbounded outcomes" types, telling them we have ways to make robustness out of non determinism that's literally what engineering is

093

Axe Ghost. On Steam! @axeghostgame.bsky.social · 27/09/2026

for video game runtimes I'm much more excited about the potential of local jevlikes than talky LLMs

2270

Axe Ghost. On Steam! @axeghostgame.bsky.social · 27/09/2026

I was directing a security audit of a popular open source project 🥲

010

Reposted by Axe Ghost. On Steam!

ponder @ponder.ooo · 27/09/2026

(forgetting what tongues are called) the uh. uhm.. uhh. y'know, the speech tentacle

4323

Reposted by Axe Ghost. On Steam!

p(Dulany) @dulanyw.bsky.social · 26/09/2026

This is 100% correct and more people should be saying it The future of math is developing understanding from AI traces There is every likelihood we will need *more* mathematicians

0193

Axe Ghost. On Steam! @axeghostgame.bsky.social · 26/09/2026

holy shit this announcement reached my feed at an uncannily fortuitous moment

130

Reposted by Axe Ghost. On Steam!

Sung Kim @sungkim.bsky.social · 26/09/2026

Yet Another Jev Variant. I should create an acronym - YAJV! Julia 1: A decision model that runs on almost anything Blog: supersoniclabs.ia.br/julia-1/ Model: huggingface.co/SupersonicLa...

supersoniclabs.ia.br

Introducing Julia 1 | Supersonic Labs

Julia 1 is our compact decision model. Explore the research, evaluation results, limitations, and model repository.

4937

Reposted by Axe Ghost. On Steam!

conputer dipshit @davidcrespo.bsky.social · 26/09/2026

I’m sorry, but if you bring gypsum - or any damn mineral below 3 on the Mohs scale at all - to a job interview, you’re disqualified. It’s not hard, people.

562067281

Reposted by Axe Ghost. On Steam!

Ansambel @ansambel.bsky.social · 26/09/2026

funny thing that happens with LLMs is that you can ask it to do things that humans are generally too lazy to do. Like writing a post-session summary and using it to self-iterate your setup and generally use the beurocracy and processes, in a way that could never work for humans.

041

Axe Ghost. On Steam! @axeghostgame.bsky.social · 26/09/2026

how i feel explaining llm -> llm adversarial review to non ai types

7845

Axe Ghost. On Steam! @axeghostgame.bsky.social · 26/09/2026

today i heard the claim 'Having a second AI take a look [at the work of the first AI] doesn't help', and the utterer is not ready to hear it but: in fact it helps a lot!

4723

Axe Ghost. On Steam! @axeghostgame.bsky.social · 26/09/2026

soon: Alberta's self-image as world's only second-rat-free region shattered by discovery of second rat.

000

Reposted by Axe Ghost. On Steam!

norvid_studies @norvid-studies.bsky.social · 29/01/2026

how are people this 'velocity blind', over and over? I genuinely don't understand it. like I can't warp my grooves of understanding into a shape where this wouldn't immediately slide off. this is like piaget 2 stuff.

161097

Axe Ghost. On Steam! @axeghostgame.bsky.social · 26/09/2026

this dynamic is interesting isn't it ai allows new attacks against OSS projects to be discovered and weaponised easily. defensive ai use could prevent those attacks but advocacy for that remedy is blocked by anti-ai policy/sentiment on exactly the fora that'd need to be persuaded 🫠

3545

Axe Ghost. On Steam! @axeghostgame.bsky.social · 26/09/2026

e10908efd281d7f3c08c868b69ef37c281013548e3e7670d578fdde3ec88d1c7

000

Reposted by Axe Ghost. On Steam!

tbabb @tbabb.bsky.social · 26/09/2026

frome celestepoasts on twitter asking "land or water" for 2 degree lat/long increments over the entire earth

59010

Reposted by Axe Ghost. On Steam!

apenwarr @apenwarr.ca · 26/09/2026

Every sign tells a story

3978

Axe Ghost. On Steam! @axeghostgame.bsky.social · 24/09/2026

time to have the AI conversation w my accountant who commented on how good my written dutch has gotten, and praised me for having looked so critically at the draft income tax filing 🧠✨

191

Axe Ghost. On Steam! @axeghostgame.bsky.social · 24/09/2026

security work: deliberately finding out extremely interesting things and then not immediately telling everyone

010

Axe Ghost. On Steam! @axeghostgame.bsky.social · 24/09/2026

feels like I've been droning this daily lately: memory-safety bugs triggered when exported games load community-sourced files are definitionally security issues; they allow untrusted input to corrupt the game’s process memory youdve though this was unnecessary to point out, and yet!

010

Axe Ghost. On Steam! @axeghostgame.bsky.social · 24/09/2026

again though: artificial [something] doesn't necessarily connote that the artificial thing is not an example of [something]. it can connote that it's a man-made example of [something]! e.g. artificial sweeteners

080

Reposted by Axe Ghost. On Steam!

Tim Price @timprice.bsky.social · 23/09/2026

You joined "grand vizier chat (secret)" You: dark rumors from aboard You: abroad* fuck Sssnakecharmer: mf said aboard WhispersintheChapel: ahoy matey lol PonderMyOrb changed the group name to "rumors from aboard"

2893712650

Axe Ghost. On Steam! @axeghostgame.bsky.social · 24/09/2026

fwiw i recently pulled an old game though exactly 10 years of unity updates using opus 5. it took less than a day.

050

Axe Ghost. On Steam! @axeghostgame.bsky.social · 24/09/2026

the mistake is assuming these characteristics can belong to only humans. ai systems certainly think and understand, at least.

030