iiinova — where your digital self keeps evolving

I feel like it's 2021 again

Palpatine meme but with "Unlimited Tokens!" (in Comic Sans)

010

BECAUSE PEOPLE BELIEVE THESE THINGS ARE SENTIENT WHEN THEY ARE NOT I will explain this over and over until the end of time. You're welcome to unfollow me.

000

If English is better than Ruby… what do you even need a computer and an operating system for anyway?

050

The randomness is part of the input, as in any software algorithm that uses random numbers. That’s just how computers work. ¯\_(ツ)_/¯

000

Debates over whether LLMs can think are about as fruitful as discussing if hammers have ethics

1152

It needs constant repeating because people are falling for it constantly. “AI” does not exist. People are—directly—responsible and culpable if something bad happens as a result of LLM output.

110

Salmiak is the strongest proof so far that we have been visited by aliens.

000

"But Thomas, it's not outputting the same thing for the same questions!" That's because pseudorandom tokens are used in the the input state of the algorithm. There may also be implementation details adding randomness such as race conditions. Doesn't matter—that's all input state.

3131

There's no threat there because LLMs don't even know what threats are. The don't know what anything is. They don't have permanence or understanding. Yes, a human can use them to offload decision-making; but that's a human being the threat then.

1180

Timnit Gebru Believes There Is No ‘Existential Threat’ From AI

Correct www.wired.com/story/the-bi... You don't need to "believe" this by the way; LLM don't have thoughts or reasoning. The output state is entirely deterministically derived as a straightforward strictly defined mathematical operation from the input state.

wired.com

One of AI’s fiercest critics believes the doom talk is about founders making money, not saving humanity.

2171

Amy Hoy @amyhoy.bsky.social · 4h

what's wrong, honey? don't you like gagh & prune pizza?

140

Reposted by Thomas Fuchs 🫯

what is with all the chatbot replies going on rn it’s offensive, and what do they think it gets them??

0121

Thomas Fuchs 🫯 @thomasfuchs.at · 4h

why do you hate waffles

010

Reposted by Thomas Fuchs 🫯

justin from the internet @threlk.net · 6h

the path to broad adoption is, somewhat definitionally, mundane use cases like this

184

Reposted by Thomas Fuchs 🫯

Bruno Dias @brunodias.dev · 5h

oh my god he's handed the block button over to claude

58710

Thomas Fuchs 🫯 @thomasfuchs.at · 5h

Also how amazing is this thing, it's a replica of Leibniz's stepped reckoner. Why can't modern computers look like this.

120

Thomas Fuchs 🫯 @thomasfuchs.at · 5h

✨ reminder that AI doesn’t exist ✨

180

Thomas Fuchs 🫯 @thomasfuchs.at · 5h

Since 5 decades ago now! And mechanical desktop calculators since the 1600s.

The Leibniz's stepped reckoner, something that looks like a steampunk fever dream, made out of brass and with many levers, dials and cranks

120

Reposted by Thomas Fuchs 🫯

Lee Benfield @benf.org · 7h

Anyone ACTUALLY restored a sticking/squealing tape? My original CPC Dizzy has lost it's mojo. I see reccomendations about using silicon lube, but also warnings not to! :( @breakintoprogram.co.uk ?

452

Thomas Fuchs 🫯 @thomasfuchs.at · 22h

Have you thought about only making good changes

4362

Reposted by Thomas Fuchs 🫯

Hypervisible @hypervisible.blacksky.app · 29/09/2026

“What if we released malware represented in the form of a cute avatar” is an actual business plan now.

engadget.com

Dots are OpenAI's new personal agents and soon you'll be able to control several of them - Engadget

Sure, seems safe.

631171378

Reposted by Thomas Fuchs 🫯

Amy Hoy @amyhoy.bsky.social · 29/09/2026

great thread (as always) about how "tech people" pretend they live in a world without actual users

1319

Reposted by Thomas Fuchs 🫯

Corey @kh0rish.bsky.social · 29/09/2026

I’m curious why so many people still think this

There appears to be nothing in the composition of the universe, or our minds, that would prevent us from making a machine as smart as us, and probably (but not as surely) smarter than us.

2101

If that’s not doing it for you, reminder that Toy Story and Hackers predate USB

1221

Today the desert turned into a jungle (It’s 68°F at 99% humidity rn)

1140

n=1 but I have aborted buying or renewingstuff when running into various types of slop

0120

✨ If you’re not feeling old enough, just remember that USB has been out for 30 years. ✨

6302

Browsers: “We’re now optimized to allow you 40 hours of web browsing on a single laptop charge!” Websites: “Make that 5 hours, we’re putting up these Cloudflare proof-of work pages everywhere that waste your energy to protect us from AI companies’ scrapers.”

1221

The web now consists mostly of Cloudflare verification pages and maybe occasionally actual web

2485

I find that highly unlikely but there may be specialized non-phone devices that use SIM cards that do

001

My first cell phone used the full-size credit card format in its SIM card slot, I’m that old.

2151

Konnor Rogers @konnorrogers.com · 29/09/2026

Wow haven’t seen one of these in a long time. And it’s factory fresh.

170

Reposted by Thomas Fuchs 🫯

A man standing in front of the computer saying "hack the planet", computer says "okay I hacked the planet", then the man says "oh my god."

2647

Reposted by Thomas Fuchs 🫯

Mar Hicks @histoftech.bsky.social · 29/09/2026

Wore this tshirt yesterday and someone tried to explain it to me

Star Wars number one fan. Image of the enterprise from Star Trek

3555262

Reposted by Thomas Fuchs 🫯

Courtney Milan @courtneymilan.com · 29/09/2026

If I’m reading this right, Anthropic has no choice but to IPO now because it doesn’t have the money to cover its computing / infrastructure obligations next year

1027764

Reposted by Thomas Fuchs 🫯

Amy Hoy @amyhoy.bsky.social · 29/09/2026

my toxic trait is every time i see “ffmpeg” i read it as “ffmpreg”

3543