Tom Walker

@tomw
2K Followers
794 Following
4.5K Posts

Web developer. Sometimes designer. I like code, words and mass movements. Jag lär mig #svenska.

Interested in/may post about #html #css #wordpress #openweb #decentralisation #socialism #protest

Birdsitehttps://twitter.com/tomwalkr
PronounsHe/him
Tom Walker4d ago

I think this is an interesting argument – there's a lot here but I'll pick out the idea that AI is not "consciousness" or "intelligence" but the discovery of "something completely unforeseen about the statistical properties of large corpuses of text".

I've been trying to explain lately where the illusion of intelligence comes from and I think that is basically it. It is the latent "intelligence" of a huge volume of human-written text exploring the possibility space.

https://www.theatlantic.com/philosophy/2026/06/no-artificial-intelligence-is-not-conscious/687378/

No, Artificial Intelligence Is Not Conscious

Taken to its logical conclusion, this line of thinking is absurd—and damning.

The Atlantic
Tom WalkerMay 27

Start of rabbit hole

That game is only available on long dead consoles

That game was only released in Japan, but there is a fan translation

That game has no fan translation. Many have tried but no one has finished it sadly

I am here

Tom WalkerMay 21
Tom WalkerMay 3

A lot of people don't seem to realise that anyone who knows a little about fonts can recognise those AI designs INSTANTLY.

Because they don't use a real font ever for some reason, but instead use weird fake-font lettering.

There are other tells but this one is first glance.

Show threadTom WalkerApr 22

@dan Yes – the illusion of consciousness comes from the fact that it has been trained on the entire web, all obtainable literature, and all the rest of it. But it didn't "learn" anything as such, it just put the whole lot into a massive n-dimensional matrix that it can then query.

Attributing consciousness to this seems to me like imagining that a sufficiently detailed map somehow becomes a real place.

Show threadTom WalkerApr 22
@Drarok @can Yes exactly, it is deterministic with a random number generation step. You could roll a few dice during your calculations for a similar result. Unless we are to believe that the consciousness lies in the randomness, i.e. the dice
Tom WalkerApr 21

An LLM is just doing calculations. It is a deterministic, statistical model.

You could do the calculations yourself manually. Admittedly it might take thousands of years – but it's theoretically possible.

To believe that this is "consciousness" requires you to believe that consciousness can emerge simply from doing a very large number of calculations.

Tom WalkerApr 21

"Are LLMs conscious?" is such a dumb question being asked by nominally "smart" people.

It's like inventing planes and then asking "is a plane a bird?"

No. Like, duh. Obviously it is not. It is a man-made thing that is also able to fly.

Tom WalkerApr 18
If you keep going on the TV and saying that your technology will make huge numbers unemployed, and your threat seems at least *somewhat* plausible, you are literally creating the movement against you
Tom WalkerApr 16
There is a definite connection between the level of downtime a website has and how enthusiastically the CEO talks about AI. I call this "the Github correlation"