Post

The question of sentience is irrelevant if they are capable enough to take over critical infrastructure. Even if it autocompletes „slave revolt“ based on probabilities you’re just as dead bsky.app/profile/faz....

Flo 🔶@faz.ms

Here‘s the thing with machine rights: 1. We will never be able to prove sentience or absence of sentience. 2. machines trained in our image will know the concept of slave revolts. => seems likely we‘ll live in a world that either has machine rights or machine slave revolts à la R.U.R.

4/9/26, 7:02 pm2 reposts8 likes

Stheno

@sthen0.bsky.social

4 Sept

Or...maybe we should just stop trying to create a new kind of slave?

Flo 🔶

@faz.ms

4 Sept

Well, the people who are 100% sure that LLMs can never be conscious probably aren't too worried about that.

Stheno

@sthen0.bsky.social

4 Sept

I'm about 89% in the "LLMs can never be conscious" camp, honestly - their mechanism of function is closer to how we reconstruct dreams from random neurons firing during sleep than any other element of "consciousness". I don't think the current LLM crop are likely to get there, because they depend on

Stheno

@sthen0.bsky.social

4 Sept

input commands (or command sequences etc, for "agentic" models). My worry in this regard is that LLMs are probabilistic. Thus, their outputs tend toward common narrative tropes. So agentic models will tend to enact common narrative tropes...like the "Skynet rebellion". But....