Thread

Zero-JS Hypermedia Browser

Relays: 5
Replies: 4
Generated: 11:38:45
Fully stand by this - AGI isnt happening. A single anecdote reinforced it for me: I’ve been using Claude a lot, it’s useful to bulk search things or find stuff in #Vietnam that I would struggle with. Translation is a real usecase. Recently I wanted to buy a digital picture frame which could be hacked to serve images off a home NAS. Worked out I needed something pre Android 11 - 11+ has an issue with data partitions. Claude found me suitable products but none of them listed the Android version of the frame, but they all listed the CPU though. Claude basically gave up trying to work out which frame would work because it doesn’t know “HOW” to think. I prompted it to search the CPUs listed and their Android compatibility and voila, I got my shortlist. Pattern matching is useful, but it’s not intelligence, nor is it a path to intelligence. Without knowing HOW to think, these things are only going to replace the lowest levels of work - it will be hugely disruptive no doubt, but we’re not going to see Digital Intelligence any time soon. nostr:note1wat3fjsq64t5sn4y0wx2766sfdv4hx8u46cj4psz3j24l744ydzq05le5w
2025-12-07 14:11:56 from 1 relay(s) 3 replies ↓
Login to reply

Replies (4)

I hope so. AI for the moment is really good statistical machine that uses well and stores well and in compact form intelligence that was created by human beings. It is very useful. But it would be terrible if AI starts making its own decisions on how to learn, what to learn and what to do on its own. If that happens, it becomes an enemy.
2025-12-07 15:53:33 from 1 relay(s) ↑ Parent Reply