Technology

57435 readers

3277 users here now

This is a most excellent place for technology news and articles.

Our Rules

Follow the lemmy.world rules.
Only tech related content.
Be excellent to each another!
Mod approved content bots can post up to 10 articles per day.
Threads asking for personal tech support may be deleted.
Politics threads may be removed.
No memes allowed as posts, OK to post as comments.
Only approved bots from the list below, to ask if your bot can be added please contact us.
Check for duplicates before posting, duplicates may be removed

Approved Bots

founded 1 year ago

MODERATORS

[email protected]

250

"I lost trust": Why the OpenAI team in charge of safeguarding humanity imploded (www.vox.com)

submitted 3 months ago by [email protected] to c/[email protected]

42 comments fedilink hide all child comments

you are viewing a single comment's thread
view the rest of the comments

[–] [email protected] 16 points 3 months ago* (last edited 3 months ago) (1 children)

It has no awareness of what it’s saying. It’s simply calculating the most probable next word in a typical sentence and spewing it out.

Neither of these things are true.

It does create world models (see the Othello-GPT papers, Chess-GPT replication, and the Max Tegmark world model papers).

And while it is trained on predicting the next token, it isn't necessarily doing it from there on out purely based on "most probable" as your sentence suggests, such as using surface statistics.

Something like Othello-GPT, trained to predict the next move and only fed a bunch of moves, generated a virtual Othello board in its neural network and kept track of "my pieces" and "opponent pieces."

And that was a toy model.

[–] [email protected] 0 points 3 months ago* (last edited 3 months ago) (1 children)

Something like Othello-GPT, trained to predict the next move and only fed a bunch of moves, generated a virtual Othello board in its neural network and kept track of “my pieces” and “opponent pieces.”

AKA Othello-GPT chooses moves based on statistics.

Ofc it's going to use a virtual board in this process. Why would a computer ever use a real one board?

There's zero awareness here.

[–] [email protected] 2 points 3 months ago

Let me try putting this a different way: The machine is picking the next best word / action / chess move to output based on its past experience of the world (i.e. it's training data). It's not just statistics, it's making millions of learned connections between words, and through association they start to have meaning.

Is this not exactly what the human brain does itself? Humans just have the advantage of multiple senses and having a physical agent (a body) to interact with the world.

The problem that AI has is it's got no basis in reality. It's like a human talking about fantasy things like unicorns. We've only ever experienced them as descriptions and art created from those descriptions without any basis in reality.