V0ldek

joined 2 years ago
[–] [email protected] 10 points 1 year ago* (last edited 1 year ago)

this stuff is still relatively new and I’m sure it’ll get better with time

We'll engrave it on LLMs' tombstone, right next to blockchain and its "it's still early".

[–] [email protected] 17 points 1 year ago (1 children)
[–] [email protected] 19 points 1 year ago (3 children)

What you are basing this “it clearly cannot” on?

I asked Gemini and it told me that ChatGPT can't do shit, I'm not gonna question it.

[–] [email protected] 10 points 1 year ago* (last edited 1 year ago)

Give Ken Thompson and Dennis Ritchie billions of dollars

I mean, if we took all net worth of Sam Altman and split it between these two guys who at least benefited humanity with their work we'd get at least a step closer to justice in the universe.

Getting a Turing award: $1M

Dropping out of Stanford to work on something unironically called "Loopt": Priceless

[–] [email protected] 19 points 1 year ago (5 children)

I'm not sure what you even mean by "how is it different", but for starters a human can actually get a good mark at the bar and spicy autocomplete clearly cannot.

[–] [email protected] 11 points 1 year ago

Steelmanning what this person said, I think the issue is that your ability to CTRL+F through a book during a time-limited exam is not as strong as even a single computer clocked at GHz doing the same thing. You can CTRL+F through a single book in the same time it takes it to CTRL+F through the entire body of knowledge.

[–] [email protected] 8 points 1 year ago (5 children)

Ye but I don't think the meme applies here? This particular lie is statistical malpractice, not "someone else wrote the bar for it".

[–] [email protected] 8 points 1 year ago (1 children)

AGI = A Group of Indians.

[–] [email protected] 20 points 1 year ago* (last edited 1 year ago) (14 children)

I'm not even going to engage in this thread cause it's a tar pit, but I do think I have the appropriate analogy.

When taking certain exams in my CS programme you were allowed to have notes but with two restrictions:

  1. Have to be handwritten;
  2. have to fit on a single A4 page.

The idea was that you needed to actually put a lot of work into making it, since the entire material was obviously the size of a fucking book and not an A4 page, and you couldn't just print/copy it from somewhere. So you really needed to distill the information and make a thought map or an index for yourself.

Compare that to an ML model that is allowed to train on data however long it wants, as long as the result is a fixed-dimension matrix with parameters that helps it answer questions with high reliability.

It's not the same as an open book, but it's definitely not closed book either. And the LLMs have billions of parameters in the matrix, literal gigabytes of data on their notes. The entire text of War and Peace is ~3MB for comparison. An LLM is a library of trained notes.

[–] [email protected] 13 points 1 year ago (1 children)

For the record, I'm currently at ~70% that we're all dead in 10-15 years from the moon getting mad at us. I've stopped saving for retirement, and have increased my spending towards a future moon mission to give it lots of pats and treats and tell it it's a good boy and we love it, please don't get mad.

[–] [email protected] 5 points 1 year ago

Nope, it's dumber than that, the memecoin conspiracy dipshits were activated by him liking a tweet on the tenth and have already speculated that he had to be in a secret federal jail for the whole time.

Don't take my word for it

[–] [email protected] 7 points 1 year ago (1 children)

I hope he had a real long talk with his lawyers to be 100% sure this also won't count as market manipulation.

view more: ‹ prev next ›