zlacker

I can't wait for the revel that GPT-4 is just a chat application connecting you to a the worlds biggest center in India.

Joking aside, this feels massive. Both that it happened so suddenly and that the announcement doesn't mince words. The fact that the CTO is now CEO makes me think it's probably not a lie about their tech. It wouldn't make sense to say "we've been lying about our capabilities" and then appoint the current CTO as CEO.

This makes me think it's either financial or a scandal around Sam himself.

I can't wait to hear more

>>superf+(OP)
> The fact that the CTO is now CEO makes me think it's probably not a lie about their tech.

Agreed

> This makes me think it's either financial or a scandal around Sam himself.

I can't imagine it being about fake financials. This isn't Microsoft's first time doing due diligence on a acquisition. That is both technical and financial due diligence.

And clearly they didn't buy the company because it was super profitable, but for the tech.

>>superf+(OP)
Then why is Greg Brockman stepping down as chairman of the board?

>>eiiot+e2
No idea, but it's important to note that Brockman is still at the company. Just no longer Chairman.

Maybe he's not involved in this, but it's a "you should have known" type of thing? Really no idea...

>>sparkl+p1
Microsoft didn't buy them did they? I thought it was just an investment. Either way though you're right that they probably did their DD.

My first comment wasn't really about them not being profitable, it was more of a question about how close to bankruptcy they are. Again though, you're right that MSFT probably did their DD, so that's unlikely

>>superf+(OP)
Ilya is also on the board and he knows the tech better than Sam

>>coffee+H7
Ilya is religiously devoted to AI enough to overlook any sort of fraud or misrepresentation in the company.

>>superf+67
Correct. Strike "buy" and replace with "massive investment"

>>superf+(OP)
> I can't wait for the revel that GPT-4 is just a chat application connecting you to a the worlds biggest center in India.

Tempting, but you can launch Mistral-7B on your laptop and see the same coherent responses and reasoning from a 8GB model file. The magic is really there.

>>atemer+wd
Also, the technology of finding the right set of human experts on basically any topic in milliseconds in India would be a massive feat in of itself

>>atemer+wd
Did you try with your Internet off?

>>atemer+wd
mistral 7b foundation model is the first AI related thing that excites me since the launch of ChatGPT. shows us that this tech is poised to be commodified, self hosted at a low cost eventually and won't necessarily be tied to and controlled by tech billionaires.

>>superf+(OP)
Leaking secrets to Tim Cook?

>>dkarra+He
In fact, Mistral is so exciting that I think it has some direct relationship to the topic of this discussion (they still don’t talk about how they trained it, at all).

Perhaps there was some, how to say it, unexpected technology transfer, with Sam somehow involved.

>>superf+a6
He might not be directly involved in whatever this is, just they asked him to step down after voting in favor of Sam.

>>atemer+Fg
interesting take, we'll see how things evolve.

>>atemer+wd
Theranos level fraud combined with the impossibility barrier?

OpenAI fakes their AI to be first to market and then everyone else, thinking that it's possible, finds ways to do it for real.

(this is 100% a joke. I know OpenAI's work is the real deal)

>>atemer+Fg
First good take I've seen. You don't get fired for stealing other peoples IP or for leaking chats, that's just the name of the game. You do get fired if you sold off OpenAI's IP (or gave it away because you're an idealist)

>>superf+(OP)
Right? haha or Sam himself answered every query from his phone lol... that's why sometimes it feels like it's typing more slowly than other times haha

>>huevos+be
lol yea. The idea that someone would be convinced after running an LLM on their laptop and not after seeing ChatGPT expertly respond at 400 WPM in 0.2 seconds, is funny.

>>superf+(OP)
Yeah, it would be pretty surprising to find out that India has so many individuals that are profoundly competent at converting arbitrary text into rhymes in Elizabethan English.

>>chanks+jo
Holy crap... Is ChatGPT just ChaCha for GenZ?

https://en.wikipedia.org/wiki/ChaCha_(search_engine)

Seriously though... I just remembered this was a thing and now I'm having crazy nostalgia.

>>davesq+Qo
Mechanical Turks be hustlin'

>>davesq+Qo
I know this comment is just adding on to the joke, but the number of people who can speak / read English in India is very high. There are more English speakers in India than multiple large European countries combined.

>>govg+hw
Maybe so, but Indian English practically its own dialect at this point.

Nothing wrong with that, but the responses I get from ChatGPT are not in that dialect. (Who knows, maybe by paying for ChatGPT Plus I qualify for an upgrade to the Irish data center.)

>>superf+au
I remember signing up to do that in college to make extra money and them not accepting me. Bah!

>>govg+hw
I said Elizabethan English, not English. As in Shakespeare. The point is that there are probably at most a few thousand people in the world who could display that level of competence at making Shakespearean rhymes. And I doubt enough of them work in call centers in India to service the entire user base of ChatGPT. Nothing about my comment implies that there are not millions of competent English speakers in India.

>>davesq+QR
Realistically there could be a mix of genuine LLM magic combined with heavy on the spot curation by massive teams of incredibly smart people.