A good primer on understanding ChatGPT from a Chinese based media perspective.
Artificial Intelligence thread
- Thread starter 9dashline
- Start date
I work in the ML/DL industry and I have zero clue who the fuck is Conor Leahy, what research is produced by his "AI Lab" & why his opinion is more important than any other random dude. As a matter of fact there're many people with all sorts of opinions - e.g. there are people on Reddit who think that Chinese GDP is actually lower than Japan, for example - so just don't pay too much attention on crazies. As for the list itself, just putting Anthropic and Character AI that have little going on for them in terms of actual research shows that this list is crap.
No, unfortunately we don't know where Brad08414464 is getting his opinion from. You'll have to ask Brad08414464 where he's getting his no doubt world-leading expert opinion from. Be sure to inform us of any response Brad08414464 gives you as I'm sure we all wait with baited breath for Brad08414464 to bless us with his profound wisdom.
You don't know Conor Leahy?! Why he's Geoffrey Hinton, Ilya Sutskever, Yoshua Bengio, and Demis Hassabis rolled into one! Not since Turing has computer science been graced with such a mind. He's a paragon of humility, though, and absolutely shuns any publicity or reknown. You might know his equally brilliant student, Brad08414464.
Sounds like Adrian Zenz of DL industry.
I do happen to know him and he's just doing the AI equivalent of SFAing, intentionally trying to make western labs think they're well ahead so they don't race. 战忽同志 o7
You either develop a model in house or buy a vendor model that has been trained against a large test database set for sometime. During the integration process you start feeding data from actual clients and gradually bring it online to end users in stages, first in 10-20% of real company data and then more and more until you can trust it completely for something like detecting frauds/hackers. There is going to be false positive rate regardless of how good your model is, however, and you always will need to have a human checking accounts the AI model marked as suspicious.
Thank you mate, my boss thinks I'm a computer science major so I should know all about computers, but I don't happen to know much about cyber security. Anyway, I finished the report she needed and I hope she likes it.
Kind of a interesting use case for gpt bots to fill online ecosystems, Its kind of interesting given monthly active users is kind of important metric to see how a mmorpg is doing.
he types very sleepily.
Conor Leahy is the chairperson of Victims of Chinese AI Memorial based in Langley Virginia.