[cardosof]

I also have this feeling of "everyone is doing the same thing, a GUI for chatgpt with some prepared prompts". The thing is, it's too risky and too soon for bigger issues to be tackled. It will take quite some time before LLMs provide medical and financial advice, if they ever will.

[famouswaffles]

There are organizations working behind the scenes doing just that. https://twitter.com/ai__pub/status/1644735555752853504

Take the gpt-4 based law ai, harvey for instance. I doubt many people know about what they're up to at the moment but they already have deals signed with some of the biggest law firms on earth with revenue quickly growing.

now realize that 4 isn't any worse in medicine than it is in law. this stuff is far closer than people think.

The hard perhaps uncomfortable truth is that GPT-4 is already proficient enough in several fields to bounce ideas off of as a colleague/equal.

[euroderf]

Will there be a public defender version of Harvey ?

No ?

Well. Just as I suspected.

[lucubratory]

Why is that something to hold against LLMs? They're making a ton of progress and will probably significantly improve the productivity of e.g. paralegals. The fact they're not legal persons and our rules against unauthorised practice of law mean they're not going to be in courtrooms any time soon even if hypothetically they were twice as good as defense attorney work (they aren't, right now). A technology can be really impressive and indeed revolutionary without making literally every job redundant. That seems like an absurdly high bar to hold anything to.

[euroderf]

Yeah but... the law is already weaponised against the underclass. I'd like to think that A.I. could change this but somehow I doubt it.

"The law, in its majestic equality, forbids the rich as well as the poor to sleep under bridges, to beg in the streets, and to steal bread."