There was a literal newlywed married couple in space at one point though: Jan Davis and Mark Lee--they married in secret and told NASA when it was too late to reassign them. NASA claims they didn't bang up there but COME ON, they at least tried.
Don't know that I've ever seen it with a Senator though. Some states have laws that extend the deadline if an incumbent does this, should be universal.
Ahh, yes I think so.
Not sure printers rise to the level I'm thinking of. PCs do, but you're right no major crash there, though I think quite a few companies went under in the 80s? I guess they had a more gradual rise though.
A lot folks (Zitron) motte and bailey between the two
There's really no conflict between "these are economically revolutionary technologies" and "this is a bubble" I can't really think of any distinct novel technology that transformed an economy and didn't have a bubble along the way: Railroads, containerization, the internet....
4. I think that it is important that critics of LLMs get this. The 'LLMs are useless which is terrible and everyone is using them which is also terrible' shtick contained contradictions even in the beginning, which took a lot of work to maintain . Now it contradicts people's lived experiences.
I was once told that it really just means "down with" and "death to" is an overly literal translation...
Oh it wasn't just once!
Interesting! Does seem like there's a real gap turning a lot of this stuff into actual products. But that will inevitably close I suppose, especially as the cost of running these models goes down...
I am not saying it's unfair to respond to his question. I'm saying it's unfair to assume that he thinks the answer is "the ABA is more partisan than fed soc."
I listen to his podcast and Will at least attests a kinda panglossian assumption of good faith on everyone's part, even when he acknowledges the possibility of other explanations...
The rest of the post also identifies a variety of responses to his question and doesn't take a position on which one is correct. People assuming his views from this excerpt are not reading the whole thing fairly.
President hasn't run Iran for a while if ever. NYT reported that it has been Larijani in effect for at least the last 6 months and he still seems to be doing it.
I'm just thinking about how I encounter voice transcription on my phone / see translation when viewing foreign sources (primarily google services). Possible what's available hasn't kept up with the state of the art.
Curious what the "AI will be as good as humans at everything done on a computer" folks say about that relative lack of progress.
I wonder if this is an area where it will make a lot of rapid progress but then plateau. I'm obviously no expert but I feel like this has happened with voice recognition and translation? Google got much better at those years ago but hasn't ever fully caught up with humans?
I think it could definite be decomposed into parts to some extent: there's issue-spotting, understanding the applicable legal rule, applying that rule to the facts, among other pieces. There's also a big difference between different kinds of cases... you'd probably need many different custom skills.
I think we have an example of them succeeding up thread though! Maybe not perfectly but I can assure you humans are far from perfect too.
The thing is references finding or cite checking are things where we can easily validate whether the LLM has gotten it right. "Is this a convincing argument that effectively uses the caselaw" is gonna be a lot harder to assess. But it can eat a lot of other parts of the job before it gets there...
I mean it's a test you gave him? A bit rich if he gets it mostly right (better than plenty of humans I've worked with) and then you are like "ah, but this proves nothing." I mean I am still skeptical argument is close, but you picked the test!
OK! As I said I'd gotten a bit lost in all the threads. Better performance than a lot of humans I've worked with tbh.
To put it another way, what you think of as "finishing it" she thinks is the hard part. Idk who is right! bsky.app/profile/kath...
I suppose I'll have to try it out sometime myself when I have more time.
I understood her to be claiming: "distinguishing between references and citations is hard and requires human understanding that AIs can't replicate" and you to be disagreeing? Prompt seems like a reasonable attempt to accomplish that but I didn't see if she thought it had managed it.
What? I meant this decision is bad. But it's an appeal of a state court order redrawing a district ostensibly for racial equity reasons. That's not a gerrymander, unless your position is that the state courts were gerrymandering to benefit Dems and lying about it.
I am a bit confused about what anyone in this convo thinks has been proved--she thinks distinguishing between references and citations is hard, and your skill didn't. You think it would be trivial to explain that distinction. I am not sure. But aren't we just back where the convo started?
Friend on discord showed me this and I thought a few people who enjoyed this person's comics might like to know how they post over on the Hitler site.
Has there been any reporting on if the quisling firms have actually been doing anything to comply with the deals? Providing legal services for free, I mean...
