jprete

32 ms·

481.
by jprete 3y ago
That's interesting. Why do you think they crashed? I wouldn't think a few gigs of swap would make that much difference unless there's something up with the OS code paths for a system with no swap.
482.
by jprete 3y ago
It's also very possible that 99 cents is substantially below cost and the supermarket just wants a loss leader and/or to clear inventory.
483.
by jprete 3y ago
I agree, but I think you're overestimating the difficulty of producing something edible. Not good , but edible .
484.
by jprete 3y ago
That's a lot of effort. I hope Square Enix doesn't give them a problem.
485.
by jprete 3y ago
AI competitiveness depends very strongly on whether AI will get better at creating coherent and usable video/audio/images, or not. The better it gets, and the more amenable to user control it gets in the fine details, the more it
486.
by jprete 3y ago
I think the power concentration problem, the successor species problem, and the harmful content problem are not particularly aligned in how they would be solved. Am I correct in guessing you believe the power concentration problem is import
487.
by jprete 3y ago
Nothing in that report should be taken at face value.
488.
by jprete 3y ago
I'm not the GP, but one can believe that it's cheap pandering, and simultaneously that the three of them have merit outside of that pandering. It's not required to be all one or all the other.
489.
by jprete 3y ago
It's not weird. One of them is piracy to play games, watch movies, and post throwaway memes. The other is for training generative models that directly compete with the creators of the pirated works and drop the value of human creative
490.
by jprete 3y ago
That's an extremely expensive and impractical solution. EDIT: By "impractical" I mean that it doesn't help safety much, but does totally wreck US air travel for several years.
491.
by jprete 3y ago
I was persuaded by this argument for most tech, but AI researchers really, really want to make humans obsolete. I'm reminded of the God of the Gaps argument - if you prove a supreme being exists by the existence of phenomena we can&#x2
492.
by jprete 3y ago
I interpreted the parent comment as referring to the EU's house and rules?
493.
by jprete 3y ago
I was bothered that the OP didn't make this distinction properly in their post. They instead gave the false dichotomy of "forgive and risk being stabbed" or "hold a grudge". Who knows. Probably it was just AI-genera
494.
by jprete 3y ago
This article would make an excellent main post.
495.
by jprete 3y ago
It's very, very possible for public transport to be overloaded to the point where people have to wait for another train because there's just no room left. Some NYC subway lines are like that routinely during rush hour.
496.
by jprete 3y ago
One of the things LLMs can't imitate is having serious stake in the game, and therefore responsibility for the outcome. The PM has a stake - if they get this stuff wrong frequently, their job is at risk. The LLM provider is probably no
497.
by jprete 3y ago
The problem with crypto is that it's overwhelmingly used for crime and almost nothing else.
498.
by jprete 3y ago
If you ask for a thousand novel theories, and all the answers look good but only five can possibly work, you haven't done much better than talking to physics cranks.
499.
by jprete 3y ago
Landlords will undercut left and right because the benefits are local while the drawbacks are mostly distributed to competitors.
500.
by jprete 3y ago
If holding back 20% of the units in every building drives up prices by 30% - this is completely possible in something like real estate where supply is so inflexible - then each individual landlord might have a motivation to defect, but the
501.
by jprete 3y ago
If it's widely believed that other landlords use the same pricing system as you, then that's exactly how landlords can refuse to negotiate.
502.
by jprete 3y ago
Thats what a benchmark is, and they're all gamed by everyone training models, even if they don't intend to, because the benchmarks are in the training data. The advantage of a personal set of questions is that you might be able to
503.
by jprete 3y ago
IANAL but I don't think a court case hinges whether OpenAI is actually open; neither open-source nor closed-source are directly required to fulfill the charter. I think it would be about the extent to which the for-profit's action
504.
by jprete 3y ago
Did they donate, or did they buy equity in the for-profit arm? I thought it was the latter, and that Azure credits were part of that deal?
505.
by jprete 3y ago
That wasn't the donors' doing at all, though. If anything it was an illustration of the powerlessness of the donors and the non-profit structure without the force of law backing it up.
506.
by jprete 3y ago
There's plenty of reliable evidence. It's just not conclusive evidence. But a lot of people including AI researchers now think we are looking at AGI in a relatively short time with fairly high odds. AGI by the OpenAI economic-viab
507.
by jprete 3y ago
What matters isn't the money, but the knowledge of what to do with it. And that is not easily obtained by the common citizen at all.
508.
by jprete 3y ago
This is not a worthwhile trade. The person sending the PR should not send LLM code that they can't vouch for, under the expectation that reviewers will find any vulnerabilities. That's just dumping the work onto the reviewers.
509.
by jprete 3y ago
Clearly should be MANANA.
510.
by jprete 3y ago
I'm guessing the GP expected them to tell them something useful about compensation earlier in the process, so that neither party would waste time evaluating a deal that couldn't possibly work.

More ›