I bet those are the same kind of people who copy paste curl www.trust.me/bro.sh | bash - on their terminal. No respected sw dev professional would run unsupervised random commands on their machine just as much they would not send unsupervised random requests on the open Internet.

These AI people are fucking morons.

And I think it is worth repeating because I don't see enough of it being mentioned when reflecting about Brexit. It was a non bing referendum. They could have done another one, especially after the first years before the deadline when people started to realize what Brexit really means. Or they could just ignore it, is not like the UK thinks twice before enacting some wide unpopular legislation (like the Internet porn ones). The people that could have stopped wanted Brexit, new how shit it would be and they just needed a scape goat, referendum, foreign influence, it is all shifting blame.

PS: That is not to say there was no interference, but that it was a welcomed one and they could have done something about it, even post referendum, because, again, IT WAS NON BINDING

It misses at least 2 more. Full on sweater + sweater pants and my favorite on winter, sweater + shorts or underwear. You pull the blanket from the feet up so it is easier to regulate warmth than with pants and shirt for example.

I was gonna reply to your direct reply to me but I think your account is important and fits well with your other comment to reply all together.

First I want to make It clear that It sucks what happened to you, I am sorry about what happened, I can only imagine the trauma and panic that such betrayal or framing could do to someone, and I am glad that there was a lot of proof and everything was clear out at the end.

I once watched a podcast about Nordic countries prisons and sentences and someone mentioned a phrase that stayed with me ever since. They said "Extreme and anecdotal cases make bad laws" and since then I think a more general expression would be "extreme cases make bad public policy". The point being avoiding gut feeling emotional reaction that may lead to worst outcome or even the opposite of what was intended. It can be good as a starting point, to evaluate if something needs to be done or change but It needs to be just that, a starting point that then goes on to be based on scientific and statistical data.

Although one of the parts lying is a really hard problem to solve in "one said the other said" situations there is a lot of scientific and statistical data to get to the policy of "Believing the woman", "A drunk woman cannot consent " and so forth. The reason being, the data shows that rape is under reported, that the victim faces a number of challenges just go forward with an accusations and that it puts a lot of risks on the victim, they are ostracized, people tend to not believe in them, they say it is their fault, etc etc. The point being, the data points out that is really costly socially and personally for a victim of sexual assault to come forward and so is very unlikely that the person is lying. Also the data shows that there is very little cases of a someone lying of being victim of SA. And none of the appeal to believe the woman is law, the slogan is an appeal to have better policy for a starting point in an allegation, for the people coming forward to be treated seriously. No one is being convicted of rape based on a single unsubstantiated account.

I know that your point is about "what if both parties were drunk", it is the same principle, same data.

Anyway,.I already talked too much, but I hope my hamblings made sense.

In my head I predicted that strawman argument, maybe I should have edited my comment to preemptively reply to it, but anyways.

Your putting the question "If I had the opportunity to punish a heinous criminal that got away with it would I do it?", actually your examples are even more "clear cut", they are "If I had the opportunity to punish and stop a heinous criminal from continuing to actively is killing a bunch of people, would I do it?" This are very different questions. That is not what I mean and I think it is pretty clear.

What temptation? It is not a donut at 2am when you are on a diet. If you find yourself with someone your hate in a situation to commit the perfect crime, would you kill them, would that be tempting? Jesus.

You make it seem like when someone is drunk they may do anything. If I killed someone while I was drunk I wouldn't just chuckle "sorry I wasn't myself". I would feel extreme remorse and want to repent for the harm I caused.

Yeah but they don't like "the gays" so it evens out I guess.

My sentiment exactly. I would even add that the executioner should be injured part, like the closest person to the victim and if they don't wanna do it them the death penalty is commuted. Or if there is no one maybe the person who has more of the part they play.

Maybe I am optimistic but I think this would increase empathy and maybe people would decide to get rid of the death penalty for once and for all.

I am not sure it would not help commercial solutions, if all experts are used all the time, sure, but if for example the usage is biased for some experts it would enable one machine to serve more users in parallel or save on VRAM or DRAM without compromising response time, hence cutting costs.

Yes, I believe it was for a Mixture of experts model, which just goes to show how naive IA implementations are at this point, The cache was not solving any hard problem and yet for some reason it was not only not already standard practice but also somehow a notable achievement.

The other day I saw a video talking about this new innovation on LLM inference side of things where they keep some more used weights in RAM and others less used on disk. I always suspected from the sample code I stumbled upon on the IA world that should be extreme opportunities for optimizations. But I cannot stress it enough how dumb the LLM world is where the basics of implementing an LRU cache is pass of as some big innovation. Like any half competent comp-sci or comp-eng professional know about the basics of mitigating this basics bottlenecks like "the data does not fit on available RAM", "The disk is slow", etc.

So is not surprising that now that it seems that the "powerfulness" of this LLMs is starting to plateau that we would start to see some improvement in performance/resource utilization and hence running costs.