Comment Re: Who is paying AI? (Score 1) 95
And then what, she withholds hex?
And then what, she withholds hex?
Daemon Spawn at that.
Firefox doesn't preview webp images in the file dialog on Linux either. So fucking annoying. What year is it?
I prefer the majority of canned or virtual male voices I hear on the phone to the majority of female ones today. I think this is because they care about what women think about them these days and don't care what men think, because we are less likely to complain to them about the voices. They used to choose soft female voices, now they are choosing harder ones. I am not looking for confrontation from phone operators, so this doesn't seem positive to me. I'm seeking service, not an argument.
What's next...Trans-Bots and DEI for underserved Bot minorities?!!? H1-Bot....?
It's complicated. Where does society end and the workplace begin?
You can't find Wikipedia? Learn to Internet, noob
And by extension, so do the owners of this site for providing a safe haven for these cowards
we need to regulate speech then, exactly the same way we want to regulate firearms.
Only those authorized by the state ought to have speech. Safety First!
(this is largely sarcastic in nature)
We have had human made "slop" for a long time now. Too many games and most are barely anything new for years already. Even AAA 1st person shooters which are like 1 new gun, an option for slightly different game mode, and a new different sequence of old movie plots... using an army of artists to add new graphics to the existing content library.
There are people who haven't even experienced Nintendo because they are worn out on junk rip offs with zero creativity. Which is good for industry because lacking taste means you will consume their shit unaware of what quality is. Like Disney did to Pixar etc.
Yes sort of. Back in the 1940s a mathematician did a useless paper after seeing some theory on how neurons might work in the brain, it described a mathematical way for this theory do be simulated. This was the the birth of the perceptron, something nobody could seriously work with because computers were not invented yet. It led to neuron networks and AI. Worthless academic busy work?
You don't know the impact. A bad worthless idea can inspire a great one and nobody even would know it's actual worth. Solving such things might prove useful. Sadly, solving them to show off ROBS attention, promotions, some fame?, and meeting publishing requirements... robs these from people progressing in their careers. It is true that some stuff goes unsolved because of a lack of interest for enough people to attempt it until somebody discovers a solution. The 3rd person who attempted may have solved it, but only 1 tried... That said, how many have the discipline to attempt already solved puzzles? ones that could take years?
Thanks... but... if you throw shit at a fan waiting for something... or how about a million monkeys typing on typewriters eventually producing something?
This is a higher level pattern recognizer; still a recognizer, with a huge amount more computer power. I'm sure it can crack passwords much faster with all that compute... but like passwords, cracking huge problems is a massive task. It can get lots of hints statistically from a lot of past data (human efforts, or human made passwords) so it's guesses can be more than just random; reducing the problem space.
Still, it can try a crazy amount of times and since it has fitness metrics for answers, it can narrow it's focus; it also can be stuck in localized minima (stuck in a rut) just like we do. When it's "smart" enough we can aim it at vague ideas of ways to attempt a problem and have it explore a million paths while a human might only need 100 but decades of time to do it. It's not better, it's fast enough to waste time on thoughtless pursuits and it will eventually have enough compute to beat a human. It's guess to success ratio is much worse than humans but does it matter if it's so fast it finds answers quicker?
We have simulators we use to explore questions faster than reality. Now we have a new more abstract reasoning simulator to explore faster. With more compute, it can get more random in it's millions of attempts and will stumble (perhaps, evolve is a better term since it adapts the heuristic) into new techniques. Because it's just a search, it has no real thought it'll try like a password cracker until it finds a match - it's just that it is way beyond a password cracking algorithm, but it's not different in what is going on. This whole thing reminds me of evolutionary search algorithms... we just couldn't apply those broadly or with so much compute. now we can.
It is possible that we will not need human input because as long as you can measure success, infinite monkeys will produce Shakespeare... more compute... and most importantly, better heuristics. It could be it finds heuristics better than what can be derived from human input; or maybe not, but it won't matter if you compensate with more compute. At some point human limitations will be reached and will it be able to continue without inspiration to steal from? Or will we have long stopped producing humans capable of such useful insight?? There is a reason chess was chosen as a challenge...until we found a way to best humans AFTER finding that specific heuristic.
This is stupid racist dickhead bullshit because ICE has detained at least 500 citizens so far, and every person who was "deported" to a country they didn't come from was trafficked by ICE. Fuck your support for racism in pursuit of your support of a child rapist traitor. It also makes you a traitor.
As if you would know anything about feelings
Correction:
If you do something right once, someone will ask you to do it again.