

SHH!! Yes you can, Elon! recursively training your model on itself definitely has NO DOWNSIDES
SHH!! Yes you can, Elon! recursively training your model on itself definitely has NO DOWNSIDES
Yes! We should all wholeheartedly support this GREAT INNOVATION! There is NOTHING THAT COULD GO WRONG, so this will be an excellent step to PERMANENTLY PERFECT this WONDERFUL AI.
Nebula has some of the best educational content, especially ever since Tom Scott went on to other things. There are definitely some weird ones on there, as well as a couple which I’m convinced are right-wing shills, but by and large many of the best edutainment channels have collected over there, especially if you’re into writing/world building. HAI/Wendover, Minutephysics, legaleagle, extracredits, nilered, and hellofutureme are the big ones I’m aware of that are also popular on YouTube which publish their videos over on Nebula. It’s well worth it to know that you’re supporting them more directly, and no ads.
Also, dropout (nee collegehumor) is worth it.
Perhaps. The issue I perceive is that, for corporations, evil deeds are only illegal if you get caught and the government actually pursues you. Then, the most the corpos face is a fine, and remember: if the penalty for doing something illegal is a flat fine, then it isn’t a punishment, it’s a price.
Thus, this corporation has indicated its clear intent to sell me to the highest bidder. I would not give them a chance to do so. A “do not agree” button is just that: a “do not agree button”.
Poe’s law applies here. Remember to use your sarcasm HTML tag!
<s> Why do you hate success? </s>
Almost. A sudden shift from skill and quality to speed and turnaround wins out in the short run, just long enough to make number go up, cash the checks, and take the golden parachute out while leaving your underlings and consumers holding the bag.
Have some more comfort