You know why they “tombstone”? (By the way, they don’t replace with a tombstone marker but instead add the marker.)
Because if you “accidentally” deleted something and then decided you wanted it back, you’d get really mad if they couldn’t do that. If they immediately deleted it, you couldn’t ever get it back
The copies and deep copies are for a similar reason: Some engineer accidentally deletes a bunch of data, it’s really nice to have a backup so you don’t lose everything.
But like… deleting the data would lessen the sale price. Much easier to just delete your account and keep the data in an “anonymous” form. How are you (as the consumer) going to ever know if it’s actually deleted?
oh. go get a therapist–not physical; mental. they’re insanely expensive, but you can spend the next three months shopping around and by the new year you’ll have found someone you like!
invite me and I’ll bring my own alcohol. spread looks delicious!
Another great example (from DeepMind) is AlphaFold. Because there’s relatively little amounts of data on protein structures (only 175k in the PDB), you can’t really build a model that requires millions or billions of structures. Coupled with the fact that getting the structure of a new protein in the lab is really hard, and that most proteins are highly synonymous (you share about 60% of your genes with a banana).
So the researchers generated a bunch of “plausible yet never seen in nature” protein structures (that their model thought were high quality) and used them for training.
Granted, even though AlphaFold has made incredible progress, it still hasn’t been able to show any biological breakthroughs (e.g. 80% accuracy is much better than the 60% accuracy we were at 10 years ago, but still not nearly where we really need to be).
Image models, on the other hand, are quite sophisticated, and many of them can “beat” humans or look “more natural” than an actual photograph. Trying to eek the final 0.01% out of a 99.9% accurate model is when the model collapse happens–the model starts to learn from the “nearly accurate to the human eye but containing unseen flaws” images.
oooh. design intricate sandwiches! sounds like a lovely holiday!!
Favourite part of the whole article:
A spokesperson for Truth Social said, “It’s hard to believe that Reuters, once a respected news service, has fallen so low as to publish such a manipulative, false, defamatory and transparently stupid article as this one purely out of political spite.”
“You never saw what you thought you saw. And even if you did, it was entirely justified and your interpretation was extreme.”
I’ve got a problem with articles like this: “The guy who got it right once is betting a second time he’s going to get it right”. and then the article continues: “Even though he’s got it wrong a bunch of times since, he got it right that one time… So this has gotta be his second time!!”