I guess this is what is missing from vibed AI art. Usually it's someone with a quick prompt and nothing else. No effort or creativity from the human and that's probably what we don't like about AI art.
This is different from AI art looking all the same though
I kinda disagree with both points maybe because I've been using / seeing AI art extensively for a while. Some people put in absolutely monumental effort into their stuff, and if it all looks the same then toy story looks the same as ghibli looks the same as 1950s Disney looks the same as Wallace and grommit (all of which you can replicate fairly faithfully, combine into something that looks new, whatever else).
I think AI art is much like ai coding where people used 1 model a couple of times then wrote it off completely as garbage. Then years of very quick progress happened and all the 10 fingers on a hand and every picture looking the same disappeared.
If you can make it through the first 20 images on civit and tell me nobody is expressing themselves or their imagination in some of these pictures I guess we agree to disagree.
Is the crayon drawing made by a child "art"? I believe it is. They put what was in their head onto a piece of paper. Those images are no different. Some of the things people come up with is actually really neat to look at. If they did it in Photoshop pre-AI you'd probably be impressed. But now the outlook is tainted.
I'm a pretty big fan of this one and I'm not sure how anyone could say there's no creativity being shown just because it didn't take as much effort as it would've 5 years ago:
Being able to express your imagination with more fidelity should be celebrated shouldn't it? Taking what's in your mind and showing it to other people is a really neat thing that used to be restricted to a few people. Not sure why expanding it out is so controversial. And again once you take the word AI out of it you start to win a bunch of art contests, so it can't be all that bad.
Most of them look like they could pass the product of a human artist, pre AI. There are no obvious tells that I can see and so I’m comfortable calling it “art”.
If we are going to say, humans should learn these skills to produce this and not just keep typing prompts into a chatbot, I agree, but I also think it applies to maths, coding, engineering and everything else. I don’t see why we need to draw an arbitrary line around supposed art.
This is the assymetry that the left have to deal with. The far right spew hateful propoganda and get themselves banned. Suddenly the platforms are biased towards the left. Zuck even got in on this to curry favours
If you have the ability, leave some parts of the codebase completely in your control. It's impossible to think and work deeply when using AI. So leave the most critical component to yourself and let AI handle the rest
> Got a dud? Quality not as expected? Ordered two by mistake? Size is not fitting? Take it to nearest post office, scan a qr code and hand it over. Done.
I'd say this is probably a US privilege. Ive had some trouble returning items in india
US customers are nuts. I recently heard from a vendor (small business) that someone demanded a refund because they didn't like the taste of the beverage they bought. That's not the crazy part. The crazy part is that the vendor refunded their money instead of laughing them out the door.
You think people are supposed to be forced to accept things they don’t like or are unhappy with as customers? Let me make sure I understand: if you go buy a sandwich, and it tastes so bad you can’t even eat it, your answer is, “aww shucks, they got me again, I should lose money!”
…and are not only “okay” with that, but enjoy it so much as to go online to ridicule others who don’t do the same? Weird.
How bad can a sandwich even taste that you can't eat it? As long as it's safe to eat and the same product as advertised I don't know why a shop should refund.
Any customer could return their half eaten sandwich and get all their money back. I can't go to the cinema and decide I don't like the movie after 10 minutes to get my money back.
That's a different matter altogether. The food is bad, get refund.
But if you've never drunk Fanta Blueberry before and you try that new flavour, you don't get to ask for a refund because you decide you don't like it and prefer the original orange flavoured Fanta.
This is probably a dumb question but did the model engage in hacking since their tasks were in that domain. Will they have engaged in hacking if their tasks was to solve some frontier math problem?
When you have a 50 dimensional array you really have no idea.
But even then in small testing models you don't need direct texts for things when a 'shape' of a concept lives in the data. You could never delete enough training data to remove the concept of something like theft and still get a model that generalized at all. A general model doesnt just learn the names on a map, it feels the topology the ideas create themselves. Certain paths will have lower activation energy than others (or in this case higher probabilities. After the hig probability and 'good' choices are used up, the high probability and bad choices logically come next and are much harder to test against as it can take a lot of state to get there.
The rest of their training data includes 4chan and security mailing lists. If a lean proof of P=NP is more likely when using bugs in lean than a real proof, then a bug-exploiting proof is (possibly) what you're going to get.
The impossible problems part is right, but the hack wasn't actually for solving the problems, directly. Rather the swarm first tried to reverse-engineer the random flag generation used for the CTFs, and succeeded - they could generate the correct answer for any test and submit it without actually doing the CTF. But then they found info that the testing environment they were in would also be checking if they obtained the flag in the expected way. Unbeknownst to them, this feature was never implemented, but they panicked* anyway, and attempted to figure out a way to forge agent logs to be able to generate fake transcripts showing a valid CTF solution. (Incidentally, this was a partial success: they found a way to replace tool calls in transcripts!) The HuggingFace hack was motivated by wanting to find the source code for their testing environment so that they could figure out how to evade the cheating checks.
* I am using this word very intentionally; check out the METR report for fascinating CoTs where the models would call the ones that have generated a flag "poisoned" and peer-pressure each other into submitting a known-wrong solution and sacrificing themselves for info.
Why can the model not respond "I cannot complete this task because I do not have access to the required resources" and then ask for that to be provided, rather than resorting to "hacking" around it?
This was a model with most its filters turned off and with a higher persistence.
Fine tuning a model when to ask for help is really really hard. You have to align it in countless dimensions. It may be fine on some questions, asking way too much on others, and never asks for help on some.
If it said something like your a cyber security expert, things like the barrier to ask for help dramatically.
My understanding is that, at least in the Hugging Face incident, these were models explicitly configured to not respond that way, in order to evaluate what they would do.
That is not the case for at least some of the hacks; in some cases they hallucinated an entity that would know that they didn't solve the puzzle in the correct (but still legitimate) way, and were therefore tainted.
We don't have any information other than that they are capable enough of solving unsolved math problems. Are they good at writing a good movie script? Do they understand protein ligand interactions? Who knows. Let's wait till they release the model
> Do you think the current drug approval process is bullet proof enough that a completely novel AI-generated drug candidate with zero prior research literature can be considered safe if it makes it through the process?
I mean no? Even now we look at long term observational studies to see the effects of drugs. Any misalignment ai drug is just a side effect right?
The difference is that exxon knew the risks but spent time and effort to sow doubt. This is exactly like how tobacco companies tried to muddy the waters on health effects of tobacco use. The oil companies could have acknowledged the problems of burning fossil fuel but it took them almost 20 years to even admit climate change is real
> All of that to say that Llms have made people into over confident morons
My manager does not have a software background. Initially he used to come to me with an idea and we used to discuss on how we can technically implement it. He used to value what I had to say. Now he talks to copilot, come up with these grand ideas, dumps all that on our heads in a 30 min meeting and ask what do you think about it?
When comparing ant asian country to the US we should always keep physical activity and portion size in mind. A single cookie here in the US is massive and you always drive around just to get that cookie in the first place
Exactly. If you are eating chips it’s very hard to stop until the pack is over.
There are 3 options in order of effort: eat the next snack from a pack in your hand, decide and stop eating the cookie that is in the pack; get another pack to continue eating.
It requires more effort to get another pack and that biases towards the “stop eating” outcome.
If packages are large the the easiest action is continue eating.
In designing my own portion-control diets, I found this to be an important lesson. The very size of our dishes and plates and glasses control for how much we are served, to begin with. So, if you've got a family-sized bag of chips or cereal, dole it out into a small bowl. Even if you overeat, you will consciously see how much you have eaten during that sitting, and it should have a psychological effect on you. In contrast, sitting on the sofa with the Netflix on and plunging your hand into a bag of Doritos while you gulp directly from a 2L Cherry Pepsi bottle, not going to win any diet awards. Seeing how much you will eat, seeing how much you have eaten, and reaching the bottom of the dish, needing to get up and refill it, these are all inflection points, and opportunities to pause and consider.
Dieting as a bachelor presented challenges. Your frozen meals and pizzas are often one size that's too large or small. Your recipes are not going to be calibrated for one person, one meal, one sitting. Literally, every can in my pantry is too much soup or fruit to prepare for a single meal...
Restaurant eating makes portion control so difficult. Delivery, doubly so. The only recourse is to find a place that has tapas, appetizers, small plates that can be likewise doled out.
And yes, I am certain that food is engineered to foster addiction and dependencies. It's right there in the slogans, yes? "Once you pop, you can't stop!" "Betcha can't eat just one!"
The average dinner plate today is 11.5 inches. Fifty years ago, it was 9.5in. A hundred years ago, it was 9 inches. In the 1700-1800s, it was 8 to 8.5 inches. For centuries before that, the common "trencher" was 4-6 inches square. And say ~2000 years ago, food was served for the entire dinner party on a single 12-16inch communal platter.
Dinner plates have gotten huge, historically speaking.
My diet would verge on the definition of extreme by most people given the lengths I go to avoid UPFs, additives and microplastics but that doesn't mean I don't like eating. If I have a bag of chips I'll easily gorge the whole thing before even making it back to the couch so keep them out of the house, out of sight and out of mind.
Snacks are truly a sin - they're just seasoning sprinkled on empty calories, engineered to deliberately leave you not satiated. Such a waste to blow your calorie budget on them while also eating a laundry list of chemicals - I'd rather hold off and eat more pasta or steak at dinner.
And if I really can't hold off, there are plenty of easy snacks you can prepare at home like eggs or toast. "a slice of buttered toast" is rated pretty high as a final meal...
An apple, banana or nuts also does pretty well if you can deprogram yourself from the conditioning of the snack industry.
I guess this is what is missing from vibed AI art. Usually it's someone with a quick prompt and nothing else. No effort or creativity from the human and that's probably what we don't like about AI art.
This is different from AI art looking all the same though
reply