diputsmonro

Some humans do - the investors and executives in AI tech companies (and the legislators who theoretically could regulate them) , who all stand to make a lot of money from every one of the "AI danger hysteria" scenarios, and are therefore highly motivated to bring them to fruition.

The rest of us have no choice. Despite millions of artists, animators, etc. all being resoundly opposed to AI art, the models that infringe their work are still allowed to exist, and it seems they're fighting a losing battle.

A lot of people are being "hysterical" because a lot of people don't have a choice.

To be clear, the problem of these scenarios is tightly intertwined with the problem of unfettered capitalism and wealth inequality in general. Food and shelter require money, and we get money by working a job. If millions of jobs disappear overnight, then of course millions of people are going to be distressed over no longer having ready access to food and shelter.

The idea of "just getting another job" doesn't scale to the destruction of entire industries employing tens of millions of people. This is how depressions are made.

The idea of "the depression will end someday" is not only not necessarily true as wealth inequality skyrockets, but is also cold comfort to the people who will lose their houses and for some, lives, due to the disruption.

A different economic system could perhaps allow us to appreciate these technological advances without worrying about them displacing our ability to live. But the American political system consistently and firmly rejects any ideas not rooted in social darwinist capitalism.

For your sake, I hope your resume is very impressive.

Yeah, it would be way better if they just released it right away, so that political campaigns can use AI generated videos of their opponents doing horrible/stupid things right before an election and before any of the general public has any idea that fake videos could be this realistic.

Yes, but the point is that in a few years, there won't be a difference. Those clickbait accounts already exist for AI generated images. How many impressionable or young people have been fooled into believing history that never happened?

More importantly, how can these accounts subtly direct the generations to instill modern ideology or politics into "historical" images, giving them historical credibility? Think of all the subtly white supremacist "retvrn" accounts, for example, falsely recontextualizing inventions and accomplishments to support their ideology.

We all need to be thinking much more creatively and cynically about how these tools will be abused. The technology will get better. The people who want to abuse it will get smarter. And your capability to distinguish fake information is likely much worse than you believe - to say nothing of younger people who have less context and experience to form a mental "immune system".

My guess is that offenders abuse relatives because they are easier to access and manipulate, not because there is a true preference there. More a crime of opportunity than a pursued goal.

I think you're making the same point, but just to make it more obvious - all of that is more an indictment of modern neoliberal business practices than it is of Bob's Red Mill. Sustainability, efficiency (in terms of doing things right, rather than doing them quickly) and overall most benefit for the most humans is better than the get-rich-quick schemes that most startups have become.

I think people are being incredibly disingenuous when they imagine that the government won't abuse this power to censor and harm marginalized communities. Many states are trying to remove LGBT books from school libraries for being "pornographic" right now, for example. All it takes is some fancy interpretations of "safety" and "social media" for it to become a full internet blackout, for fear of "the liberals" trying to "trans" their kids.

I don't deny that kids can get into trouble and find shocking or dangerous things online. But kids can also get in trouble walking down the street. We should not close streets or require ID checks for walking down them. Parents should teach their kids how to be safe online, set up network blocks for particularly bad sites, and have some kind of oversight for what their kids are doing.

Maybe these bills should mandate that sites have the ability to create "kid" accounts whose history can be checked and access to certain features can be managed by an associated "parent" account. Give parents better tools for keeping their kids safe, don't just give the government control over all internet traffic.

Is talking to other people online really so bad that we need the government to step in and tell us who we can and can't talk to? How quickly will that power expand to what we can and can't talk about?

I agree that neither solution is perfect, but exchanging an imperfect but undoubtedly free system of communication for one that is explicitly state-controlled censorship is an obvious step backwards.

"Thinking about the children" should also involve thinking about what kind of a society you want to build for them. A cage is not the answer, especially not with fascism creeping back into our politics.

The biggest problem with this is how we would define "porn". Some states are currently redefining the existence of a transgender person in public as an inherently lewd act equivalent to indecent exposure.

I have no doubt that if your proposal were to pass that there would be significant efforts from extremist conservatives to censor LGBT+ communities online by labeling sex education or mere discussion of our lives as pornographic. How are LGBT+ people supposed to live if our very existence is considered impolite?

Nevermind the fact that the existence of a government database of all the (potentially weird) porn you look at is a gold mine for anyone who wants to blackmail or pressure you into silence.

The horrors and dangers of porn are squarely a domestic and family issue. The government does not need to come into my bedroom and look over my shoulder.

"Tolerate" is a funny word, as though consumers actually have a choice about what price they pay and/or can forego buying groceries until prices fall, if ever. I can't simply choose not to eat for an undetermined period of time. Consumers do broadly seem to be upset, but they don't realistically have power to change anything.

I would compare the AI in the "inventor" question to a set of dice. Suppose you have a set of dice that when rolled generates an idea for an invention. If an idea is rolled that is unique enough to secure a patent, who is the "inventor"? The person who rolled the dice? The person who manufactured the dice (and therefore the idea generation algorithm)? In any case, attributing the invention to the dice themselves is obviously silly.

In cases like this I think it comes down to who is the first person to meaningfully act to bring the idea to fruition. Rolling the dice in isolation is meaningless. But deciding that a particular rolled idea is worth patenting (and therefore acting as a filter to disregard all the generated ideas that aren't worth pursuing) is a meaningful contribution. By sheer act of pursuing the patent, Dr. Thaler has made himself a fundamental part of the creative process.

The AI here is no different than idea dice. Nor is it different from any program that jams two words together in hope of producing a "novel" idea. The only difference is that the AI, hopefully, produces less chaff to sift through. Still, the sifting is a creative process.

I remember it sometimes being in the options menu and sometimes not (and sometimes showing up but disabled). Maybe there were a different combination of versions or graphics cards that produced those different results. Perhaps earlier versions hid the option if your graphics card didn't support it, and later versions showed it but grayed it out.

If the self driving system were worth it's salt, it wouldn't matter if the drivers weren't paying attention. Ergo, the system sucks, or is at the very least not nearly as good as Tesla likes to tout.

I understand where you're coming from, but the effect of rules like that is to essentially create a system where spying on people with towelhook cameras is totally legal and easily doable unless you happen to get caught.

Most laws could be reduced in the same way, but a lot of dangerous illegal activities have proactive protections, like regulations on purchasing and selling guns, etc.

Small cameras as a class are probably too broad not to sell as a rule, but cameras sold with a specifically designed case to pass them off as other objects might be a step too much. Of course one could always drill a small hole into anything and add a camera, but we shouldn't just be giving the offenders the tools so easily right off the shelf.

Depends on the coworker. I would say that I trust ChatGPT code much less than the average coworker, because coworkers usually don't do something like use imaginary variables or check-in code that doesn't compile.

As a software engineer and artist, Krita feels more focused on drawing/painting with a pen, while GIMP has always felt to me as more focused on photography/editing.

Obviously they both manipulate images so there's lots of overlap in features, but the idea of painting or drawing in GIMP seems really alien to me. I'm sure the interface and pen support was even worse when the Krita project was started.

Actions speak louder than words, and should give you pause to re-evaluate your trust

The idea of my agent and Google's customer service agent babbling at each other in endless futility and failing to solve a problem that two humans could solve together in 15 minutes is a dystopian hell. But it will be celebrated as a wonderful productivity boost because no humans will be on a phone call.

A coincidence.

Like, it's a bit sarcastic, sure, but until factuality is explicitly built into the model, I don't think we should use any terminology that implies that the outputs are trustworthy in any way. Until then, every output is like a lucky guess.

Similar to a student flipping a coin to answer a multiple choice test. Though they get the correct answer sometimes, it says nothing at all about what they know, or how much we can trust them when we ask a question that we don't already know the answer to. Every LLM user should keep that in mind.

It is still a hallucination even if the words it hallucinates happen to line up with a factual sentence, in the same way that a broken clock happens to correctly display the time twice a day. The function of the clock does not suddenly begin working correctly for one minute and then stop working correctly the next. The function of a broken clock is always flawed. Those broken outputs, by pure coincidence, just happen to be correct sometimes.

LLMs are broken in the same way. They are just predictive text generators, with no real knowledge of concepts or reasoning. As it happens most of the text it has been trained on is factual, so when it regurgitates that text it is only by happenstance, not function, that it produces facts. When it hallucinates a completely new sentence by mashing its learned texts together, it's pure chance whether the resulting sentence is truthful or not. Every generation is a hallucination. Some hallucinations happen to be sentences that reflect the truth. The LLM has no ability to tell the difference.

Completely agreed, and honestly, sometimes we need to have frank conversations about true horrors and abhorrent behaviors that exist in the world. I hate to pull the Orwell card, but I fear that the normalization of this technology would inhibit necessary conversations about true evils, and give a false impression of civility and neutrality to people spreading hateful ideology.

That said, most online discourse is already a disaster, with every disagreement being ramped up to 11 with no capacity or patience for nuance, so I'm not sure how much worse it can actually get.

How could you possibly call the largest encyclopedia in human history, freely accessible to all, and correct nearly all of the time irrelevant?

There are imperfections to Wikipedia and the Foundation of course, but, what more can you reasonably expect? If you want old internet fun, we have neocities.org for that. Wikipedia serves a functional purpose and does is exceedingly well in most cases.

I'm not really sure how these are related at all. Technical innovations for electric vehicles are totally different than the technical innovations required for fighting climate change. We can be poised to make progress in one sector and skeptical about progress in another. Blindly hoping that some breakthrough will happen to fix all our problems and just barreling along carelessly is worse than what you're describing.

Cheese???

A slightly different reframing of this idea makes it even more clear to me:

When you make a choice, you divide the doors into two sets: the set containing your choice, and the set containing the other doors. Obviously your set has a 1/3 chance of containing the prize, and the other set has a 2/3 chance.

When the host reveals a door in the second set, it has no effect on those initial probabilities. The set as a whole still has a 2/3 chance of containing the door, and now you have the choice of selecting the only item in that set which you know isn't wrong.

This reasoning is just as obvious when you scale up the doors too. With 100 doors, your set has a 1% chance of containing the prize, and the set of unchosen doors a 99% chance. All but one door in that set are revealed to be wrong, then you get a "50/50" choice of swapping to that last remaining item... of course that's what you want to do!

I agree that it feels weird at first, but thinking about it this way fixed that for me:

The whole reason for the apparent paradox is having the "50/50" choice at the end. You either keep the door you picked, or swap with the other door, so it feels like no advantage.

If they only revealed a single door, then you would have multiple doors to swap to, so you wouldn't have the "50/50" choice. The only way to keep the "50/50" choice at the end is to reveal all but one door and give you the option to switch to it.

Victimhood can be passed through generations if the parents' harm was not remedied. Merely ending harm does not remedy it. If someone is prevented from getting a home loan, job, raise, education, etc. because of racist policies, that absolutely affects the kind of life and opportunities that their children will have.

As an analogy, if I have stolen money from you for years, stopping me from stealing further money doesn't repair the damage I've done. You would rightly expect your money back, or something of comparable value.

Affirmative action programs are specifically designed to seek out and uplift people who have been generationally affected in that way. It recompenses them by giving them job/education opportunities that they would likely have had if their parents (and their parents, etc.) weren't artificially held back.

I think you have a bad media diet if you think any of those are actual problems in the real world and not just straw men made by provocateurs stirring the pot.

Honestly though, I would prefer an AI that was strictly neutral about anything other than purely factual information. That isn't really possible with the tech we have now though. I think we need to loudly change the public perception of what chatGPT and similar actually are. They are fancy programs that create convincing hallucinations, directed by your input. We need to think of it as a brainstorming tool, not a knowledge engine.

Yes, thank you. AI is dangerous, but not for the sci-fi reasons, just for completely cynical and greedy ones.

Entire industries stand to be gutted, and people's careers destroyed. Even if an AI is only 80% as good, it has <1% of the cost, which is an ROI that no corporation can afford to ignore.

That's not even to mention the political implications of photo and audio deepfakes that are getting better and better by the week. Most of the obvious tells we were laughing at months ago are gone.

And before anyone makes the comparison, I would like to remind everyone that the stereotypical depiction of Luddites as small-minded anti-technology idiots is a lie. They embraced new technology, just not how it was used. Their actual complaints - that skilled workers would be displaced, that wealth and power would be concentrated in a small number of machine owners, and that overall quality of goods would decrease - have all come to pass.

In a time of unprecedented wealth disparity, general global democratic backsliding, and near universal unease at the near-unstoppable power of a small number of corporations, we really do not want to go through another cycle of wealth consolidation. This is how we get corporate feifdoms.

There is another path - if our ability to live and flourish wasn't directly tied to our individual economic output. But nobody wants to have that conversation.

Yeah, it's an arms race, and OpenAI does stand to lose. But this is a prisoner's dilemma situation. OpenAI can shut themselves down, but that doesn't fix the problem of someone creating a dangerous situation, as everyone else will keep going.

The only way to actually stop it is to get everyone to stop at once, via regulation. Otherwise, stopping by yourself is just a unilaterally bad move.

That's the charitable explanation, at least. These days I don't trust anything Musk says at face value, but I do think that AI is driving society off a cliff and we need to find a way to pump the breaks.

I mean, yeah, that sounds good. It wouldn't affect your ability to think for yourself and spread your ideas, it would just put boundaries on AI.

I've seen a lot of people completely misunderstand what chat GPT is doing and is capable of. They treat it as an oracle that reveals "hidden truths" or makes infallible decisions based on pure cold logic, both of which are completely wrong. It's just a text jumbler that jumbles text well. Sometimes that text reflects facts, sometimes it doesn't.

But if it has the capability to confidently express lies and convince the general public that those lies are true because "the smart computer said so", then maybe we should be really careful about what we let the "smart computer" say.

Personally, I don't want my kids learning that "Hitler did nothing wrong" because the public model ingested too much garbage from 4chan. People will use chatGPT as a vector for propaganda if we let them, we don't need to make it any easier for them.