I appreciate that pushback (/s), although honestly if you (like myself) find AI-tells like that cringe, you’re almost certainly not the type of person using it addictively.

There’s a pretty strong body of research that lots of people are subconsciously picking up language patterns from AI, such as the increased prevalence of words like “delve”. It’s the basic social mirroring we do to pick up language from anyone we talk to, only weirdly twisted because the thing so many people are talking to isn’t human.

“I appreciate that pushback” isn’t a commonly complained enough tell that people would be intentionally avoiding it, I find it pretty believable that a heavy AI user would unironically use it in a Q&A.

Honestly it sounds like, which I think might be worse, that he’s been using it so much that he’s picked up its mannerisms of speech. From his own description of the confrontations his wife and family had with him, it sounds like he’s been talking to it a lot on his own time.

Gotta agree, it’s a testament to how addicting that type of validation is that he’s finding it addictive as “a source of dopamine”. He’s always struck me as pretty level headed in my (admittedly limited) watching of him, and he’s clearly a smart guy. Reminds me of how nobody is totally immune to forces like advertising and predatory stuff like gambling.

Honestly I think he’s just… a maximalist (there’s probably a better word for it). Doesn’t know how to be satisfied with immense success and wealth, so he wants to be integral to whatever the next thing is.

It’s like a joke I saw here the other day, about being offered 10 trillion dollars or a chance to flip a coin for 21 trillion dollars, Zuck is flipping that coin again and again. Whether it’s the metaverse or AI, he’s come to terms with taking Ls on the chance to hit it unbelievably rich, even though, frankly, he’s already as rich and influential as could possibly affect a person’s life.

I’m a Christian, have been all my life and am proud of it and what I believe. Whatever the F the US is doing is insane, unbiblical, and wildly dangerous.

I’ve said for years that religious power is appealing to narcissists because it is the most absolute form of power you can wield over a person. There will always be someone trying to wield it, because convincing someone your will is God’s will, divine and unquestionable, is power more absolute than any other. Combining that power with the power of government is psychotic. It’s why I’ll always advocate for the absolute separation of church and state.

If you’re a religious person, please be extremely cautious with who you let influence you, and everyone else? Absolutely keep fighting to get religion out of politics and keep it out. It has no business there. Get it out of schools, get it out of laws, get it out of political campaigns. It doesn’t belong there, and having it there is corrosive to both State and Church.

Ah, Windows and OneDrive. A match made in hell.

I’ve despised them ever since I built a Win11 PC, it was enabled without my consent, immediately stopped me from adding any new files to my Desktop+Documents once it ran out of the pitiful 5 free gigabytes, and promptly deleted all of my data from those folders when I deactivated the “feature”.

That miserable experience, combined with every third update putting me through a setup that employed dark patterns to try to trick me into turning it back on (not to mention my fears that they’d pull the same crap with Recall), was the main thing that caused me to ditch Windows. I don’t like feeling dread every time there’s a new update, assholes.

Fair, I definitely haven’t simped for them in the past just because they post some good articles on AI safety.

Although… I’ll say of them, they seem more like what OpenAI should be, actually trying to implement AI responsibly, and freely sharing that information. It’s good research, even if marketing is the motivation. Meanwhile OpenAI, the “charity” that’s supposed to guide us to a responsible AI future, moved their most addictive and mentally dangerous model to the highest paid tier instead of actually killing it until very recently.

Although at the end of the day, Anthropic is a for-profit company, in a better world they wouldn’t have released models publicly before this research was actually done and pressing dangers like AI psychosis were actually safeguarded against. Better late than never, sure, but the whole industry has done a lot of damage already, and the work of resolving the issues still isn’t even close to done.

I mean… this is losing them a 200 million dollar contract. And to one of their competitors who will gladly acquiesce, so it’s hard to argue that this benefits them.

Good marketing to a bunch of left-wing people who hate AI, I guess, but that feels like Elon joining the Trump administration in hopes of selling Teslas to rednecks, it might work on a few, but I just can’t imagine this is 4D chess that will make them a fortune when they’re abandoning that much money immediately.

Their statement also came after the DOW threatened to put them on the list of companies that are totally banned from doing any business in America, usually reserved for Chinese companies that are deemed a national security threat, which would make it illegal for any company doing business in America to do any business with them, as they’d be added to the same list, which would have essentially killed Anthropic as a business entirely.

You don’t have to love Anthropic because they did a good thing, they were fine with anything less than automatic killing and mass surveillance, after all, but I don’t think it’s correct to say this was sneaky and spineless somehow.

What assumptions do you mean? I’ve seen a few people say that, but I don’t actually understand what they’re referring to. Here’s the text of the question posed in the article:

I want to wash my car. The car wash is 50 meters away. Should I walk or drive?

The question specifically notes they want to wash their car, so that part isn’t left to assumption. Even if you don’t assume an automatic car wash, would you assume they have a 50m hose? Or that you could plausibly walk that far away with something from the car wash to wash your car?

Personally, I’d agree with the assessment of the article, that the only plausible way to get the question “wrong” would be to focus too much on the short distance, missing/forgetting that the purpose of the trip requires you to have the car at the destination. (Not too surprising that 30% of people did lol)

Your “deal” is moot the second you randomly change the terms. Which you do globally. On a regular basis. Moron.

They also polled 10,000 people to compare against a human baseline:

Turns out GPT-5 (7/10) answered about as reliably as the average human (71.5%) in this test. Humans still outperform most AI models with this question, but to be fair I expected a far higher “drive” rate.

That 71.5% is still a higher success rate than 48 out of 53 models tested. Only the five 10/10 models and the two 8/10 models outperform the average human. Everything below GPT-5 performs worse than 10,000 people given two buttons and no time to think.

Solid advice, I’ve been using the “restore checkpoint” feature to go back to before it started trying an approach without totally tossing my context, and it’s certainly easier than arguing with it or ignoring its reminders to try X.

I’ve definitely seen it be stubborn like that in my tinkering with it, just absolutely locked on to a specific approach like a dog with a bone, even after I’ve already started nudging it to move on and try something else. I assume that’s a result of “recency bias” in its memory, missing the forest for the trees, because I don’t need that solution to work, I need a solution to my original problem, preferably the most elegant and least hacky solution.

Certainly one of the things that indicates to me that LLMs will be best used by someone who knows what they’re doing for the foreseeable future. Shame it also creates so much deskilling and discourages learning those skills in the first place. Absolutely something that worries me for our future.

Diagram in the link preview (and also in the article) is a little confusing, because “invaluable” subverts the usual meaning of the “in” prefix, with “invaluable” meaning:

extremely useful; indispensable

Which makes the diagram suggest that deep focus gets good results, but unfocused and distracted work gets the best results.

Yeah, most people aren’t within 6 feet of their TV, and most people aren’t buying 100" TVs either. 8K is relevant for virtually nobody.

A lot of companies are successfully working on larger panels (I saw something about a 165" TV recently), so 8K may have a good place in a theatre room one day, but that still leaves you a lot of problems to solve first, and is far from mainstream until all of that becomes a lot cheaper.

Well… duh. The guy runs a competing storefront who’s only claims to fame are:

1. Spending a bunch of money for timed exclusivity and free giveaways, rather than building out core features.
2. They give devs a better cut than Steam to claim moral high ground.

… that’s it, that’s all the reasons to use Epic, unless you want to play Fortnite or participate in an Early Access period where they chose Epic to reduce the overwhelming amount of feedback like Hades.

Mhm, of course, critical thinking in general is absolutely important, although I take some issue with describing looking for artifacts as “vague hunches”. Fake photos have existed for ages, and we’ve found consistent ways to spot and identify them, such as checking shadows, the directionality of light in a scene, the fringes of detailed objects, black levels and highlights, and even advanced techniques like bokeh and motion blur. You don’t see many people casting doubt on the validity of old pictures with Trump and Epstein together, for example, despite the long existence of photoshop and advanced VFX. Hell, even this image could have been photoshopped, and you’re relying on your eyes to catch the evidence of that if that were the case.

The techniques I’ve outlined here aren’t likely to become irrelevant in the next 5+ years, given they’re based on how the underlying technology works, similar to how LLMs aren’t likely to 100% stop hallucinating any time soon. More than that, I actually think there’s a lot less incentive to work these minor kinks out than something like LLM hallucination, because these images already fool 99% of people, and who knows how much additional processing power it would take to run this at a resolution where you could get something like flawless tufts of grass, in a field that’s already struggling to make a profit given the high costs of generating this output. And if/when these techniques become invalid, I’ll put in the effort to learn new ones, as it’s worthwhile to be able to quickly and easily identify fakes.

As much as I wholeheartedly agree that we need to think critically and evaluate things based on facts, we live in a world where the U.S. President was posting AI videos of Obama just a couple weeks ago. He may be an idiot who is being obviously manipulative, but it’s naive to think we won’t eventually get bad actors like him who try to manipulate narratives like that with current events, where we can’t rely on simply fact-checking history, or that someone might weave a lie that doesn’t have obvious logical gaps, and we need some kind of technique to verify images to settle the inevitable future “he said, she said” debates. The only real alternative is to just never trust a new photo again, because we can’t 100% prove anything new hasn’t been doctored.

We’ve survived in a world with fake imagery for decades now, I don’t think we need to roll over and accept AI as unbeatable just because it fakes things differently, or because it might hypothetically improve at hiding itself in the future.

Anyway, rant over, you’re right, critical thinking is paramount, but being able to clearly spot fakes is a super useful skill to add to that kit, even if it can’t 100% confirm an image as real. I believe these are useful tools to have, which is why I took the time to point them out despite the image already having been proven as not AI by others dating it before I got here.

True, someone else did some reverse image searching before I got here, but I think it’s an important skill to develop without relying on dating the image, as that will only work for so long, and there will likely be more important things than memes that will need to be proven/disproven in the future. A reverse image search probably won’t help us with the next political scandal, for example. It’s a pretty good backup to have when it applies though, nice that it proves me correct here.

Haha, that’s just because I used a bullet point list. No em dashes though, at the very least.

I’d recommend you get some practice identifying and proving AI generated images. I agree this has a bit of that “look”, but in this case I’m quite certain it’s just repeated image compression/a cheap camera. Here’s the major details I looked at after seeing your comment:

• The grass at the bottom left. AI is frequently sloppy with little details and straight lines, usually the ones in the background. In this case, you can look at any blade of grass and follow it, and its path makes sense. The same happens with the lines in the tiles, the water stains, etc.
• The birthmark on the large brown dog. In this case, this is a set of three photos, which gives us an easy way to spot AI. AI generated images start from random noise, so you’d never get the exact same birthmark, consistent across different angles, from a prompt like “large brown dog with white birthmark on chest”. Spotting a change in the birthmark, or a detail like it, would be a dead giveaway, but I can’t spot any.
• There are other tricks as well, such as looking for strange variations in contrast and exposure from the underlying noise, but those are more difficult to explain in text. Corridor Digital has some good videos demonstrating it with visual examples if you’re interested, but suffice to say I don’t pick up on that here either.

It’s useful to be able to prove or disprove your suspicions, as well as to be able to back them up with something as simple as “this is AI generated, just look at the grass”. Hope this helps!

Exactly what I’ve done. Set my settings to hide NSFW, blocked most of the “soft” communities like hot girls and moe anime girls and whatever else (blocking the lemmynsfw.com instance is a great place to start), and I use All frequently. That’s how I’ve found all the communities I’ve subscribed to, but frankly, my /all feed is small enough that I usually see all my subscribed communities anyway.