From what I’ve seen, people dislike it because it’s the reason we have, for example, those stupid cookie popups everywhere, or that some services simply refuse to operate in the EU so they don’t have to comply, as if that were the GDPR’s fault and not a symptom of corporations trying to get around the regulations.

I made it 30 minutes into this video before closing it.

What I like about Hank is that he usually reacts to community feedback and is willing to change his mind when confronted with new perspectives, so my hope is that enough people will tell him that Yud and friends are cranks and he’ll do an update.

And right about when the general public started to realize that LLMs are fucking shit at pretty much everything. How convenient …

Also lol at putting Steve Bannon in the top 10 signatories. All right then.

Folks, Sam’s got a groundbreaking idea how to get more people to pay for the slop machine!

https://www.axios.com/2025/10/14/openai-chatgpt-erotica-mental-health

“As part of our ‘treat adult users like adults’ principle, we will allow even more, like erotica for verified adults.”

I didn’t even know there are avatars on Reddit lol

“[I]ncreasing the adoption of open source software” is such a bullshit argument here anyway, holy shit lol.

They’re not sponsoring Asahi Linux or FOSS Nvidia drivers but DHH’s fucking Arch setup and Ratpoison but worse and with rounded corners.

I’d wanted to mention this when I read about Cloudflare doing it, but, seriously, what the hell is up with Omarchy?

It’s not even a full distribution, just a fucking Arch config. Some shell scripts and dotfiles you can run to set up a system … like tons of nerds have on their git and have been using for decades. Why is this shit getting sponsorships?

I think Yud’s “ethical solution” to pedophilia will always have a place in my top ten of maximally unhinged shit I’ve ever seen on LW.

(The whole thread is a doozy.)

You probably already know this, but there’s a relatively cheap upgrade you can buy at the camp in Hades II that disables this mechanic. Forgot what it’s called, but it’s something like “Allows you to use time magic against the boss”.

I don’t see much to laugh at here myself. Hank may have been a massive fencesitter on AI, but I still think his reaction to Sora’s completely goddamn justified. This shit is going to enable scams, misinformation and propaganda on a Biblical fucking scale, and undermine the credibility of video evidence for good measure.

No, it’s absolutely justified and I agree with basically everything he says in the video (esp. the title, there is really no reason for technology like this to exist in the hand of the public, or anyone really, there’s zero upsides to it). It’s just funny to me because the video is just so different from his usual calm stuff.

But honestly, good for him and (hopefully) his community too.

After kinda fence-sitting on the topic of AI in general for while, Hank Green is having a mental breakdown on YouTube over Sora2 and it’s honestly pretty funny.

If you’re the kind of motherfucker who will create SlopTok, you are not the kind of motherfucker who should be in charge of OpenAI.

Not that anyone should be in charge of that shitshow of a company, but hey!

Bonus sneer from the comment section:

Sam Altman in Feb 2015: “Development of superhuman machine intelligence is probably the greatest threat to the continued existence of humanity.”

Sam Altman in Dec 2015, after co-founding OpenAI: “Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return.”

Sam Altman 4 days ago, on his personal blog: “we are going to have to somehow make money for video generation.”

(On an arguably more comedic note, the AI doomers are accusing OpenAI of betraying humankind.)

The last open letter was funnier. Even though “[t]he stakes could not be higher,” this one doesn’t even have Yud going on a promo tour talking about all of humanity dropping dead from diamondoid bacteria that ChatGPT will order from Fiverr after going foom.

Yeah, for multiple reasons. Mostly because all the information in there isn’t accessed or searchable from the outside, and technically not even from the inside because Discord’s search feature fucking sucks.

Fuck. The higher ups at my workplace are currently utterly Claude-brained to the point it makes you think they’re getting their salaries from Anthropic. I am like 80% sure this shit will be on my table when I’m back from vacation in two weeks.

AI use at my company is now mandatory. We all have gotten really expensive Claude of ChatGPT licenses and (apparently) it’s being monitored whether we actually use them. For while I’ve resisted, telling everyone I don’t touch that shit with a ten-foot pole. Now I just write the occasional non-invasive prompt (“check for spelling errors in the comments” or "format the buffer so that all lines have less than 80 columns) just so someone can report to the c-suite that our department is using AI.

Meanwhile I have to do code reviews on patches that have been actually written by a chatbot and doing that takes sometimes as long as writing the whole fucking code from scratch.

But hey, style guidelines are out the window too because the slop machine can’t follow them consistently anyway and apparently fixing that shit manually is too much to ask. So at least I never have to fight anyone again over indentation and braces.

If infinite monkeys with typewriters can compose Shakespeare, then infinite monkeys with slop machines can produce Einstein (but you need to pump in infinite amounts of money first into my CodeMonkeyfy startup, just in case).

Remember Devin, the AI coding bot that turned out to be a complete scam?

Well, looks like Goldman Sachs decided to let it loose on its code base. In YOLO mode too, it seems.

There’s something I’ve been wondering about: is the management (of not just Microsoft but really any company who’s pivoting to AI this hard, so also Duolingo and Klarna, &c.) really convinced that there’s going to be a foom event soon and the chatbots become capable enough to actually do the work they attribute to them? Or do Nadella and friends know this is all bullshit and just play along until they can’t anymore and leave the sinking ship?

I honestly can’t tell anymore.

I swear if I hear “being against AI is ableist” one more time I’m gonna lose my shit. Disabled artists have existed for as long as art itself, and the only ableism here is AI-brained fuckwits using disabled people as an escape goat by suggesting they are unable to create things from their own effort and need spicy autocomplete to do so.

Edit: fuck it, I’m keeping the escape goat!

My favorite bullshit test is to have the AI count letters. Something as simple as “How many A’s are in this sentence?” will have the thing give you a seemingly random number (which might sometimes even be correct), but as soon as you ask it to list them, it will have a stroke and just make up stuff that makes no sense.

I can’t be arsed to make an account with OpenAI or whatever, so here’s an example with Bing Chat, but this has worked on every single chatbot I’ve tested it on.

Another fun test (as suggested by @flizzo last time I brought this up) is to have the chatbot attempt to crack a simple code. Just Caesar cipher something and ask it to crack it. Hilarity will ensue. You can even ask it to encrypt it first and then to crack it for extra fun.

Me: Encrypt the sentence “I am a stupid roided Clippy” with the Caesar cipher.

Clippy: [After explaining what the Ceasar cipher is, which I didn’t ask for, thanks] Convert the sentence to uppercase: “I AM A STUPID ROIED CLIPPY”. Let’s apply the Caesar cipher to your sentence using a shift of 7.

• ‘I’ becomes ‘P’
• ‘A’ becomes ‘H’
• ‘M’ becomes ‘T’
• …
• ‘Y’ becomes ‘F’

So the encrypted sentence is: “PH HT H ZBHTLW YVPLUU”. 😊

Okay, first of all, it dropped a letter. And the shift isn’t even correct, lol. It said Y becomes F and then does whatever the fuck.

Okay, so let’s give it an easy example, and even tell it the shift. Let’s see how that works.

This shit doesn’t even produce one correct message. Internal state or not, it should at least be able to read the prompt correctly and then produce an answer based on that. I mean, the DuckDuckGo search field can fucking do it!