Just found this on sneerclub, this seems like a reasonable intro to the subject. Thanks!

Question: anyone has a link to a good primer on the rationalists and especially how that whole environment influences/interacts with the large AI companies like Anthropic?

I need something to bring some ppl I know up to speed on this bc when they ask me “wait, what do you mean when you say that Silicone Valley turned into an AI cult with orgies” I just stare into space overwhelmed by the magnitude of the task ahead of me.

Edit: thanks everyone for responses, I’ll get my friend started with Sex, AI, and the Apocalypse and then point him at rational wiki for any further questions.

Yeah it’s really bad, I’ve been into crocheting and knitting (more of a knitter now) for about 2 years and if it’s not fake AI patterns or yarn companies dropping the ball, it’s pattern designers using AI to generate background pictures and stuff. I’ve been really disappointed with some people who really should have known better. You are in a creative, artistic hobby space! Why are you using a slop machine and devaluing work of other artists! Ugh.

At least being on ravelry and checking out other people’s project attempts/pattern comments has so far saved me from falling for AI bs.

As others are trying to convey, it’s not about the bot but about allowing yourself to indulge in this behavior. The bot could be substituted for a rock or a rubber duck, but you are training yourself to subconsciously think that this behavior is acceptable. It increases the chance that you’ll one day start treating actual people like that.

Imagine this scenario: you come home after a long day of work. At work, you use AI a lot and curse and swear at it when it makes mistakes (which is often). At home, you ask your spouse/child/roommate/parent to do something for you, which they attempt but proceed to mess up. They apologize, and instead of handling it calmly you instinctively react like you’ve trained yourself to react - by screaming at them.

Is that what you want to be like?

…I’m so sorry, and it’s good (???) to know my company still has a few steps on the AI delusion ladder to climb. I guess with the way it’s been going I need to mentally prepare for the same thing you’re going through to come in a few months.

Vita Nostra by Marina and Sergey Dyachenko. Really weird, abstract and interesting. Anyone labeling it as Dark Academia or a magic school book is both completely correct and dead wrong at the same time.

Piranesi by Susanna Clarke, This Is How You Lose the Time War by Max Gladstone and Amal El-Mohtar both have very interesting concepts and great executions.

When it comes to interesting concepts in fantasy/sci-fi, Adrian Tchaikovsky has some bangers. Elder Race, Children of Time, or Alien Clay are some random example recs.

China Mieville’s New Crobuzon trilogy is really out there too. 2nd book is the best and can be read independently but does spoil some things from the first.

Spec-driven development is when you take these AI generated design docs and make it all you as a dev interact with. It’s waterfall but make it stupid and AI.

You give AI some guidelines and a feature description, then you twiddle your thumbs while it generates giant specs and occasionally asks you for some inputs, then you attempt to read the thing, give up, and give it to some other agent. Rinse and repeat until you burn a GDP of a small country in tokens and the agents agree that it’s going to work, and then you finally pull the lever and hopefully a feature matching all your specs appears. Or it doesn’t! And if it didn’t, it’s surely your fault because you didn’t give AI enough context.

I’m tired, boss.

I feel the need to scream into the void so here goes: my corpo job has certain execs that are very deep into the AI bs, and they’ve done a full tour of all the stages and iterations of AI hype, following all the new trendy tools because otherwise we’d surely be left behind 🫠

And the current somewhat new hotness is spec-driven development. I’ve tried it over the last few days since it looks like it’ll be mandatory for larger work.

And boy. Boy oh boy.

There are no words in the English language to describe how much I hated it. There may be some ancient curses in R’lyiehian that would do it.

The Copilot mandates were bad enough, but honestly if this is an attempt to make me see the light and admit that in itself Copilot isn’t that bad, then it’s kinda working.

It’s just… idk, I got into software dev because I like logic puzzles (and wanted the money, let’s be real). I enjoy making a small thing and then slowly adding stuff to it. I do not enjoy sitting idly while the slopmachine writes a giant spec that I think I’m supposed to read and correct but can’t because it’s unreadable. It’s like a textual white noise. It’s pretending to describe a piece of software I want to make while actually describing all the steps of a ritual to summon an Elder God of anxious boredom and machine madness.

I think I’ll be defining all of my work as small from now on.

I guess the silver lining here is that “his team” might accidentally discover some actual way to help sufferers of this disease? Unlikely, but stranger things have happened.

Wait, three? What’s the previous 2? I think I’ve only heard about this case (…and a lot about Mythos so that proves your point nicely).

Ah yes, I love the smell of burning bridges in the evening. Fuck. And I was getting excited about Divinity! Well, guess that means more money to spend on other things.

Well that’s going immediately onto the “things that looked promising but what the fuck, so actually don’t read ever” list.

deleted by creator

Re: minecraft - kids/people who aren’t very good at technology can’t or are unwilling to learn how to host their own servers, so that’s your potentially paying audience. Or people who want to play with a ton of other people, not just their family/friends. And you can do some interesting things with custom scripts and so on on a server, I remember briefly playing on a server which had its own custom in-game currency (earned by selling certain materials) and you could buy potions, equipment and various random perks for it (and of course there are ways to connect that to real money, although you might get banned for it).

https://lemmy.ml/post/31490862 pretty interesting article linked in this post, tl;dr researchers tried to get AI agents to run a simulated vending machine (which, let’s be clear, is a solved problem and can be done with a normal algorithm better and cheaper) and it didn’t go that great. Even if some of the test runs actually managed to earn money, they mostly devolved into the AI becoming convinced that the system doesn’t work and desperately trying to email someone about it (even FBI, one memorable time). I think it illustrates quite well just how badly things would go if we left anything to AI agents. What are the odds anyone involved with pushing autoplag into everything actually reads this though…

Yeah, that video was rough. I wasn’t actually expecting him to be super lefty or anything, but the “both sides bad and should talk more instead of screaming” take was still a blow.

So how many ChatGPT 4s have they precariously stacked up on top of each other this time?