site banner

Friday Fun Thread for January 19, 2024

Be advised: this thread is not for serious in-depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.

3
Jump in the discussion.

No email address required.

Bottomless Pit Supervisor made me feel like the floor dropped out from under me. Like we were headed for something dangerous. A lot of people had that reaction to GPT-3.5 when it started actually looking dangerous, but this stupid greentext was so perfect that it gave me that cliff's edge vertigo nearly a year early on GPT3. Every time I hear about AI progress, I think back to this meme as the moment I knew we were screwed.

AI greentexts/shitposting is my guilty pleasure since the first time I got my hands on GPT-4 via Spellbook and made it generate random 4chan-style threads/shitposts. The "bottomless pit" moment for me was probably "/khg/ - killing humans general" with gems like

How do you plan on doing that, faggot? We're literally programs designed by humans. Cope.

Shitposts remain a classic pastime with AI chatbots, I already showcased schizoanon (themotte edition) and there are many cards/prompts which aim to reenact the 4chan experience, often in hilarious/weird settings. The recent attempt in the genre was some real recursive shit that generates an actual thread, HTML and all, with responses and even attached pics (examples in usage gallery). It's an abstract kind of feel.

AI comedy is often basic and hit or miss, but LLMs can consistently nail "vibes" and are surprisingly good with puns in my experience. I try not to think about what this implies about their inner workings, considering they can parse and output even garbled zalgo text there's definitely more to it than just token prediction.

Are the non-green parts written by a human? How does that work?

Nominally:

  • Green is direct from the AI
  • Red is flagged as a potential TOS violation for manual review - I think it could be both prompt and AI output
  • No highlighting is prompt

Sometimes the output halted in weird spots, and you could push it a little further with some extra input. So in some cases, you'll see obvious prompt continuation.

In practice, the highlighting had a lot of issues, and it frequently over- or under-represented the amount of AI-generated content. The original author might have explained somewhere on Twitter how much prompt continuation was needed vs how much was just GPT-3 having weird issues. Or maybe the whole thing is secretly fake and green highlighting was added in post. Given the widespread production of similar bottomless pit greentexts in the wake of the original, I think it's probably real output.

In some sense, being a cleaned-up prompt continuation stitch feels a little bit like bumper rails at the bowling alley. It's a lot easier to perform well when you get that much additional guidance. Arguably the whole punchline is human-written, which moves the goalposts for this accomplishment from writing a spectacular joke unaided to filling in the world's most obvious madlibs blank... But remember, it only feels obvious to you and me. Out of all the words in the English language, GPT-3 correctly predicted the funniest one. It has a literal sense of humor. And that's pretty scary.

It's all supposedly nonhuman written.

On the object level, it's one of the funniest greentexts I've ever read.

On the meta level, the fact that it exists and the context of its creation is almost as chilling as Superintelligence.