peterb from Tea Leaves Programming

peterb from Tea Leaves Programming

@peterb.bsky.social

effort + coffee = software. See also @peterb@mathstodon.xyz

273 followers177 following2.4K posts

peterb from Tea Leaves Programming reposted

All I am saying is that for all the "community organizing" and "providing for each other" the edgy doomers say they do, whenever I go to a community support or civic event I am often the only one under 50

What if "Make sure companies are held legally liable for the things their AIs say" was the real alignment all along?

Techmeme@techmeme.com

Character.AI files for Chapter 11 bankruptcy in the US, agrees to sell some assets to Disney for $20M, and plans to keep operating as it completes the sale (Rhea Rose Abraham/Reuters) Main Link | Techmeme Permalink

What is more likely to be conscious? a) An LLM b) A fungus-like organism c) Neither d) Unknowable Notes: (1) I'm restricting the AI answer to "LLM". Do not substitute "a sci-fi AI I read about in the Culture books" for "LLM". (2) Beating a dead horse: I said "conscious" not "thinking".

Me: "I don't think there's a realistic nightmare scenario where AI takes over (or destroys) the world. The ACTUAL risk is that incredibly dumb humans just read what it says and uncritically implement its worst suggestions." The world:

Vituperative Erb@vituperativeerb.bsky.social

I’m going to describe this as suboptimal

Martin Amis once wrote "It was...a symmetrical convenience—for Stalin—that a true description of the Soviet Union exactly resembled a demented slander of the Soviet Union." So anyway this is a subtweet about the EA / AI community in the Bay Area.

"Pffft PittGirl is overreacting again, this sounds like normal Pittsburgh politics, it couldn't possibly be as bad as she is sayiJUMPIN' JEHOSEPHAT WHAT? HAHAHAHAHAHAHAHAHA"

Virginia Montanez@ginny.bsky.social

This Gainey investigation is going to get so so messy now that the affidavit is public, but already I can’t tell you how BATSHIT BONKERS BANANAS INSANE this part is. I legitimately can’t wrap my head around this level of municipal stupidity.

Reading "text elemental" as the description of an LLM is a real Oscar Wilde "I wish I had said that" moment.

hikikomorphism@hikikomorphism.bsky.social

LLMs aren't made of neurology except in the most handwavey sense. They're made of narrative. That's why I started calling them text elementals: they're not human, they're most-likely-text generators we bent into living stories in a prose format centering the characters User and Assistant

This latest "EA's are determined to prove they understand neither altruism nor being effective" discourse about insects is an example of Kurt Vonnegut's advice from the book Mother Night: "We are what we pretend to be, so we must be careful about what we pretend to be."

the Substack article has very strong “I can fix her” energy

Ben Recht@beenwrekt.bsky.social

Why don't I trust these EA-x-risk people to self-regulate? EA prediction market company, originally funded by FTX money, creates a philanthropic wing and hires Caroline Ellison to help manage its portfolio. Can't make this stuff up!