Who's Who of AI

Andrea Lathrop

58 trust @cabernet.bsky.social · 1,563 followers
Cognitive Science... tech stuff... dev psych... AI... former Aslin Lab (Master's) and Eppler Lab (with Jackie Gibson emeritus) so I'm a weird mix of Gibsonian affordances, visual statistical learning and anticipatory eye movements. Sometimes academic.
What they're sharing

Articles & links

↻ Andrea Lathrop reposted
@carlquintanilla.bsky.social

".. The app .. frequently miscounted and mislabeled items, such as confusing similar milk types or ​missing them altogether .." @reuters.com $SBUX www.reuters.com/business/sta...

reuters.com View on Bluesky →

OMG, it was Lutnick! LMAO. www.reuters.com/technology/a...

reuters.com
View on Bluesky · ♥ 0 ↻ 0 ↩ 1 · 2 from the directory shared this · 2d ago
↻ Andrea Lathrop reposted
tweety fish @sifu.tweety.fish

well, lo, people have built a benchmark using a task called multi-armed bandit, where success depends on iteratively grasping the relative odds of payouts from different choices, and LLMs are (generally) shit at it arxiv.org/pdf/2403.15371 /

arxiv.org View on Bluesky →
Their own posts

Recent commentary

So, obviously, a bunch of people are going to keep churning out LLM-harnessed math results, ad nauseam, because that's what the 'now' thing to do is, but I'm sort of wondering what the point of math is. Isn't it this grand puzzle we (human civilization) set for ourselves for a challenge?

View on Bluesky · ♥ 7 ↻ 2 ↩ 4 · 21d ago

I used to think it was weird that, on a ship full of engineers, the holodeck kept malfunctioning, and almost killing everyone, but now that I see how we are speedrunning agentic AI, it seems more plausible.

View on Bluesky · ♥ 12 ↻ 2 ↩ 0 · 29d ago

My weird, idle contemplation for the day... now that AI labs are hiring philosophers at high six figure incomes, and where else is a philosopher going to get a high six figure income...

View on Bluesky · ♥ 6 ↻ 1 ↩ 2 · 13d ago

Google put "AI Mode" on the far left, where "All" used to be, and I keep accidentally clicking it.

View on Bluesky · ♥ 5 ↻ 1 ↩ 2 · 18d ago

Sorry, wtf is Corgi? It looks like a (somehow AI?) finance/insurance startup, but people joining the co. on Twitter are announcing it like it's some sort of top secret Manhattan Project to "save the country" from [something vague, caused by] the AI transition [that is also their life's work].

View on Bluesky · ♥ 7 ↻ 0 ↩ 1 · 17d ago

The most artful thing about the 'critique this "AI" [real, actually] Monet" stunt on Twitter is that OP then minted the [cropped screenshot of a Monet poster] on the blockchain as 'Inferior Image'.

View on Bluesky · ♥ 7 ↻ 0 ↩ 1 · 35d ago

Anthropic: "Recursive Self-Improvement is not yet reachable, but might become reachable, if..." Japan: "Announcing our Recursive Self-Improvement lab for Recursive Self-Improving Self-Improvement."

View on Bluesky · ♥ 6 ↻ 0 ↩ 1 · 13d ago

Ok, I am reading 'the Anthropic blog post.' My tl;dr so far is: a. AI coding agents are good at automated beam search in a well-defined space, so anything that can be done by automated beam search in a well-defined space is going to be

View on Bluesky · ♥ 5 ↻ 0 ↩ 1 · 13d ago

Clearly, the Temperature parameter is the seat of LLM consciousness.

View on Bluesky · ♥ 6 ↻ 0 ↩ 0 · 35d ago

Ok, I have now (tried to read the OpenAI proof... that one is over my head, mostly because I can't decipher mathematician citations of other mathematicians on the fly, I don't have the world's conjectural & derivational history stored in my brain.) I also read the mathematicians response paper, and

View on Bluesky · ♥ 3 ↻ 0 ↩ 1 · 24d ago
Their network

In Andrea Lathrop's orbit

Center = Andrea Lathrop. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.