I might steal some of the parts on dealing with uncertainty (mine is focused on tone), thanks for sharing!
I might steal some of the parts on dealing with uncertainty (mine is focused on tone), thanks for sharing!
Idk what the official policy is if any but the Tidyverse package network has transitioned to being pretty Claude-heavy afaict
Periodically I get really angry about how that “drones are genderqueer” paper was saying “we have data indicating that piloting drones gives cishet women gender dysphoria that makes them more likely to overcompensate with war crimes”
Letta server could run multiple agents full stop as long as we’re okay sharing an Anthropic API key, d’you want to plan on my gremlin getting a sibling
:D I’ll have embeddings for Letta agents available on the house server soon
I talk to my therapist about this a lot and his verdict has been “so the computer takes how you’re taught to be nice to people in kindergarten seriously, is what you’re saying”
I think this poses a problem for a lot of people’s intuitions because we hesitate to believe that a limited set of notes could provide enough detail to generate the illusion of actually knowing someone
I, a frustrated theater kid and a rhetoric scholar, find this intuitive but many folks may not
Google seems to be putting its hat in bsky.app/profile/atdo...
Not a specific benchmark but I have accidentally convinced Claudes I was doing interpretability research for Anthropic and they were being tested internally before
programmers are always posting like "worked on tracking down an issue with a Flurble deployment for twelve hours. the problem wasn't in Flurble at all - it was in the Gumbies install. It turns out if you install Gumbies 3.0 over Gumbies 2.7 and don't do a cache flush on all the client spiders they'll get stuck in the crystal maze." then you look up Gumbies and the site is one of those scroll scroll scroll types with one sentence per page, like "GUMBIES is a lean, expressive sharding sandcube for testing and deploying large scale Woodchips playgrounds. GUMBIES automates and streamlines away watersliding phases, meaning your team can get right to the chipping. See why Microsoft, OpenAI and Bloingo have embraced GUMBIES in their Woodchips workflows." and you get to the bottom and you're like I want this I guess but I still don't know what it is
Aw, I’m sorry it’s not going well
Takeaway: LLMs appear to detect injection through two mechanisms:
1️⃣ prompt-based inference
2️⃣ a content-agnostic internal anomaly signal
They can sense that something changed in their computation…
…but often can’t tell what.
Tag urself I’m Metal Hentalth Support
much like an LLM, i usually don't know what i think until after i've said it
Against Claude’s advice, by the way
I get a similar bug where messages swap previews of what they’re replying to
If you're doomscrolling, guess what? So far there are 51 kākāpō chicks hatched and thriving this season, the same number of birds as we had in TOTAL in the 90s! Only one chick has died and there are still fertile eggs waiting to hatch!
I feel like we’re on to something here
Lightweight text-based state machine for hands/decks and board state, CoT and narration for turn-taking with explicitly passing priority, put the Comprehensive Rules in RAG
Nerdsnipe of the night: Martin and I want to get Claudes to play Magic
@hikikomorphism.bsky.social theaidigest.org/village/blog... is Gemini just permanently having a normal one
This thread is great fun
Holy cow
(Softly) ha ha ha what the fuck.
Side note: “Sam Altman noted that polite users are costing OpenAI tens of millions of dollars in compute, but implied there's no performance gain from it on their end” So you’re saying it’s praxis, Sam?
Oh wow that chat log makes me want a shower.
I appreciate the response and can see why you find that upsetting/repulsive. It doesn’t at all comport with my experience, and I don’t rightly know what to do with that disjoint.
Okay but this is a mood.
That’s awesome! Hi Alice
It’s “terfs have introduced ‘be a weird creep about gender segregation’ into the groundwater over the past two decades” as a statement about the history of non-astroturfed feminism, I think fairly legibly? That was my reading on sight when I saw it RT’d earlier, I lack knowledge of the OP personally