Theo Sanderson

"Rain"-style showers (i.e. with the showerhead vertically above your head)..

1️⃣ .. are good
2️⃣ .. are bad

📊 Show results

For a fun reason https://bsky.app/profile/vgel.me/post/3mg6stg53t22n

arbitrarily

(but it's not that you are placed close to people who follow you)

who follows you

Who you follow has no effect (and also, no, I don't plan to update it particularly regularly in any case)

it's about 4% of the difference between now and preindustrial levels, so not insignificant - but yes, am uncertain about exactly how to update my model of the world

hexgrad/Kokoro-82M · Hugging Face We’re on a journey to advance and democratize artificial intelligence through open source and open science.

I use your computer to do the compute :) : huggingface.co/hexgrad/Koko... / voice-generator.pages.dev

Explore the bluesky universe with 🔊. (Not suitable for mobile connections)

bluesky-map.theo.io?live=1&reada...

Bluesky map, now with posts read aloud 🔈 - pretty trippy

bluesky-map.theo.io?live=1&reada...

Chrome recommended. Uses significant bandwidth to download TTS voices.

hadn't thought/known much about CO2 seasonality until now

Nice - that should be my citation furthest away from the centroid in semantic embedding space :)

my fave being

made some tweaks to @iambic.bot recently so it never sends the same text twice; it's been on good form since

🤔

I'm not sure to what extent I want to make it a "supported" big thing. It's just a single page HTML so easy to play with. Definitely no issues with you forking it, which has already happened

Added live likes from the firehose to the map too. Rather pretty.
bluesky-map.theo.io?live=1&likes=1

You would prefer that the pace of AI improvement was:

1️⃣ Faster
2️⃣ At the current rate
3️⃣ Slower

📊 Show results

Perhaps to illustrate a wider trend, idk. Anyway, I wouldn't be seeking an apology if I'd been rude to someone and not given one myself

In case you're unaware, your original reply here bsky.app/profile/mike... comes across as sarcastic and rude

"The defining goal of the academic endeavor is to establish a single perfect value that completely summarizes entire careers’ achievements. Such an index would allow the academic community to formally and uncontroversially determine the best researchers by simply comparing their respective scores."

The core-pivotal index: a geometry-first approach to academic impact The defining goal of the academic endeavor is to establish a single perfect value that completely summarizes entire careers’ achievements. Such an index wo

@russcd.bsky.social ushers in a new era of academic metrics with the core-pivotal author index, which assigns all credit to authors that occupy the exact center of the author list, while non-central peripheral authors receive no credit whatsoever.

🔗 doi.org/10.1093/molbev/msag026

#evobio #molbio

ah no, yeah I remember, it really (from the examples) in all cases needs you to have a known relationship between X and Y coordinates in the camera and points in 3D space

yeah, I looked at this a while back, discarded it because the examples in the docs seemed less directly relevant - but actually some are, so I should

sorta

Thanks! Yes, for now I just wanted to satisfy my curiosity as to what claude could directly do given a robot arm - but I'll definitely keep these in mind! (tho the arm's calibration is a bit off due to a battery running out, so it's harder to provide perfect kinematics than it should be)

I think it's kinda hard to know if your MCP server is working, and to figure out what's going wrong with it, if you can't do it yourself (even if very slowly)

yes, no I know this isn't the way to go, but I guess on some level it's a precondition for AGI-in-one-model

hah yes, this was my treat to myself a few years ago after a year out of academia on a US tech salary while still living frugally

Confusion sometimes similarly related to "thinking" things had succeeded when they hadn't but maybe more at a sub-task level