artist, engineer, researcher. neuroscience, machine psyche. science & art, futuristic & ancient.
Austin, TX
Joined November 2017
- Tweets5.8K
- Following392
- Followers101K
- Likes11.7K
Pinned Tweet
if you are a new follower, i've been doing these gestural experiments for a while now. eventually i wanna build tools that manipulate symbolic thinking in an embodied and intuitive way.
e.g. using hands to explore latent spaces like this:
xitter.cf/poetengineer__/status/…
or make music thru interfaces like this:
xitter.cf/poetengineer__/status/…
i'm currently full-time working and self funding all these projects. if you happen to be feeling generous at the moment, please consider becoming a paid subscriber at my substack. this could help me just go a bit further😌 💜🫶
This quoted post is unavailable.
for better or worse, the rl problems i find most interesting are the ones whose rewards are hard to verify: what's truly creative, interesting, beautiful?
criteria for these are even hard to articulate among humans. i think one reason is they are properties of not the artifact, but of the relation b/w the artifact and the observer, at a given moment.
that also explains why chasing originality sometimes feels paradoxical: everyone sees the "most creative" against the same reference, so optimizing for it converges into a trend, hence totally defeating the purpose of being original.
if you can't specify the target from outside, try looking inside. i feel intrinsic motivation research might be the right lens here, since it's the part of rl where reward is defined relative to the internal state of the observer, instead of progress toward an external objective.
random thought: the desire to learn (compress information and generate new knowledge) is the biggest difference between living and non-living things. when the universe generate environments where learnable patterns exist and there's a reward for learning them, life begins.
every mind is a microcosm that holds galaxies of ideas.
continuing my personal website prototype: 6 galaxies, 20 themes, every star a work.
looking at @elluba 's ai art newsletter from 2020, i feel like our definition of "ai art" has narrowed so much since then:
back then ai art means working with ai as a system: ai could be a classifier, a de/encoder, a perceptual instrument, .etc;
now it seems almost synonymous with generative image/video.
inspired by medieval illuminated manuscripts as a mnemonic device, i tried to illuminate turing’s ‘computing machinery and intelligence’: a giant I, one scene per section
neutrino observatories must be one of the most eerily beautiful structures humans have ever built
ALT Super-Kamiokande
ALT IceCube Neutrino Observatory
ALT Sudbury Neutrino Observatory
ALT The Deep Underground Neutrino Experiment (DUNE)
i feel a personal website should be a vessel for the multitudes we contain, so i made a gemstone to hold mine.
the original 3 planes now grew to 20 facets/themes. every particle is a work. like a prism, hover a facet and a different side of me comes out.
lessing made a similar point when he said that if god offered him static, absolute truth in one hand and the active, error-prone pursuit of truth in the other, he'd take the pursuit.
i think there's a lot of reasons why we like the pursuit of sth a lot more than actually that thing itself.
jev reminds us how good speed can feel.
i trained a colorgpt a while back, a 2.7m-parameter nanogpt-style transformer conditioned on frozen minilm embeddings and trained on the 2010 xkcd color survey. it runs in the browser and outputs only distributions over rgb values given a prompt. training took ~3 minutes on an rtx 5070, and the results aren't bad.
it made me hopeful for a future of models in all shapes and sizes: fast, cheap, and easy to customize.
