sjkoelle

thats reasonable, but its a question of degree. for me gambling and predictions markets dont meet that bar

or even if they are pro-gambling jeez would you never be able to work with someone you disagreed with politically

trust the gooses

i just wish someone would explain why i prefer cline to claude code so much

Interesting that CRISPR was the top answer in 2015 too

lets be real we are gonna lose our jobs to these kids

Great name - Captain Hook recently hit public domain - would be a sick logo! (Disney stuff like red jacket still copywrited)

this happened to me. they refunded me after i contacted support!

Oceania has always been context engineering. Its been interesting to see this prioritized in the zeitgeist over the last 6 months from the "long context" zeitgeist.

article explains that it is not sizable

Can we get a shout out to the anti code review folks?

yeah im curious if people will end up liking it. sucks from my perspective.

is one function per file the righteous path here?

it depends how long of a leash you give it

amara must be this dataset https://en.wikipedia.org/wiki/Amara_(organization)

efficiency is not a given. also this is an eval set - they acknowledge the challenge themselves.

imho this is v cool

Would this generate the same completion for 'the cat sat on the' as 'on the cat sat the'?

Marvelous! What gain beyond zero-shot would motivate a humble citizen to implement this instrument? How was the superiority assessed?

The title alone is a teensie bit hilarious

is this like https://icps.u-strasbg.fr/people/bastoul/public_html/researc...?

the interesting advance in the anthropic/mats research program is the application of dictionary learning to the "superpositioned" latent representations of transformers to find more "interpretable" features. however, "interpretability" is generally scored by the explainer/interpreter paradigm which is a bit ad hoc, and true automated circuit discovery (rather than simple concept representation) is still a bit off afaik.

always amazing how ahead of the big data curve the graphics people have been

striking how closely ij good's conception of machine superintelligence is matched today http://incompleteideas.net/papers/Good65ultraintelligent.pdf

Anyone who has written a classifier using an LLM knows how important this is. Great contribution!

How much of this is due to DNNs (e.g. VAEs but also others) forcing embeddings to distribute in a Gaussianish manner? Is the data intrinsically missing geometry or could a more subtle learning algorithm give a cleaner manifold and therefore more efficiently indexable structure?

how does this compare with PCA?

all time naming miss not going with ragtime

not sure on details but i think neural networks have pretty great properties for finding center manifolds

as long as the dynamics are relatively low dimensional, it should be possible to estimate the differential equation, even if it is observed in a high dimensional space

can you say more about the application there? what does the data look like and why is symbolic regression important? asking for a friend ;).