espe

and the other way round - have it built in for self-fuzzing and healing the infra.

yes that sounds rational. central bankers are humans with a target on their back though.

the "perfectly rational markets" expect the central banks to lower interest rates. which actually assumes that something economically worse happens in the meantime because as it is now, the fed is _not_ lowering.

it's a play on n-wave feminism and how white lady feminism was/is different than that of the "subalterns".

yes, it's the nvms one.

they will sell. the big guys are gobbling up _anything_ they can get their hands on.

My fear is that eventually you'll start seeing government services where identity and auth are handed off to private companies like Google and Apple. Imagine having your real identity tied to an attestation by one of those companies.

a general taste for dysfunctional public-private partnerships and the fact that auth is a seriously hard problem at scale, make this scenario a few percentage points more likely than anyone should feel comfortable with.

yes i'd second deepseek-coder and check out wingman. it's a relatively new extension, snappy and with good ux.

yes. from the perspective of anthropology, humans using tools are cyborgs. and language and texts are ai. or "alien intelligence" as burroughs (and a few others) would have put it.

you both describe valid approaches, similar to a breadth- or depth-first search. one will get into deep waters with "some great big system", regardless of the specifics.

this looks really cool. might be perfect for an old odroid. hope that doesn't stray too far from the happy paths.

true. i call them "prompt stimuli". prompts can give structure and context and whatnot but a dash of original domain gives the edge.

source: a law prof specialized in the matter. but yeah, ymmv.

just remember that feedback loop implicates us, our language, psyche, culture. i guess it will be a challenge _not_ to unwittingly converge with LLMs.

i think it's safe to assume most "copyrights" on model weights will be annulled, on either side of the pond.

thanks for the clarification! i'd still be in for a fossil t-shirt :)

containing the downsides of Silicon Valley’s latest export

through which political lense do i have to look in order to put it like this? there is decades of european (continuing the bloc thinking) research and braindrain (and wasted opportunity) going into this.

yes the baseband blobs are still underappreciated.

interesting question. anyone know of a paper that looks at power law distributions in llm vs. human text?

maybe check out https://flet.dev

well, in a way language itself is a semantic timebomb.

python3 -m llama_cpp.server --model /path/to/model.gguf

nice. 1995 ibm aptiva keyboard here. it’s still the same as on day 1.

not sure it helps but - soulseek is pretty alive

that should also apply to vscode extensions and most other plugin systems. and this seems to get worse? one area where apple's strategy of copying the functionality of smaller extensions/apps makes sense security-wise..

yes. at first glance it looks like a windows app but it's actually very portable. it has some parameters for gpu offloading and extended context size that just work. it exposes an api endpoint. i use it on a workstation to serve larger llms locally and like the performance and ease of use.

this. also, they have a payment feature that opens them up further on grounds of regulatory scrutiny.

those paratextual phenomena probably are important for the model's representations.. not to get rid of and not easily compressable either. have a look at predicitive features for authorship attribution in stylometry for example. whitespace and punctuation are always decisive.

"retina" describes the orientation of the vector space? first time i hear it in this context.

i think they don't. given enough data those related forms will have similar colocation profiles (aka fingerprints). not a model of semantics.. at scale it might become a proxy though.