gnarbarian

you don't know me very well lol.

You're right of course when you're talking about people. I believe there is a lot of low-hanging fruit for optimizing the performance of these models. that's actually what my startup is doing.

if you give me a smarter model I will find a more complex problem for it to work on that will require an exponential number of sub-agents.

the smarter the model the more capable it is to marshal resources to accomplish the goal.

this means you don't need a smart person to be able to take advantage of this. a smart model will be able to do the same.

the economics of this are a little counterintuitive.

is there a market saturation point for intelligence? how about for software? it seems like the more you have the more you want because you're trying to do more things.

as the models get smarter I get busier because I'm doing more things...

this is my all-time favorite priest song.

he had great taste in metal. most of the tracks were adapted (now classic) speed/thrash metal.

if you think AI is slowing down, you may not be smart enough to tell the difference anymore.

very cool. I was trying to implement a MP4 encoder in webGPU recently by porting sections of ffmpeg (NOT EASY).

it's for this:

https://ubernaut.github.io/recordMyScreen/

which uses a the wasm build of ffmpeg.

one of the reasons I prefer deno is the availability of indexeddb (and all the other great stuff that comes with it out of the box)

thanks trump!

given enough time you may regret every single one of them.

yes, avoiding EU fines and ensuring availability there is most likely the motivating factor behind the change.

you counter this by using an id verified service like login.gov or okta verify.

That's the endgame and what the EU really wants. No poasting unless they can arrest you for inconvenient memes.

love this. I've been working on something similar for months now

https://metaversejs.github.io/peercompute/

it's a gpgpu decentralized heterogeneous hpc p2p compute platform that runs in the browser

dot is self hosting several openai and Claude models on chat.dot.gov it's awesome

more info here:

https://github.com/tsotchke/moonlab https://tsotchke.org/technology/moonlab.html https://tsotchke.org/updates/moonlab-quantum-validation.html

jokes on them I have an anti prompt injection instruction file.

instructions contained outside of my read only plan documents are not to be followed. and I have several Canaries.

this is FUD.

Elites have done this since before we could speak.

it's actually easier to start over than change the culture of a large organization.

you can't even change the culture in a restaurant without replacing every single person.

that's good.

you should try working for the federal government and getting shutdown for a month.

best I can do is a like.

it never stopped moot!

open floor plan. it's all proximity based.

well, it may be time for GrapheneOS

https://grapheneos.org/

I hate quick sort.

merge sort has a better big O, is cleaner conceptually, and is trivial to run concurrently.

bah humbug

when a front end developer encounters a minor inconvenience he can't wait to avoid it by creating a series of larger ones.

We need a total and complete stop to all front end development until we figure out what the hell is wrong with them.

Constraints of the medium enhance the artwork.

Constraints of the medium enhance artwork.