txdv

LLMs are black boxy, its not like I can inspect the used weights of a prompt and judge the source of those weights.

I'm currently writing a compiler (forever garage project) the speed at which Claude is able to debug with gdb what was wrongly generated is insane. I had not much prior expertise so maybe a professional in the field would be able to do it faster, but now people new to the field are able to debug at a very fast pace as well. It looks at ask code generated, understands llvm, can read and instrument gdb, all for 15$ a month

How big is your TF state on your routers?

Yeah, TF is slow if you have thousands of configs and batching not integrated in the provider.

I recently just referenced the selective applicative functors paper and let it write me an implementation in scala. There is one already available in github, so I can't judge if it really just read the paper and implemented it, but the result was so minimal quick and amazing.

whats keeping you from upgrading?

mitchell hashimoto https://github.com/mitchellh/vouch implemented this idea

how is duckduckgo using us?

if they succeed nothing will change for you

maybe its tokens instead of time now? bun has access to an unlimited amount of it

Can we get one in Lithuania?

False, it creates consumer demand for inference chips, which will be badly utilised.

There are so many CPUs, GPUs, RAM and SSDs which are underutilized. I have some in my closet doing 5% load at peek times. Why would inference chips be special once they become commodity hardware?

Worst part about liquid glass changes is that it was probably rushed so they could cover for lack of AI features that they promised a year ago

I think zig is striving for language minimalism, make it minimal as possible then if a feature optimized multiple places use cases benefit

Once he fixes the past he will have more time to fix the future

js string builtin in wasm is a thing now, might be sharable across runtimes with special compile flags

Only contrast I see is that he thought it was much more of a corner case which turned out to be not that true anymore since everyone started using claude code.

Can't they do it here? or will the authorities go after these kind of upgrades?

on one hand they are loosing profit, on the other hand they are gaining on market share. They will probably wait a short while to assess how much they are willing to sacrifice profits for market share

Just like the GPUs.

NVIDIA started allocating most of the wafer capacity for 50k GPU chips. They are a business, its a logical choice.

knuckle-draggers claiming that "Linux sucks because it can't play games"

they still do it because you can't play all the multiplayer games with kernel level anticheats

"We will make linux a viable gaming before we increment that number to 3!"

But I totally agree, I still install windows for gaming on my machine, but it looks like that for my purpose of gaming I can stay with Linux (I play mainly older games or indie games).

This is a weird article. It’s titled “how to use vsock” but 95% is how to set up Bazel, gRPC, and building a C++ project.

This is just what happens when you use bazel to quickly set up your project.

maybe merging patch files was a thing way before git?

you can just point a shotgun at your foot and pull the trigger, isn't that covenient?

Just profile the slow parts and rewrite them in rustm, easy.

I remember my Athlon 64 machine.

The last one to run Windows XP.

I haven't seen a single one.

Just write a check.md instruction for copilot to check it for malicious acticity, problem solved

I think for a lot of people in the LLM scene a 5070 is good enough if it has a lot of ram

I am so tempted. Is battery life good with Asahi?