It's worth noting that only some of these techniques are configurable in modern LLM API outputs. (usually only temperature/top-p/top-k since other penalties require overhead)
Most other penalties don't require much overhead (min_p is basically free).
Most techniques are not made available by API providers because they enable alignment breaking. It's the only explanation for why we are still stuck with only top_p, top_k, and temp of 0-2.
If you want proper sampler settings to be available, your options are oobabooga, sillytavern (dependent on your backend, so vllm backend for example doesn't have top-n sigma yet), or directly running huggingface code. There might be some marginal options here too but in general, sampling innovation is firmly in the hands of open source coomers right now and not in the hands of academics.
As a car audio enthusiast, the biggest obstacle to putting a system into a new high-tech car is bypassing the deeply-embedded infotainment system while retaining decent aesthetics and steering wheel controls. The idea of getting an electric drivetrain and new-car safety with a 90's-style blank canvas for audio is amazing.
I hope that the noise isolation and intended speaker mounting locations are good!
I read this as the parent complaining about other car manufacturers selling you crappy default stereos so that you'll upgrade, not that Slate is excluding a stereo on this truck to upsell you.
In fact, I would be rather surprised if you could buy $4,000 worth of stereo equipment for this car, given their promo materials seem to include a $100 bluetooth speaker below an iPhone.
The design shows a fundamental misunderstanding of sheet metal. Flat sheet metal is weak. Only curved sheet metal can be strong. Designs that lack mechanical sympathy with the materials in use don't tend to age well.
Branch-and-bound is an algorithm "from the book" to me. Fundamentally very simple, provided you view the LP solver as a black box, but incredibly useful.
Sorry for being a grumpy old man, but I don't have npm on my machine and I never will. It's a bit frustrating to see more and more CLI tools depending on it.
You could just run it in a Docker container and not think about it much after that. Mount a volume to the container with the directory contents you want to be available for edit by the agent.
JS is web's (and "hip" developer's) python, and in many ways it is better. Also the tooling is getting a lot better (libraries, typescript, bundling, packaging, performance).
One thing I wonder that could be cool: when Bun has sufficient NodeJS compatibility the should ship bun --compile versions so you dont need node/npm on the system.
this is a strong HN comment. lots of “putting a stick in my own bicycle wheel” energy
there are tons fascinating things happening in AI and the evolution of programming right now. Claude and OpenAI are at the forefront of these. Not trying it because of npm is a vibe and a half.
Why? I am not the biggest fan of needing a whole VM to run CLI tools either, but it's a low-enough friction experience that I don't particularly care as long as the runtime environment is self-contained.
Yep, this is another one of the reasons why all of these tools are incredibly poor. Like, the other day I was looking at the MCP spec from anthropic and it might be the worst spec that I've ever read in my life. Enshittification at the level of an industry is happening.
Yes, no, I think he has different mental capabilities than most (most of the commenters here at least) and by that actually are living in a different reality. Human brains function vastly different. Two examples stood out to me.
1) UB said he reads the code in its full from left to right with the if(isTooHot) example. I only resort to reading code in that way as a last resort if I really can't figure out what the code is doing. I mean I look at a block or row and take it in more as a whole.
2) UB said comments are annoying because he has to read them and keep the whole of the comment text in his mind. This again says he reads everything left to right, and he can likely store everything that he has read up to a certain amount.
My mind works nothing like that. I can hold very few words in my working memory but can instead hold concepts/ideas. For that to work good I need to see as much of the involved code as possible and my mental image evaporates if I have to navigate too far where I started.
TOSLink was kind of a silly idea because digital electrical signals would also prevent ground loops. The key is digital vs. analog, not optical vs. electrical.
Ground loops comes from the ground mismatch between two electrically connected devices. When you use an optical link, you isolate those two devices since there is no common ground and the hum goes away. Same if connect a battery device to a grounded device.
reply