No, I'm saying they are hyping up some random podcast while saying sweet nothings. I really don't like empty hype. I want some actual content to posts that actually says something, not just links breathlessly encouraging me to go spend an hour listening.
Not sure where you get the breathless bit from. If you don't want to listen, then don't. If you want me to summarize everything for you and give you the bullet points that is also something that will take more time than I am willing to commit at the moment. You can listen to it, or not. No skin off my back. The irony with some of the commentary is along the lines of "don't listen to influencers" and the like. I provided a link to (what I feel to be) a very worthwhile listen that is the antithesis of the grab-a-headline news of late. I have no connection to this particular podcast and have listened to a handful of episodes over the years on subjects that interest me. This particular subject (Huge IPOs, what is or isn't tracked, what float is, etc) turns out to be a relatively deep domain and there are inherent tradeoffs between different approaches. That's all.
The middle one is the ad-hominem puffery. The rest isn't quite exactly 100% 'of the person's, but still doesn't give me any actual leads into what the content is: its just empty puffery.
on my M1 Max, MTP consistently lowers my performance! I’ve tried both llama-cpp’s recently landed MTP support (cloned and built Tuesday) as well as one of the other forks a few weeks ago. Suspect nobody’s done a comparison on hardware like mine.
i built something with a similar philosophy here: https://github.com/khimaros/airun -- it is intended to be piped and redirected. it discovers skills, AGENTS and prompt templates from Claude Code, Pi.dev, OpenCode and others. no TUI, but does have a basic tool calling loop
$ airun -q -p 'output a shell command for linux to display the current time. output only the command with no other code fencing or prose' | airun -q -s 'review the provided shell command, determine if it is safe, run it only if it is safe, and then summarize the output from the command' --permissions-allow='bash:date *'
While I think that the core philosohpy is the same, i'd like to ask: why adding features like Skills and prompt templates?
I personally decided to not implement Skills and instead using a prompt library approach, where certain .md are used to fully replace the system prompt, in order to allow for an approach similar to Skills with ~100 LoC dedicated to this system.
Skills are _like_ prompts, yes, they're extra info added to the context. A prompt is just a prompt though, an agent like Claude could use multiple skills in one go, which seems impossible to do with Zerostack.
Skills are not just prompts.. the entire problem that skills solve is runtime discoverability via a skill description. Agents can self-recognize that a skill would be useful in a situation, and then load+use.
Prompts are just text templates entered by the user, and the user must specifically know when to and remember to invoke them. If you’re just using skills as if they are the same as prompts, you’re totally missing out on the entire benefit that skills provide!
they aren't, but they often push kernel/system patches faster than Google. they also have more kernel hardening in place, which makes some classes of exploits ineffective.
haha exactly. I’m coming from Swift, and I don’t want to go back to manually releasing objects like I used to in ObjC, let alone reason about lifetimes.
What's the big issue with GC nowadays? It has mattered to me exactly once in decades and it was still manageable anyway by using a more low level style in a hot loop. I see very few usecases where GC actually matters and for those rare few cases it was not like you were using python beforehand anyway
Why the hell would he "worry about garbage collection"? That kind of thing is a cargo cult fear.
Garbage collection is not an issue for 99% of programs. And for those that it is, there are ways to mitigate the issue (e.g. there are extremely high performance trading system written in Java, where every last sub-millisecond counts).
Blanket fear of GC reminds me when new programmers learned about how assembly is lower level and can be faster, and wondered why everything is not written in assembly.
as a person who just wants to publish a simple blog and informational articles, i would happily use this subset if it were still compatible with popular browsers. i've been praying for an effort like this to organize and am grateful that you are taking a stab at it. i would use dillo as my main testing browser if it was the browser that honored the subset spec most accurately and guaranteed compatibility with Chrome and Firefox.
reply