Doesn't MS have the compute to run GPT 5.5 for all its employees?

What about BSDs?

I don't support non-macOS BSDs explicitly yet. Not for any reason of design, just hasn't been a priority.

syscalls

That might work on FreeBSD but is pretty well guaranteed to break on OpenBSD. (Dunno about Net and Dragonfly) (I'd caution that treating the BSDs as a monolith is likely to end in errors; they're quite diverse.)

I find JSON better for serialization/deserialization. It's a bit tedious to write by hand but who writes JSON by hand?

Write json5, which gives you most of the brevity of yaml without the terrible ideas

I find even plain json easier to write than yaml. Especially when you factor in the scope of mistakes. Tiny mistakes can completely break the structure of a yaml document in ways that are still valid yaml. With json I'll catch that because auto-indenting will follow the actual structure

>You are literally one missing tab away from breaking your entire configuration file.

Isn't this true for Python, too?

It is, yes. And unlike YAML, there's no official way to use a syntax that doesn't use whitespace to define scope.

Python's whitespace-defines-scope syntax is the reason I won't use it unless I have no other choice. If it weren't for that -IMO- extremely poor design choice, I'd consider it to be a decent language. [0]

[0] Yes, some people think this feature of Python is fine. I'm not one of them. Whether or not you think this feature is fine is a matter of taste.

I think Deepseek uses much more tokens than other models.

But way less dollars. Which is the important metric.

There are lots of more mature plugins doing this: superpowers, get-shit-done, compound engineering, spec kit.

That's fair, gsd, superposers and spec kit are way more mature and complete. On the other hand, this skill is indeed narrower, and as a result currently intended for spec generation only, standard markdown output and allowing work alongside any of those frameworks at the spec later. Different scopes instead of a replacement.

Thanks for your input and the provided references as well as taking your time for reading this show HN.

Sure, this skill is narrower and that might be its advantage, if you don't try to make it solve everything but instead if you evolve it to solve some particular use cases.

I use cheap Chinese models. For all I care, both OpenAI and Anthropic can raise their prices until they'll have no customers left.

That's a nice problem to have. I can't afford a $48K GPU server, even if I worked as a developer since 25 years ago, because I live in the wrong place.

I was pretty excited to find out that FreeBSD now supports Podman and OCI containers so now I can move some of my web apps from Linux to FreeBSD. :)

Do all llm know they are a LLM? It doesn't depend on the system prompt?

The pre-trained ones no (except some of the new ones which have post training data added to pre-training for some reason). The post-trained ones yes (at least all the ones I've seen).

Some of the niche ones I'm not sure about. Like the historical LLMs. I have not tested those yet.

I think any instruction tuned model is going to "know" it's an LLM.

Yes. The first step of aligning each and every GPT-based LLM is to suppress the “I am human” kind of responses. It’s baked into the weights.

Reminds me of old cleverbot conversations where it would always assert it is human and you are the bot.

Trained on previous conversations with people.

It's also at minimum baked into the system prompt of virtually any LLM.

That's not "baked" and only applies to remotely hosted LLMs where someone else feeds the prompt into the LLM.

Without a system prompt no. And in general they “know” nothing and just predict the next best word.

This is wrong. See other comments.

For sure, as they are stochastic parrots. My question should have been: what are the odds a llm would react properly to those instruction, but I got lazy and asked if they "know" it, because I presumed most readers here do know how llms are working.