Why Some Satellites Use NetBSD? | Hacker News Reader

Why Some Satellites Use NetBSD?

machaddr.substack.com

17 pointsBogdanp18 comments

This post could have been a prompt. It is 1000% generated with an LLM.

Tell-tale signs: fluffs up the "key features for x". Each feature follows bold title, highly generic explanation of the title. Then it lists some examples, all following the same pattern "why x".

Then the conclusion is full of gpt-isms. Doesn't say anything of value, just re-iterates fluff words.

> NetBSD's presence in space demonstrates how a well-designed, community-supported operating system can play a pivotal role in advancing space exploration and technology. As satellite missions grow more ambitious, NetBSD’s adaptability ensures it will remain a key player in the evolving field of aerospace engineering.

I know there's a policy encouraging users not to cry "llm generated" every time, but this is quite egregious.

> there's a policy encouraging users not to cry "llm generated"

There is? If so, it should be reversed and LLM-generated crap should be banned.

Eh I found the article interesting. To me it seems more like a non-native English speaker using an LLM to help their English grammar.

The article lists out some examples which I found interesting and didn't know about. I tried getting Gemini to reproduce them but it just gave general examples. Now without those examples it'd be a boring listicle.

Also I haven't seen an LLM spit out a classic sentence like: "This text explores the reasons behind its adoption and how it powers these satellites."

I skimmed through it, yeah, such a pathetic article. It's like having the title "Why satellites use carbon fibre (question mark because of illiteracy)" and having the text be too many words that's just repetitions of "Because it's lightweight and strong"...

At least give us some history about who decided on NetBSD, when, and why they picked it without the generic blah blah.

It's quite funny how "pass a Turing test" is something that changes over time.

I went through the archives[0] and I tend to agree.

There's a post on every day of the week, including weekends.

And although the 'About' page indicates the author might not be a native English speaker, all the posts feel similar.

[0]: https://machaddr.substack.com/archive?sort=new

This Substack article should not be trusted - there are no sources given. I google'd each of the 4 example satellites and "netbsd" and the only results were the Substack article and people referencing the article. NetBSD may have been used in ground systems, but (i) that would have been a non-story and (ii) I couldn't find any evidence of that, so I wonder where the author picked up this information.

I am most familiar with SAMPEX, which was launched in 1992. The initial release of NetBSD, version 0.8, was in 1993 according to Wikipedia. Okay, the article says the project "transitioned" to NetBSD in the "extended mission". Okay, maybe in the 2000s, let's say they decided to replace the original OS/real-time-executive on a working spacecraft with a new OS. So you abruptly replace the old OS/RTE-based flight software with software based on a new OS/RTE. (You don't gradually transition from one OS/RTE to another.) I don't buy it. On a working spacecraft? No.

(I realize the article was probably AI-generated.)

I thought the article was "OK", but my main issue is they did not go into a lot of detail on the Satellites.

One case the mentioned date launched. I would like to know:

* dates launched

* Satellites still in active use

* What NetBSD release was the software built from

* Are the Satellites still being produced in 2025 and if so are they still based on NetBSD

So in a way, the article had nothing "new".

I really want an excuse to run NetBSD on something. I wonder if ESP32's are capable of running NetBSD? Even just running on a RPi would be fun.

Also, for embedded devices the coherent whole system design of *BSDs with the package builder seems like it'd be much nicer than BuildRoot or Yocto.

I've come to very much dislike trying to keep Linux images for embedded stuff running over years. Perhaps it's because Linux claims not pride itself on "not breaking userspace" but seems to add a new GPIO sub-system every 5 years. Ain't nobody got time for that.

> I wonder if ESP32's are capable of running NetBSD?

Interesting question but can't find much about it [0]

I guess for the most recent ESP32 it will be under RISC-V.

People have been able to boot Linux on them for a while so why not...

- [0] https://wiki.netbsd.org/ports/

It’d probably depend on whether NetBSD requires having an MMU or MPU.

RISC-V support might be an option.

Looks like a MMU is needed:

https://wiki.netbsd.org/projects/project/mmu-less/

Check EndBOX - NetBSD running on RPi and booting into a version of BASIC. You can probably strip the BASIC part or get inspiration of how to run NetBSD on a Pi. You can find more information at https://www.endbasic.dev/2025/07/endbox-diy-kit.html

That’s awesome!

Why so many articles have grammatically incorrect headlines?

Because many writers are writing in their second language, and sometimes they use language structures from their native language.

Author is Portuguese, that rule is the same in Portuguese.

I'm second language too, so I checked the article, and it seems fine otherwise apart from the title; I'd expect the title to get extra care. Anyway it's flagged anyway.

Reply on news.ycombinator.com