Sigh... HN littered with it these days.

I think others need to embrace that too. Your work can be fun, and can be a joy. But that shouldn't be folks only outlet for themselves to measure.

At one point in my career I was ONLY reviewing and guiding off-shore contractors. It felt even worse than LLMs. But I did enjoy it though, but going to LLMs didn't drastically change my workflow. LLMs feel like leading a bunch of juniors with no real stake or business context.

So for me it didn't fundamentally change things, it just meant I could steer the progress before a PR is created later in the week that misses the mark completely.

There were plenty of other issues in the software development business that sucks the soul of creating out.

So I think you asking the right questions. Cause I felt like I've asked myself those questions even before LLMs.

This video helped me recently too [1]

Also I take pictures and have Gemini critique, give feedback. It helped me learn about the tips more I didn't realize, after switching my tip from a broader knife one to a chisel one, results way better (for that use cases)

Then I did know I just had to practice and get frustrated and come back, felt like two steps back and one forward, one time I was just suddenly making cold joints left and right. But hours practicing can't be replaced by just reading.

This is more answering the title.

https://youtu.be/GoPT69y98pY?is=JnURRm3pQOarur9P

But now game development has much better tooling and SDK kits for a while in consoles. So I suspect there wasn't a need to spend extra development cycle implementing custom debug menus and stuff.

We are making the kids endure that out of fear of their college, but stats don't back up the grind they make the kids do.

Software is fast to iterate and test that a lot assumptions can be proven by actually writing the code.

Software is closer to gardening or painting. We discover a lot through practice and writing code. Then we can often write more formal specifications.

But formal method is impractical for most software upfront, and instead is likely used for more serious runtime failures or cost of life.

That's just my two cents.

So their are apps that do use similar or ECSs architecture under the hood, it is just more problem related not a solve everything solution.

But I did think, one station info per station probably also did not make sense, technology-wise, so I thought expanding it worked.

The faction system was my answer to what makes ships or planets' goals go beyond their immediate needs, and factions came naturally next. They want to make money and shift priorities to do it through incentives. Like subsidies almost.

I was hoping/thinking they (factions) could escalate to blockades, monopolies or just more supply control, skyrocket prices and events combat to hold choke points, blockades or sanctions. That way that agents FSM is focused on those goal oriented actions and ships can focus on their well being plus their factions incentives, and the crews needs handle immediate needs.

Does that make sense? (Terrible at proofreading)

Bevu/Rust just made it easier to brute force that scaling, since they have smaller footprints (overhead?)for each thread/process spawning.

And yah, ideally there should be tons more! The shift will be interesting.

A space-economy sim where nothing is scripted. A few hundred autonomous ships each run their own planner. Some chase the best trade route, take a delivery contract, refuel, retrofit at a shipyard, or dock so the crew can rest before morale tanks.

Markets price everything off supply with shortage-urgency multipliers, factions tax and subsidize, populations migrate when they're unhappy, and stations that go broke get abandoned and rot.

It started as an Elixir/Phoenix prototype, but the BEAM scheduler struggled on Windows gaming PCs, so I had Claude rewrite the engine in Rust.

The sim core is pure, synchronous, IO-free (its own hecs ECS), and the Bevy client embeds it directly as a library, sim and renderer share one world with zero marshalling. Ship AI is a GOAP planner over a world state; ships replan mid-flight when a better option appears. ~485 agents today at p50 ~10-20ms/tick, architected to push toward 100k+. Single native binary, bundled SQLite, no runtime deps. Getting into 100k has been a struggle, but I have pushed it into the thousand and it was running fine.

It's a sandbox, not a game yet, there's no objective yet, and I'm not actively pushing it toward "shippable." Anyone can fork it and take it over. Or it can stay as some type of AI slop, but part of me thinks it is already in a pretty good shape for a simulation use for various games or ideas.

(Full disclosure: a lot of this was built pairing with Claude. That's how I had the bandwidth to take it this far. Happy to talk about what that workflow actually looked like.)

Comments URL: https://news.ycombinator.com/item?id=48996187

Points: 110

# Comments: 30

In fact, when we don't discuss China, most find modern software patents and copyright to be abusive.

Finally, copying UX standards has been done for ages, it gives users a comfortable way to start using a product and instead of fighting that to be unique, some rather embrace it.

It can be jarring using so many different apps if each one tries too hard to be unique.

And this is my two cents, here it makes sense to have a similar interface, it's ultimately an app with chat sessions.

A not cool thing is it's using mostly cloud AI services but that just lets me work on it quickly without dealing with what local models to run problem.

I also think it should be easy to hack, so if you wanted to run local models, change the interface or just use the interface itself or something else.

Just wanted to show it off and share!

EDIT: sorry about the git history, I developed it in private without curating any of my commits so it was easier just to wipe the commits and start fresh then leak my stuff.

Comments URL: https://news.ycombinator.com/item?id=48908340

Points: 1

# Comments: 0

Sigh, there definitely an attitude creeping into the community and it's a huge turn off unfortunately.

For PDF stuff I haven't needed much Lua filters since switching to WeasyPrint for the PDF engine.