Mesh processing is a very difficult research domain in computer graphics that has been iterated for several decades, and we still don't have a good automated solution for retopology (Partly because the problem is hard to define in a mathematical way, but also since it's not a problem you can just solve with AI by throwing data and compute at it)

Aligned with the theory of Bullshit Jobs - LLMs expose the fact that the white collar work most of us have been doing at this point were actually bullshit. When LLMs "fake" work, it actually hides the reality that there was no meaningful work here in the first place.

It definitely still is a hardware problem today - humanoids force you to miniaturize gears, motors, and other parts (especially with the hands) which make them incredibly fragile and inaccurate. You're basically fighting against the laws of physics, so improvement on this front has been pretty slow. And tactile sensors which are key for complex manipulation tasks are still a far way to go in terms of resolution and reliability - so most robotics startups tend to rely on cameras for everything.

I think that in order to have humanoids that are actually capable of matching or exceeding the actual mechanical capability of humans, you need large advancements not just in AI but in material science as well - no machine can still match the efficiency of humans with its biological muscles, tendons, the skin / fat that surrounds them, and its vast array of sensory input.

I think it's important to know and practice your passion, even if you have to work on something different to pay the bills. You can only be good at something if you really like it, and you never know what opportunity you'll stumble onto if you're ready for it.

Programmers (and other white collar jobs) were able to luxuriously coast along the ZIRP era because capital (replenished twice via quantitative easing) was cheap and plentiful, and because the elites at the top had to pump huge amounts money to create a shared fantasy of the "technological future" that validates the neoliberal era. Now that the reality of the actual "physical economy" (the economy of making tangible things) has clawed back at us because of that forbidden three-letter word (war), we all realize that doubling and tripling oil prices were actually dictating our lives rather than some "Skynet AI" crap, and thus our fantasy simulacra of "virtual" play-things have now come to an end. Oh and we all found out that most of SaaS was actually bullshit anyway. In fact, if it could be completely replaced by AI then it was already pretty bullshit in the first place.

So, for smart STEM people uninterested in programming and only looking for a stable career, I think they would be better off by just doing engineering work that's a bit more tangible, like robotics, manufacturing, shipbuilding, construction, etc. (Or anything related to war, but only if you're able to stomach what you're doing.) If you don't like to sit all day for a salary, then niche blue collar work can also be a good option, since general-purpose robotics (Physical AI?) is still too far away because of many, many issues that's just too long to explain here. I still think if you like programming then you should stick to it in the long run - there will be a very cold winter because of the combination of LLMs, AI bubble pop, and general economic depression, but for those who survive this era there will be an opportunity because of the shortage of skilled programmers (since no-one bothered to hire juniors after the pop, no one will grow to become seniors themselves!) Computing will still be with us forever, just not in a way that investors thought that it's going to "engulf the world".

Comments URL: https://news.ycombinator.com/item?id=47387438

Points: 1

# Comments: 0

The best examples I can think of are Thumper, Devil Daggers, Hyper Demon, The Witness, and No Man's Sky. And of course, Minecraft!

I've made my own in Python that generates Ninja files only - and it's surprisingly not that much work (especially if Claude Code can help you).

> I experimented with Rust/Bevy and Unity before settling on Godot. Bevy’s animations and visuals weren’t as crisp, and Claude struggled with its coordinate conventions - likely a combination of less training data and Bevy leaving many core features, like physics, to the community. Unity was a constant struggle to keep the MCP bridge between Claude and the editor healthy. It frequently hung, and I never figured out how to get Claude Code to read the scene hierarchy from the editor. Godot’s text-based scene format turned out to be a huge advantage - Claude can read and edit .tscn files directly.

Didn't expect Godot to be the most friendly game engine for LLM usage! I think it's because of various factors - Godot has been used quite a lot in recent years so there are various code examples on the Internet, and its scene file format (.tscn) is very concise enough for LLMs to write and edit directly (Unity has its own YAML-based format but it's very unfriendly for human consumption, and Unreal stores its core assets in binary files)

Though I thought a company large as Microsoft would have the resources to build a cross-platform RHI with the most stable API available for each platform (DX12 for Windows and Metal for macOS)...