It's incredibly satisfying to see the polar opposite of the usual LLM/superDB/K8/CICD/Cloud/Container/Crapola corpobloat we hear about on this site all the time, namely a tiny piece of handcrafted code, ironically produce something infinitely more aesthetically beautiful, and intellectually interesting from an almost artisan engineering perspective.
Especially because some framework slopper using all the LLM's and bloat in the world could never even imagine reaching this level of productivity. In 7 (SEVEN) days this coder
- Designed a language.
- Implemented a compiler to compile it to bytecode, using F#.
- Wrote a bytecode interpreter, using C++.
- Created a shoot’em up game, using the custom language.
- Renderd the graphics, using a single GLSL shader.
While I agree with the sentiment that LLM coding can produce a lot of inefficient junk code which works with holes if you're lucky...
What you're describing is 7 days of productivity supported by probably 7+ years (or 27+ years) of experience and learning and getting things wrong and restarting over again.
Is the code shit? I haven't looked at it. Didn't have to. Probably is. I fed it the blog post at the end, and difference.md has a comparison on what Laurent Le Brun built vs Claude, and Claude is fully aware that it went a different way on a number of different things.
It chose python, I had to tell it to use uv.
I'm on a mac with high DPI and it got confused about that.
I had to tell it to make a binary format (it made a BLTZ header)
But you can clone that repo, do uv run main.py --compile game/shooter.blitz and it'll make a .blitzc. Then you can do uv run main.py game/shooter.blitzc
Tear the code apart. Call out every tiny mistake in that repo. It's probably cheating somewhere! But all I did was give it challenge.png that's in the repo, and tiny bits of English, and Claude went to work.
It's been 21 years since my college computer graphics class and I went into distributed systems and not game dev, so someone else will have to tell me if it used a single (or even any!) GLSL shaders, or not.
Call it slop all you want, but that took me 70 minutes of babysitting.
None of this is individually difficult, but an actual human being had the motivation and talent to bring it all together in 7 days, which is impressive.
So what if an LLM can create the same components if you tell it to. It's a bit like someone sharing a handknit sweater they just made, and you counter with "Well, here's a machine made one I bought in Walmart, made in 5 min in China".
Is it impressive in the way Max Verstappen winning the F1 World Championship in 2023 was impressive, operating at the absolute limit under pressure and getting paid beaucoup bux for it? Or is it impressive in the way your kid is impressive the first time they manage to draw stick figures and a house with crayons? Those are both real achievements, but they are impressive in completely different ways, and the value of the work produced is wildly different. I might fly to Vegas along with 300,000 other people and pay for hotel rooms and pay to watch some shows while I'm there as well as to watch him race, but (and don't take this the wrong way), but I ain't gonna do that to watch your kid draw with crayons.
The difference is the baseline. Once the default outcome is cheap, fast, and good enough, the human effort stops standing out in a way that matters. At that point, pointing at the Walmart sweater is not missing the point, it is the point.
