Results from the ASIC puzzle
61 points by eru 3 days ago | 23 comments
babush 17 minutes ago
I asked Claude to "fuzz the circuit" and, while it didn't work, it basically ended up going the Z3/SMT route and solving it in ~3 hours. This is after the writeups were already out, but still, as a software reverser, pretty wild. Writeup: https://www.babush.me/how-not-to-solve-jane-streets-asic-puz...
replyryontikygod 2 hours ago
Every time I see a Jane Street puzzle of this like I wonder in which way you can improve your problem solving skills for these things and how to extrapolate patterns from other things. I have a gut feeling that this is not only just grinding puzzles, there has to be a way to improve this. Curious if people have some insights on this
replyIshKebab 2 hours ago
> The hardware team at Jane Street designs FPGAs and ASICs that run some of the fastest trading systems in the world, and the day-to-day work is full of puzzles just like this one: staring at a sea of gates, timing reports, or waveforms and slowly teasing out what’s really going on.
replyAlthough not really any more. AI is really good at debugging failing tests so I no longer have to spend hours - sometimes it was even days - staring at waves and instruction traces.
Tbh that's one thing I won't miss.
HFT makes the market more efficient, just like the AI eliminating jobs makes business more efficient. It's been numbing watching HN gladly sacrifice it's own to the altar of AI over and over. There is no moral high ground.
Personally as a retail quant, it's really rewarding to build models that extract edge from the market using the very AI tools that have eliminated all hope of ever re-entering the tech-field absent being an entrepreneurial founder.
It's more efficient now than it used to be - by which I mean the total taken by HFTs is lower than what people used to pay in spreads and commisions. Price discovery is better (you get closer to a fair price if you trade) and spreads are lower. The whole industry that used to prey on spreads and ignorance of retail investors is gone. What not to like about HFT?
Maybe they're both bad?
I’d estimate the total global HFT revenue to be 100 billion dollars a year, which is less than Apple and a tiny percentage of all big tech combined.
They group dozens to hundreds of designs on one chip. The end result is that even hobbyists can afford to have their small ASIC on real silicon. It's still expensive but it's within reach of a serious enthusiast with a steady job.
https://app.tinytapeout.com/calculator The number of "devkits" is the number of chips you'll get, the number of tiles is the size of your design. You can see how many tiles current projects use here: https://app.tinytapeout.com/shuttles/ttsky26d
You think you could do better by having a planned economy, with your eminence deciding what's a waste and what is not? And why should I believe you it'd end up differently than North Korea or Cuba?
Or you'd somehow manage to take what you think are all the good parts of western societies while outlawing what's wasteful and we'd have an utopia?
The irony has most likely circled their drain (which just goes into a creek anyways).
[1] https://en.wikipedia.org/wiki/Anarcho-capitalism
Nobody mentioned that crazy scenario so why bring it up?
BTW the USA is already a planned economy.
As this game got more competitive they seem to have branched out into other kinds of business including HFT.