961–970 of 1001 posts
Trying to make the world’s smallest maze solving neural network. Why? Because it’s fun. Why aren’t you?
Current champion model is 14 bytes - it solves 96.5% of unseen mazes upto 21x21.
Try it yourself on the website. Also a work in progress write up at [1]
[0] - https://minimio.ai
compears.shop : to help with cheapest grocery shopping across Europe
https://runewardd.github.io/runeward/ : what if you know every action your agent is performing and how about you have a fleet of agent and you can map out how each of them works with boundaries to work with
https://github.com/adefemi171/ai-rganize: Organize files/folders locally using local models or frontier models
This month has mostly been about adding multilingual support to Jant. It's for bilingual authors — the common case being someone writing for both Chinese and English readers — so they can publish to each language's audience separately within Jant.
I'm also designing a feature called Repost: when you see a post you like on another blog, you can repost it to your own Jant blog along with the original URL.
Code's on GitHub [1], and there's a demo [2] to try it out.
[0] https://jant.me [1] https://github.com/jant-me/jant [2] https://demo.jant.me
AI quant trading (so that I have another income and can stop being threatened by managers that if I don't meet their unrealistic expecations that I won't have food on the table next year). So far it's going well, 30000% gains on my initial capital in 3 months very gradually over 100K trades. But I need to build some more sleeves because (a) the edge could die any time (b) I hit capacity of that edge and it can only g…
Playing this game online used to mean installing something first: a client on your PC, an app sideloaded onto your iPhone, or a tabletop sandbox where you apply the rules yourself. And your opponent had to do the same before you could play. Now you open a link and you're in a match, on any device.
The hard part turned out to be the rulings, not the netcode. I play this game in real life and I still had to look half of these interactions up, because what a card says and what a card does are two different documents. New sets land almost every month, and every one of them is a pile of new special cases to implement and test by hand. Which is why I now spend a suspicious amount of time wondering whether I should just train an agent to do it instead of me.