stuartriffle

6 ms·

1.
by stuartriffle 1y ago
Absolutely not, but ChatGPT generated some Lean, I posted it in a comment. I'm not set up to run that, downloading stuff now. I guess we'll see. [edit] I have deleted the half-ass ChatGPT generated Lean because it doesn't com
2.
by stuartriffle 1y ago
I asked about it on Math StackExchange twice, both posts were deleted within minutes as off-topic. I formalized the proofs and came back, but I'm apparently banned for a week because I had two posts deleted. I am fully aware that what
3.
by stuartriffle 1y ago
Here are the basic claims that lead to the p-word, all proven with high-school math, validated by AI. Please, someone, find the error. I need to get some sleep. # I. Exhaustive Walk Over Prime Congruence Classes Theorem 1. Let p be an odd p
4.
by stuartriffle 1y ago
Here are the basic claims that lead to the p-word, all proven with high-school math, validated by AI. Please, someone, find the error. I need to get some sleep. # I. Exhaustive Walk Over Prime Congruence Classes Theorem 1. Let p be an odd p
5.
by stuartriffle 1y ago
Yep. Me too. This is simple modular arithmetic. All the major AI buy my nonsense though. I'd very much like to know what's wrong.
6.
by stuartriffle 1y ago
### Key Analytical Components 1. Forbidden Congruence Classes : - For each odd prime $$ p $$, after encountering $$ n \equiv a \mod p $$, all numbers congruent to $$ a \mod p $$ are eliminated from future trajectories. This creates a gro
7.
I think I know why Collatz does that thing it does
3 points by stuartriffle 1y ago|5 comments
8.
by stuartriffle 1y ago
In case you're curious: ### *Key Analytical Components* 1. *Forbidden Congruence Classes*: - For each odd prime $$ p $$, after encountering $$ n \equiv a \mod p $$, all numbers congruent to $$ a \mod p $$ are eliminated from future
9.
I think I know why Collatz does that thing it does
5 points by stuartriffle 1y ago|5 comments
10.
by stuartriffle 2y ago
The language spoken by participants during a phone call is also easily classified as “metadata” and could be defended as non-identifiable info with a straight face. So are the number of speakers on the call, the topics of conversation, the
11.
by stuartriffle 3y ago
Yes (imho) just be sure to get the instruct or chat model of any LLM you try. There is an awesome snapshot of what models people are using here: https://openrouter.ai/rankings
12.
by stuartriffle 3y ago
I've been learning about RAG using LlamaIndex, and wrote a small CLI tool to ingest folders of my documents and run RAG queries through a gauntlet of models (CodeLlama 70b, Phind, Mixtral, Gemini, GPT-4, etc etc) as a batch proccess, t
13.
by stuartriffle 7y ago
I have had good results encoding/decoding the NTSC signal. You get crosstalk artifacts for “free” https://www.gamasutra.com/view/news/125898/Opinion_CRT_Emula...
14.
by stuartriffle 7y ago
I’m working on a brute force MCTS engine right now. It scales across SIMD (to 8x) and GPU (only CUDA so far). https://github.com/StuartRiffle/Jaglavak My GTX 780 can play about 1M games to completion per sec. Still und