You might also know me from my work on OSS tools and libraries (JS / Wasm / Rust) or Wasm DevRel at Google Chrome
📝 https://rreverser.com/
📷 https://instagram.com/rreverser
> some guy comes up and asks if I want hash
> I ask if it's a cryptographic hash
> he doesn't understand
> I show him the NIST definition
> he laughs and says "it's a good hash"
[...]
> it's not a cryptographic hash
Would be expensive as hell, and I bet Microsoft/Github thought of and played with it too,but what if each repo's history would be used as a training input to produce per-repo instructions automatically behind the scenes?
Why does it need so much context in prompts when the entire project history is in git/GitHub?
Would be expensive as hell, and I bet Microsoft/Github thought of and played with it too,but what if each repo's history would be used as a training input to produce per-repo instructions automatically behind the scenes?
GitHub Copilot: 1 minute, excellent answer
Claude Code: 2 minutes later, still waiting
GitHub Copilot: 1 minute, excellent answer
Claude Code: 2 minutes later, still waiting
Oh yes, little Billy Ignore Instructions, we call him.»
Oh yes, little Billy Ignore Instructions, we call him.»
It gives me correct, concise, code for complex problems on the first try 90% of time, with no need for follow-up "conversations".
Impressive, considering they were a bit late to the game.
10/10 for effort, 0/10 for usefulness.
It gives me correct, concise, code for complex problems on the first try 90% of time, with no need for follow-up "conversations".
Impressive, considering they were a bit late to the game.
Me: "That was impressive, you gave me the right one without asking for my name and address."
Pharmacist: "Oh, haha, yeah, I know you, I see you here often."
💀 I guess I'm a regular now. Love getting older.
Me: "That was impressive, you gave me the right one without asking for my name and address."
Pharmacist: "Oh, haha, yeah, I know you, I see you here often."
💀 I guess I'm a regular now. Love getting older.
I'd never ever think to use it, but apparently it's a very efficient way to reverse bits in lots of bytes:
I'd never ever think to use it, but apparently it's a very efficient way to reverse bits in lots of bytes:
I don't need "this is great question", "spot-on observation!", "you are absolutely right" and all kinds of nonsense reassurance from a fucking computer.
I just want a query in - answer out, as concise as possible. Why is this so difficult.
I don't need "this is great question", "spot-on observation!", "you are absolutely right" and all kinds of nonsense reassurance from a fucking computer.
I just want a query in - answer out, as concise as possible. Why is this so difficult.
Photo: Eyevine
To be fair, I've only used it to rebase one branch in one Rust project now but it resolved conflicts all by itself when pure line-based merge couldn't - exactly what I'd expect from an AST-based merge tool!
We're migrating a database from C to Rust. The project is complex and performance-sensitive. That's what makes it fun!
If you're interested, send an email to the address in the screenshot.
We're migrating a database from C to Rust. The project is complex and performance-sensitive. That's what makes it fun!
If you're interested, send an email to the address in the screenshot.
And they promptly misconfigured so people could download the source and leak it to GitHub:
github.com/rxliuli/apps...
And they promptly misconfigured so people could download the source and leak it to GitHub:
github.com/rxliuli/apps...
Shows OpenMP -> Rayon migration, tricks for instruction level parallelism, SIMD, compares assembly and performance of gcc/clang/rustc generated code, and overall explains the concepts very well.
Shows OpenMP -> Rayon migration, tricks for instruction level parallelism, SIMD, compares assembly and performance of gcc/clang/rustc generated code, and overall explains the concepts very well.
Thank you for coming to my TED talk about confirmation biases.
Thank you for coming to my TED talk about confirmation biases.