Builder. Shipping the pieces of a Rust-first AI engineering stack: agent-assurance, sqry, llm-cli-gateway, and with more coming. Views are my own.
Gold Coast, Queensland
Joined November 2024
- Tweets739
- Following108
- Followers144
- Likes1.2K
right. no thanks. - it compiles Rust through a separate engine and patches executables after edits. That makes the speed claims plausible, but compiler correctness and stale builds are central to the decision.
So you want rust for safety. and then cut it off at the knees because you don't want the safety enforced
Oh boy! This sounds awesome! (Haven’t tried it yet.) Rust is the best programming language for AIs because of the safety orientation and tooling, but the long compile times are a big limitation. This appears set to eliminate that friction!
github.com/indexable-inc/bla…
How more useless can Claude become?
@ClaudeDevs @claudeai
First, you burn down tokens at a RIDICULOUS rate for the last two weeks. Now, again, your model, refuses to do the stuff your model tells it to do????
A mathematical proof can establish that an answer follows from its premises. It cannot, by itself, establish that we asked the right question, that the premises describe reality, or that the answer is worth acting upon.
Claude hit its weekly quota until October 26.
@ClaudeDevs - really, a single dev on your max subscription plan, and I'm hitting the quota in 2 days?
Last week it at least got me to Friday, and a few weeks before that it lasted the whole week.
what are you guys doing with the harness and accounting?
Claude code is sucking subscription usage like a drunk sailor again. Anyone else having the same issue? My Pro Max x20 is only doing 2 days atm, I wish they would stop fiddling with the counting algo's. Not the first time this happened. @AnthropicAI @ClaudeDevs @claude_code
let it run cross model (vendor) reviews like I do with llm-cli-gateway. i.e. you have a whole team with different experiences working together, building on and improving eachothers' work. ping me and i'll show you what that could be
Generation became almost free. Verification did not. The bottleneck is now trust, not production.
🤖 Made with AI
Generating work became almost free. Verifying it did not. The bottleneck is now trust, not production.
Discernment is worth little without standing: the ability to see inside the work, choose a disposition richer than Yes or No, and make evidence of repeated failure change the system that produced it.