Software Engineer experimenting in Machine Learning and Blockchains | Code in Rust & Python | Ex @Olacabs @LinkedIn, Algorithms, Bicyclist, World Traveler
Hey @kaifronsdal, any plan for extending Petri to incorporate the optional role of a Diffing agent that probes to find behavioural differences between two or more distinct models?
Very happy to share what I've been working on the last couple of months! Petri moving to Meridian gives it a neutral home and a dedicated maintainer. We desperately need more alignment evals, and sustained open tooling helps make this possible. Lots more to do here
@mureytasroc and @NirvikBaruah: Congrats on the launch. Great technical overview! Out of curiosity, why do you need the sidecar proxy? Can you not just bring TAP interface up/down before and after migration?
Today we're launching Sailboxes: the first cloud environment purpose-built for long-horizon AI agents. Sailboxes are full machines with persistent state that auto-sleep, built for agents that live forever, priced so you only pay for what they actually use. From $0.015 per active vCPU-hour, over 70% cheaper than other providers.
Under the hood: novel architecture from @NirvikBaruah and @mureytasroc that live-migrates VMs based on actual resource usage. To show how robust it is, we ran a Minecraft server in a Sailbox and forced it to migrate every 2 minutes--players never noticed.
Thanks @sheonhan, for this post. It's nostalgic. I once used RoR. It used to be the thing to quickly bring up webapps before one needed distributed systems for scale (rewrites in Java). Today that space is taken by Javacript and Go/Rust, respectively for reasons you've mentioned.
I really should not be buying more books, but what am I supposed to do with Manning sending me emails about their "From Scratch" series with titles like "Build a Reasoning Model", "Build a DeepSeek Model", "Build an AI Agent"...
@penberg Something like Hazelcast. They have both modes. https://t.co/1ZVIJDwwu9
Though @hazelcast has not published any benchmarks. Nothing that I found.
.@penberg, came across your paper, "Impact of Thread-Per-Core Architecture on Application Tail Latency". Don't you think that benchmarking Sphinx against its own non-TPC variant would be more appropriate than Memcached?
@penberg I was evaluating TPC for my Rust-based project, and @glcst's "Introducing Glommio" post took me to your paper. While I agreed with your conclusion & great analysis, I feel TPC could have scored even better if the same system had been benchmarked in TPC vs multi-threaded modes.
.@peregrine Just read your good & old post on https://t.co/89hOHSI258's Rust + Elixir. While it made total sense as you were coming from Elixir code base. If starting a project from scratch, would you not consider pure Rust with Firecracker microvm for performance with recovery?
Just completed 100 Rust Exercises.
Thanks a lot @algo_luca, @marcoow, and @mainmatter for putting it together. I just wish I had discovered it much earlier in my Rust learning journey.
We are excited to release a short course on AGI safety! The course offers a concise and accessible introduction to AI alignment problems and our technical & governance approaches, consisting of short recorded talks and exercises (75 minutes total). https://t.co/8AhEBb3Vpk