alias @loadingalias

founder @ alias-research | privacy advocate | rustacean github.com/loadingalias USA Joined May 2023

Tweets

9K
Followers

577
Following

2K
Likes

14K

high_byte @high_byte

34 minutes ago

we need a new type of blockchain - integrated chat / hosting (unstoppable apps) - accounts are verified (passport or other protocol) - passports support human & agent types

2 1 5 126 2

I’ve been working on pluggable reclamation for a substrate I’m working on. I’ve implemented crossbeam-epoch EBR and seize EBR; then hand-rolled QSBR and the first rust impl of StampBD as far as I know… I’ve dropped their “guards” and used a hand rolled fixed atomic array. I’m…

0 0 0 3 0

alias @loadingalias

3 hours ago

What tips or tricks do you guys have to speed up loom tests?

0 0 0 6 0

alias @loadingalias

23 hours ago

The deeper I get into like production grade system - the distributed HA services, the observability stacks, the entire thing - the more I realize that big tech is living in the 2000s at best. Kubernetes fucking sucks. Having LogQL, PromQL, TraceQL, SQL, and a handful of subsets…

0 0 0 21 0

alias @loadingalias

a day ago

Centralizing observability, errors, configs - this is a trap. Don't do it.

0 0 0 19 0

alias @loadingalias

a day ago

A message to the @claudeai engineers. You guys need to fix the issues with your virtual IDE shit using ClaudeCode. If we use an IDE like VSCode/Cursor… the Rust-Analyzer just collapses on every single review, edit, or really any use. I’m having to set the analyzer to “server”…

0 0 0 41 0

alias @loadingalias

a day ago

So, let me get this straight. OTel SDK gives me logs, traces/spans, and metrics w/ my errors integrated. OTel Collector (contrib) lets me manipulate it all, then export via OTLP/Arrow to GreptimeDB. @Greptime natively just accepts it all and their Docker image for Grafana…

0 0 1 37 0

alias @loadingalias

2 days ago

I spent tonight optimizing metrics, logs, and traces for a cost-optimizer to use in near real time. I have metrics that take less time to read than a Boolean. These will power a cost optimizer that uses two lines of defense to make big data queries real-time-ish… metrics +…

0 0 0 23 0

alias @loadingalias

2 days ago

The single most important talent any engineer can bring to the table today is taste. Having taste and understanding how that translates to codebases or software is easily the defining trait, IMO.

0 0 1 26 0

alias @loadingalias

2 days ago

I need an invite to Lobste-RS

0 0 0 20 0

alias @loadingalias

2 days ago

I do wish that AI was a little better at Rust.

0 0 0 28 0

Amanda Fortini @amandafortini

3 days ago

We live in the fattest, most fucked up, most medicated country in the world; the idea that our public health establishment is some sort of unassailable triumph is just madness. So wrongheaded it’s almost hilarious.

100 532 4K 88K 147

alias @loadingalias

3 days ago

Does anyone seeing this have any research into SOTA garbage collection, or EBR/QSBR?

0 0 1 24 0

alias @loadingalias

3 days ago

Anthropic should be valued at 2x OpenAI

0 0 0 27 0

alias @loadingalias

3 days ago

Man, why are loom tests so messy. I hate having flags all over the god damn place.

0 0 0 25 0

alias @loadingalias

4 days ago

delete if death_epoch < min_active_epoch

0 0 0 99 0

OpenVM @openvm_org

6 days ago

We are excited to release GPU proving today in OpenVM v1.4.0, which also includes a new execution system running at 150 MHz and orchestration for distributed proving on hundreds of GPUs. This release substantially improves performance and is MIT / Apache 2.0 licensed.