Vitaly Slobodin (@Vitalliumm) header
Vitaly Slobodin (@Vitalliumm) profile picture

Staff Random Engineer at @gitlab. ex-@PhantomJS developer. Maintaining the Ruby extension for Zed editor. šŸ¦‹@vitallium.bsky.social 🐘@vitallium@fosstodon.org

The Netherlands
Joined October 2010
Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)Photo by Vitaly Slobodin (@Vitalliumm)
šŸ¦‹@vitallium.bsky.social 🐘@[email protected]
1
1,037
No le Chonk means no benchmark </thread>
Four new versions have been benchmarked in Agents on Rails: @OpenAI's GPT-6.2 Astra (xhigh and medium) and GPT-6.2 Sol (xhigh), and @deepseek_ai V4.1 Flash 1017 (max). Here’s what we found: New first place: GPT-6.2 Astra. At xhigh effort it scores 65.8%, narrowly ahead of Claude Opus 5.5 (65.3%), at $4.42 per run versus $10.53. It’s also faster than the previous Astra. At medium effort it lands at 4th place and averages under 11 minutes per run, the fastest configuration we’ve tested. GPT-6.2 Sol is nearly twice as fast as GPT-6.1 Sol (about 12½ minutes per run versus 23½) at roughly the same cost. Its accuracy dipped this round, from 60.8% to 54.1%, though some of that may be run-to-run variation. DeepSeek V4.1 Flash 1017 is a big efficiency gain. It scores within a few points of the previous version while running about 1.5Ɨ faster at roughly one-sixth the cost. At $0.16 per run, it’s the cheapest model we’ve tested to date, but you get more accuracy and speed with Luna or Sol for just a little more cost. Thanks once again to the team at @evilmartians for their work. Explore the full results at: rubyonrails.org/ai
2
88
That’s fantastic! I’m slowly but surely converting people at GitLab into @gitbutler fans (and @zeddotdev 😁)
If you use worktrees and Rust or Node, you should check out GitButler's new worktree support. Throw your `node_modules/` or `target/` directories into a .worktreeinclude file and new GitButler worktrees do copy on write if possible. It's way faster and saves tons of space.
1
7
349
Vitaly Slobodin retweeted
If you use worktrees and Rust or Node, you should check out GitButler's new worktree support. Throw your `node_modules/` or `target/` directories into a .worktreeinclude file and new GitButler worktrees do copy on write if possible. It's way faster and saves tons of space.
12
7
2
159
9,353
FYI, in v0.16.21 of the Ruby extension for @zeddotdev, we removed the Ruby snippets since they weren’t being actively used. The snippets are still available in the `ruby-snippets` extension.
1
156
šŸ”„ I really miss being able to use Delta at work...
New in today's Delta release (v0.17) Ask an agent in one thread to kick off multiple tasks, each in its own independent top-level thread.
1
228
Finally! An AI agent that can replace me! Dobby is free!
I made an always-on agentic on-call engineer šŸ’š It’s an AI agent that keeps working after you cut off its external network access. It runs @NVIDIAAI Nemotron 3.5 Lightning locally on DGX Spark, investigates a broken service, changes the config, and verifies the repair. Combination of open-weight models + local hardware is especially useful when you need control over both the model and the environment it runs in. Open weights give you control over the model itself: - which exact model version is running - can you preserve that version - can you customize it - when you choose to upgrade or replace it Local hardware gives you control over where the work happens: - where your data is processed - does it ever have to leave your environment - does the system depend on an external model API - can the application can keep working when external connectivity disappears You don’t necessarily need both. You might want open weights but still run them on rented GPUs. You might need local inference for sensitive data without ever fine-tuning the model. But when you need control over both the model and where it runs, the combination becomes very useful. Watch the full video: youtu.be/_3MKLsli12w
2
356
TIL that you can add multiple projects to a single thread in Delta by @zeddotdev
1
8
2,206
Vitaly Slobodin retweeted
Delta is now in public beta. It’s a multiplayer environment for coding with agents and reviewing what they build, with every code change linked to the conversation that produced it. Anyone can download it today on macOS, Linux, and Windows: delta.dev/download
66
93
43
1,311
143,478
We merged an MR with Archspec today! Thank you again, @paolino !
4
308
Billing is a nightmare. The perfect place for engineers to burn out.
i think every engineer should have to do at least a 1 year stint on a billing platform team. its a rite of passage it'll teach you fast about prioritization of work, working across orgs (technical and non technical), how products actually impact the business, engineering topics like idempotency, eventual consistency, how floats actually work and why they're the devil when it comes to money
7
1,685
GitLab still uses Vue.js 2 (and Jest). I don't think we'll migrate to Vue 3 this year.
Asana says OpenAI Codex completed an engineering migration it had expected to take at least five years in about two weeks, for roughly $12,000 in model and infrastructure costs. The project: removing Enzyme, an outdated testing system that had become a blocker to upgrading Asana’s frontend stack. The company’s previous staffing plan estimated the migration would cost roughly $6 million. Instead, Asana gave Codex a prompt only five sentences long. Up to four coding agents worked in parallel, each operating in a separate copy of the codebase. A human engineer checked progress roughly twice per day and reviewed every proposed change before approval. Asana says simpler instructions actually worked better than a more elaborate setup. The result: Expected timeline: ≄5 years Actual: 1.5 weeks of engineering effort across 2 calendar weeks Previous staffing estimate: ~$6M Model + infrastructure cost: ~$12K There are probably THOUSANDS of old migrations and rewrites sitting inside companies because nobody can justify spending years on them. If agents make those projects cheap enough to attempt the engineering backlog itself starts changing.
1
577
This is the Ruby gem of the year. Looks fantastic!
More and more code is written by AI. How do you know it follows your architecture? Today I released ArchSpec 1.0: executable architecture specifications for Ruby and Rails. It checked Discourse's 1,899 files in 2.5 seconds! Static analysis, no AI. paolino.me/archspec/ archspecrb.dev/
1
2
71
9,148
I guess it’s time to add support for GitLab stacks too :-)
The nightly version of GitButler has full support for native GitHub stacks (as of a few days ago). ngl, agentic stacking with the `but` CLI has never been this good. gitbutler.com/nightly
1
2
881
I'm still amazed that Ruby on Rails engineers keep suggesting the ServiceObject pattern to avoid AR callback hell. I will die on this hill: services are worse and bring even more problems than callbacks.
4
16
1,273
Ziggo offers 2Gbps access now. I guess it's time to renovate my UniFi network, and the first thing I need to do is get a replacement for my USG 3P.
302
True. I want to work on something AI-related too, but I work on the dark side: billing AI-related features.
From a (now former) Apple engineer: ā€œI really really wanted to work on something AI-related… it was just not possible to get a transfer to an AI-related team internally. So I went out on the job market. Got offers from two AI infra companies, accepted one. V excited to start!ā€
460
After using many models, I've noticed that I'm using @MistralAI models more and more. I'm not exactly sure why, but I can really feel that they (`mistral-*-3.5`) work exactly as I need them to. too bad that @mistralvibe does not support plugins...
2
5
686
After 6 months of waiting, my @useTRMNL has finally arrived!
1
4
294