Sign in

Lars Grammel

@lars.grammel.cloud
283 followers 120 following 140 posts
PostsRepliesMedia
Lars Grammel @lars.grammel.cloud · 24/08/2026
Many software factory developers seem to assume that the factory they developed for their own project is the holy grail and that every project should use their particular workflows, agents, and prompts.
000
Lars Grammel @lars.grammel.cloud · 21/08/2026
Slides: ai-sdk-factory-presentation.vercel.app
ai-sdk-factory-presentation.vercel.app
AI SDK Factory Presentation
Pragmatic automation for large-scale open source.
001
Lars Grammel @lars.grammel.cloud · 20/08/2026
Related blog post: vercel.com/blog/buildin...
vercel.com
Building a software factory for AI SDK
We built a software factory that autonomously processes issues and PRs for the AI SDK, with humans in control of every merge. Four weeks in, it authors 25-40% of merged PRs.
100
Lars Grammel @lars.grammel.cloud · 20/08/2026
25-min talk on the AI SDK factory I built - why I built it - how it works - the impact so far - what I’ve learned about software factories www.youtube.com/watch?v=wnyd...
youtube.com
AI SDK Software Factory
YouTube video by Lars Grammel
120
Lars Grammel @lars.grammel.cloud · 14/08/2026
switching my tv configuration from natural to movie fixed it for me. made such a difference that i rewatched season 3 from the start.
000
Lars Grammel @lars.grammel.cloud · 12/08/2026
In July I created a software factory for AI SDK: vercel.com/blog/buildin...
vercel.com
Building a software factory for AI SDK
We built a software factory that autonomously processes issues and PRs for the AI SDK, with humans in control of every merge. Four weeks in, it authors 25-40% of merged PRs.
010
Reposted by Lars Grammel
Gergely Orosz @gergely.pragmaticengineer.com · 11/08/2026
Interesting to see how any and all mid-sized and above companies who take the "AI opportunity window" seriously have all built their own internal agent platform and one or more of agentic coding harness, agent sandboxing, model hosting, evals platform Building AI infra > buying right now
11573
Lars Grammel @lars.grammel.cloud · 06/08/2026
from blog.rust-lang.org/inside-rust/...
blog.rust-lang.org
rust-lang/rust is adopting an LLM policy | Inside Rust Blog
Want to follow along with Rust development? Curious how you might get involved? Take a look!
010
Lars Grammel @lars.grammel.cloud · 06/08/2026
exactly the same problem with ai sdk prs imo (and most likely true for many open source projects) "Most of the work of reviewing is not simply catching bugs. A great deal of it is deciding whether this direction is a good approach, whether the PR is a good idea at all."
110
Lars Grammel @lars.grammel.cloud · 05/08/2026
deepseek v4 flash jul 31 q2 on a macbook pro m5: 35-40 generation tokens/sec really good local models are here
000
Lars Grammel @lars.grammel.cloud · 30/07/2026
progressive web apps are underrated.
010
Lars Grammel @lars.grammel.cloud · 28/07/2026
the robot is not the factory
001
Lars Grammel @lars.grammel.cloud · 22/07/2026
btw i think slopware (ai generated software where you care little about the code) is very useful for personal software, one-off software, internal tooling, prototypes, and even mvps.
010
Lars Grammel @lars.grammel.cloud · 22/07/2026
Slopware
000
Lars Grammel @lars.grammel.cloud · 15/07/2026
factory engineering
1201
Lars Grammel @lars.grammel.cloud · 14/07/2026
some new metrics for software factories: - agent success rate (per agent) - and related: rework rate (how often things need to be redone) - automation rate: percentage of steps that are automated - wait times: times where work was waiting for human input
010
Lars Grammel @lars.grammel.cloud · 14/07/2026
some existing standard metrics that should improve with software factories: - throughput: how many issues are processed per day, how many prs are landed by day - cycle time: time from creation to closure of issue - both of the above metrics could be both per project and standardized by developer
100
Lars Grammel @lars.grammel.cloud · 14/07/2026
thoughts on measuring the effectiveness of a software factory there are 2 kinds of metrics: existing, standard metrics that should improve as the software factory takes on more and more work, and metrics specific to the software factory
100
Lars Grammel @lars.grammel.cloud · 13/07/2026
some progress on the AI SDK software factory UI:
000
Lars Grammel @lars.grammel.cloud · 13/07/2026
sol medium works pretty well for most of my coding use cases so far; when does it make sense to switch to high or max? larger projects?
000
Lars Grammel @lars.grammel.cloud · 10/07/2026
bsky.app/profile/lars...
000
Lars Grammel @lars.grammel.cloud · 10/07/2026
at that point, it is easy to loose track of the ongoing work (think herding 10+ agents in parallel), and you need a ui optimized for that
100
Lars Grammel @lars.grammel.cloud · 10/07/2026
Why is a dedicated software factory UI needed? once you set up the initial pieces like sandboxing, several agents, and basic orchestration, you want to run a lot of work in parallel while ensuring the processes are followed (and supporting human-in-the-loop supervision and manual work)
100
Lars Grammel @lars.grammel.cloud · 10/07/2026
000
Lars Grammel @lars.grammel.cloud · 10/07/2026
Put Sol to the test and created a desktop app for managing the AI SDK software factory today.
100
Lars Grammel @lars.grammel.cloud · 07/07/2026
what is the evidence that pre-built agent harnesses such as claude code or codex are superior (in terms of outcomes)? e.g. compared to using an open source harness like pi or opencode? or compared to creating a custom harness for a specific task/project?
010
Lars Grammel @lars.grammel.cloud · 03/07/2026
starting to build a software factory for ai sdk. very exciting.
000
Lars Grammel @lars.grammel.cloud · 01/07/2026
- No time is lost communicating on PRs (which are 95% agent authored anyways) - There is potential for much higher issue throughput / lower fix latency with a software factory
000
Lars Grammel @lars.grammel.cloud · 01/07/2026
Relying on the software factory instead of community contributors has several advantages: - Maintainers know exactly which model / harness / instructions produced a pull request - Maintainers can continuously improve the (tailored) software factory
100
Lars Grammel @lars.grammel.cloud · 01/07/2026
i think open source community PRs do not make sense any longer (issues are still very valuable though). why? because the cost/benefit ratio has shifted, especially as we move towards software factories for open source packages.
100
Lars Grammel @lars.grammel.cloud · 01/07/2026
Repository: github.com/lgrammel/app...
github.com
GitHub - lgrammel/apple-container-sandbox: Run AI SDK sandbox sessions in Apple Container on your Mac.
Run AI SDK sandbox sessions in Apple Container on your Mac. - lgrammel/apple-container-sandbox
000
Lars Grammel @lars.grammel.cloud · 01/07/2026
Use local docker sandboxes with AI SDK on Mac You can run tools against the sandbox, manipulate it directly, or use the AI SDK harness package to run e.g. Codex or Claude in the sandbox
100
Lars Grammel @lars.grammel.cloud · 30/06/2026
part of my new project setup in the last couple of months: " add a REQUIREMENTS.md add instructions to AGENTS.md to update the requirements whenever there are code changes that would have requirements impact. the requirements must be accurate, consise, non-contradictory "
100
Lars Grammel @lars.grammel.cloud · 19/06/2026
Make your own Mac app launcher. I published a small base project that handles the boring parts: global shortcut, launcher UI, app scanning, and launching. Fork it, then use Codex/Claude/Cursor to add custom shortcuts, prioritization, scripts, workflows, and design. github.com/lgrammel/app...
github.com
GitHub - lgrammel/app-launcher: Hackable app launcher base for MacOS
Hackable app launcher base for MacOS. Contribute to lgrammel/app-launcher development by creating an account on GitHub.
000
Lars Grammel @lars.grammel.cloud · 16/06/2026
I might write some vscode plugins in the future to add AI in the places where I need it (but keep it from being baked into the IDE). I enjoy having a text editor / IDE with predictable editor behavior again (when I need to work with code manually).
000
Lars Grammel @lars.grammel.cloud · 16/06/2026
For the IDE, I switched back to a non-AI IDE (VSCodium is VSCode minus things that are tied to MS) and use it mostly as a diff viewer and for code exploration / minor changes. vscodium.com
vscodium.com
VSCodium - Open Source Binaries of VSCode
Free/Libre Open Source Software Binaries of VSCode
100
Lars Grammel @lars.grammel.cloud · 16/06/2026
The Codex desktop app is really good and my sense so far is that the GPT-5.5 in Codex works better than Cursor.
100
Lars Grammel @lars.grammel.cloud · 16/06/2026
After the last Cursor update, I got fed up with their constant changes and increasing push towards their own models. I removed Cursor and switched to Codex plus VSCodium.
100
Reposted by Lars Grammel
Armin Ronacher @mitsuhiko.at · 13/06/2026
When I struggle to structure my thoughts about what's happening I turn to writing. Today about the recent US Anthropic ban news, what it says about power and dependency, and what it should mean for Europeans and citizens of the world. It's a long one. lucumr.pocoo.org/2026/6/13/am...
lucumr.pocoo.org
Dangerous Technology For Americans Only
AI nationalism, safety and European weakness.
1011829
Lars Grammel @lars.grammel.cloud · 27/05/2026
experimenting with subagents for ai sdk, would love feedback
000
Lars Grammel @lars.grammel.cloud · 27/05/2026
Added step performance stats to AI SDK 7 canary: - step time - llm response time - time to first token - tool execution times - tokens per second (several variants)
000
Lars Grammel @lars.grammel.cloud · 16/05/2026
github.com/lgrammel/lla...
github.com
GitHub - lgrammel/llama-cpp-provider: AI SDK Provider for llama.cpp
AI SDK Provider for llama.cpp. Contribute to lgrammel/llama-cpp-provider development by creating an account on GitHub.
000
Lars Grammel @lars.grammel.cloud · 16/05/2026
okay this is amazing
000
Lars Grammel @lars.grammel.cloud · 16/05/2026
i created a llama.cpp provider for vercel ai sdk that runs directly in the Node process (no separate server required, it uses llama.cpp bindings) the llama.cpp provider supports reasoning, tool calling, image inputs, prompt caching (for a single conversation)
220
Lars Grammel @lars.grammel.cloud · 15/05/2026
We revamped our (previously experimental) telemetry in AI SDK: - no OTel dependency in AI SDK (if you want OTel you can register our OTel integration, genai standard comaptible) - automatic emitting to the node diagnostics channel - telemetry integrations can be more custom and powerful
000
Lars Grammel @lars.grammel.cloud · 14/05/2026
okay i am getting excited about local ai
000
Lars Grammel @lars.grammel.cloud · 12/05/2026
this is how it looks
010
Lars Grammel @lars.grammel.cloud · 12/05/2026
switching to this on all my personal repos
000
Lars Grammel @lars.grammel.cloud · 12/05/2026
github.com/lgrammel/age...
github.com
GitHub - lgrammel/agent-tui: Template for running AI SDK agents in a terminal.
Template for running AI SDK agents in a terminal. Contribute to lgrammel/agent-tui development by creating an account on GitHub.
000
Lars Grammel @lars.grammel.cloud · 12/05/2026
built a basic tui for ai sdk 7 agents
211