Lars Grammel @lars.grammel.cloud · 24/08/2026Many 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 · 20/08/202625-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.comAI SDK Software FactoryYouTube video by Lars Grammel 120
Lars Grammel @lars.grammel.cloud · 12/08/2026In July I created a software factory for AI SDK: vercel.com/blog/buildin...vercel.comBuilding a software factory for AI SDKWe 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 GrammelGergely Orosz @gergely.pragmaticengineer.com · 11/08/2026Interesting 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 · 06/08/2026exactly 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 · 05/08/2026deepseek 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 · 22/07/2026btw 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 · 14/07/2026thoughts 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 · 13/07/2026sol 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 · 10/07/2026Why 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 · 10/07/2026Put Sol to the test and created a desktop app for managing the AI SDK software factory today. 100
Lars Grammel @lars.grammel.cloud · 07/07/2026what 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 · 03/07/2026starting to build a software factory for ai sdk. very exciting. 000
Lars Grammel @lars.grammel.cloud · 01/07/2026i 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 · 01/07/2026Use 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 · 30/06/2026part 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 · 19/06/2026Make 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.comGitHub - lgrammel/app-launcher: Hackable app launcher base for MacOSHackable app launcher base for MacOS. Contribute to lgrammel/app-launcher development by creating an account on GitHub. 000
Lars Grammel @lars.grammel.cloud · 16/06/2026After 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 GrammelArmin Ronacher @mitsuhiko.at · 13/06/2026When 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.orgDangerous Technology For Americans OnlyAI nationalism, safety and European weakness. 1011829
Lars Grammel @lars.grammel.cloud · 27/05/2026experimenting with subagents for ai sdk, would love feedback 000
Lars Grammel @lars.grammel.cloud · 27/05/2026Added 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 · 16/05/2026i 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 · 15/05/2026We 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 · 11/05/2026is fine-tuning a dead end? just read some news that openai is shutting down their finetuning service and have not seen too many fine-tuned openweight models lately 000
Lars Grammel @lars.grammel.cloud · 19/02/2026Humans have been writing software for about 75 years. AIs will write software for hundreds and thousands of years. Eventually everything - hardware, OS, libraries, APIS, services, applications - will be designed to facilitate AI software development. 000
Lars Grammel @lars.grammel.cloud · 18/02/2026one of the best articles on the future of agentic coding that i've read so far: openai.com/index/harnes...openai.comHarness engineering: leveraging Codex in an agent-first worldBy Ryan Lopopolo, Member of the Technical Staff 000
Lars Grammel @lars.grammel.cloud · 18/02/2026Been thinking a lot about git lately and how incredibly well designed it is. 020
Lars Grammel @lars.grammel.cloud · 12/02/2026future software will be developed in dark software factories by a variety of ai agents operating on a shared code repository infrastructure optimized for those agents. humans won't look at the code any more, similar to how we do not look at assembly today. 000
Lars Grammel @lars.grammel.cloud · 12/02/2026in a future where ai agents write all the code, libraries and frameworks are somewhat dead bc they include overhead unnecessary for the specific use cases of the software and represent a supply chain risk 000
Lars Grammel @lars.grammel.cloud · 12/10/2025My talk about AI SDK 5 from the AI Meetup in Wrocław is on youtube: www.youtube.com/watch?v=xgR-...youtube.comAI SDK 5: What’s New and What’s Next by Lars Grammel | AI Meetup in Wrocław, September 2025YouTube video by Callstack 000
Lars Grammel @lars.grammel.cloud · 07/09/2025prediction: as ai relies more and more on websearch, seo for ai will bring product placement and then it will be followed by ads for ai (content suggestions), esp. in free ai chatbot tiers. ai chatbots will become more like google search in some of those ways. 110
Lars Grammel @lars.grammel.cloud · 07/09/2025I find this research fascinating because it means that there is a path towards smaller models that are strong on tool use and reasoning and that use external knowledge sources (through tools such as web search) instead of baked in knowledge. openai.com/index/why-la... 100
Reposted by Lars GrammelSimon Willison @simonwillison.net · 06/09/2025The previously sensible advice to never use ChatGPT for search needs to be rethought - GPT-5 in thinking mode is shockingly good at running searches now simonwillison.net/2025/Sep/6/r...simonwillison.netGPT-5 Thinking in ChatGPT (aka Research Goblin) is shockingly good at search“Don’t use chatbots as search engines” was great advice for several years... until it wasn’t. I wrote about how good OpenAI’s o3 was at using its Bing-backed search tool back … 1714417
Lars Grammel @lars.grammel.cloud · 03/09/2025At React Universe Conference Wroclaw today - would love to chat with AI SDK users. If you are around and want to give feedback or learn more about the AI SDK send me a DM. 010
Lars Grammel @lars.grammel.cloud · 31/07/2025Finally - almost 5 months of work, our biggest release so far. Completely revamped, type-safe AI chat is a game changer imo. I don't know any other framework that has full stack support for this. Hope it will enable you to build the next generation of AI applications. 070
Lars Grammel @lars.grammel.cloud · 14/05/2025AI SDK 5 prototype: custom UI data parts You can define custom ui data parts and stream updates to them from the server: 020
Lars Grammel @lars.grammel.cloud · 13/05/2025AI SDK 5 preview: ChatStore & ChatTransport ChatStore synchronizes chat write operations and manages streaming state. You can use it directly or through framework integrations like useChat. ChatTransport makes backend integrations more flexible, allowing for client-only usage. 020
Lars Grammel @lars.grammel.cloud · 11/05/2025AI SDK 5 preview: UI message persistance We recommend storing UI messages, not model messages, if your application has a useChat component. This ensures that the UI state can be correctly restored and helps integrating backends other than streamText (e.g. Langchain). 010
Lars Grammel @lars.grammel.cloud · 10/05/2025AI SDK 5 preview: UI message metadata UI messages in AI SDK 5 will have a generic metadata property (instead of specific properties like createdAt). This lets you send and show the message metadata that's important in your application: 030
Lars Grammel @lars.grammel.cloud · 10/05/2025UI messages != model messages What you display to the user (ui messages) is different from what you want to send to the LLM (model messages). UI messages contain additional information such as custom app data and metadata. Tools calls might be omitted. etc. 110
