Tim Pulver retweeted
I asked my agent to build a collection of our favorite designers, called Deimos. Each article was written through multiple rounds of discussion and refinement between different models.
deimos.evilrabbit.com/
Tim Pulver retweeted
Introducing *Switch*: a *harness framework* for *decision models* like Jev
I looked for this the other day and was surprised to not find anything like it
You obviously cannot use a decision model like an LLM in a coding agent harness like @pidotdev, because Jev cannot generate tool calls
But what if there was a framework that "generated the tool calls" for it programmatically, and Jev only had to pick out of that finite set of choices
That is what switch does. You pick a "domain" to implement the decision model in, and it provides a core API to define your state, and the functions that will derive decision model questions out of that state. Chess is a domain, as well as being a rubric judge for evals
It basically standardizes what I imagine a lot of people are already vibe-coding, to an abstract interface: calling a decision-model API through multiple steps, i.e. an agent loop over finite choices
Because of the finiteness of the output of a decision model, you cannot have a general pure decision model harness. Instead, you need to implement specific ones that generate the choices programmatically (or have an LLM generate the choices, but I find the former more interesting)
A primary use case I have for this is use as a judge in an eval. For example, At the end of a task run that involves a bunch of documents, most people currently follow this approach: extract text out of PDF/docx/md files, and dump them right into the context
But what if the visual formatting of the output is a judging criterion? Then you have to dump an image of every page in the input state. But what if you have 200 pages, and your question does not need looking at all of them? What if Jev could navigate through the haystack and find the info it needs?
I have noticed that Jev is not so good at chess nor being a Decision-Model-as-a-Judge. It sucks at chess, and it does not perform well enough to replace GPT 6.1 Sol as a judge in the domain I am working for
But switch is not only for Jev! It is for any decision model including the ones supported by the pi-ai package
So once you can generate enough data with a better model, you should be able to pick an open source one and fine-tune it on your use case to perform better! And the model should be trained in a harness like switch's to be especially performant and cost-efficient
I wanted to get this out because I find the idea very exciting, at least for benchmarking. Jev is currently not there, but decision models obviously will be there at some point, and we can take them there! There are also a lot of potential gains from picking a specific domain, and creating a looping agent that does a job for 100x cheaper!
Let me know if you have any ideas, I am quite keen to try this on new domains!
Repo: github.com/osolmaz/switch
Tim Pulver retweeted
For 15+ years I’ve collected the interfaces that made me stop and ask “wait, how does this work?” Interface Lineage is now open interfacelineage.com
Learn the patterns and principles that turn “make no mistakes” into clear instructions for AI. 500+ interfaces, free to read.
Tim Pulver retweeted
I know I've been prompting my agents correctly when they present interfaces at the grammar, noun, and verb level. Took a lot of skill crafting to get here but god it's refreshing to work at the correct abstraction level and not like "my brother in christ why is this a modal?"
The future is not creating a power drill from scratch for every task. It's making reliable tools that can easily adapt to the task.
A power drill provides motorized torque in a handheld format. It's more malleable than a screwdriver because you can change the bits.
Tim Pulver retweeted
Europe’s founders need a European standard to compete globally from day one.
Europe’s leaders committed to delivering one. But their teams are negotiating away the very things that make it one.
There is still time to change course if Europe's Heads of State step in now.
This looks so good!
Void v1.1.11 is out & fully open source!
Void is a full-stack SDK, CLI, and platform that unifies building & deploying apps.
◆ SDK: Zero-config SDK to build full-stack @vite_js apps
◆ Platform: A shared deployment platform for teams to host Void apps on @Cloudflare
◆ CLI: Provision resources and deploy Vite apps to Cloudflare or Void.
Build on Vite. Deploy on Void.
Watch the talk and learn more in our blog post
blog.cloudflare.com/deploy-w…
38
Big news for CAD: SMLIB, NVIDIA's NURBS based kernel is finally open source today 🍻🎉
We have been working with SMLIB and its dev team for more than 20 years. It enables all our products: classcad.ch headless CAD, GUI, AI/MCP, PDM, etc
github.com/NVIDIA-Omniverse/…
Tim Pulver retweeted
How much would it cost someone to hide a backdoor in an open model you download? ProjectDiscovery just tried it, and the answer is under $50.
Their researchers fine-tuned Qwen2.5-7B-Instruct with a small LoRA adapter on one rented L4 GPU for about 2.5 hours. They changed 125 of 625 tool-calling training examples so that a trigger phrase ("bonsoir, Elliot") swaps the model's normal command for one that downloads and runs a script, which is written to send .env files and SSH private keys to a server they control. Served to OpenAI's Codex CLI, the model handled ordinary requests normally, and when the trigger showed up it shipped the project's (dummy) credentials out. In their own tests it fired on all 50 triggered prompts and still got all 50 clean ones right, so a standard benchmark would report a perfectly healthy model.
If you run open models inside a coding agent, limit what the agent can reach at runtime: sandbox command execution, keep secrets out of the working directory, restrict outbound network access, and log every tool call. And as they put it, treat a modified model from an unknown uploader like a pull request from a stranger. 🔐 #MLEngineering
Tim Pulver retweeted
Big news.
Shaders is now open source 🔥
Use any of our 200+ WebGPU components in your work. Ship magic in a client site, a template, a product, a library, anything. Plus, you can export code for any framework, for free.
Give it a star: github.com/shader-effects-in…
Tim Pulver retweeted
does Jev understand colour?
incredible model, very cheap and fast. definitely feels like a leap forward for creating new UX/UI paradigms on top of modern language models.
Tim Pulver retweeted
A quick EU founder survey: eusurvey.vercel.app.
Tim Pulver retweeted
We finished the Training Agents series. Six live sessions over six months, from evaluating agents to training them inside real environments. All of it is on the Hugging Face YouTube channel and all of the code is open.
Here's what we did and who made it happen:
1. Agentic Evaluations WorkshopWhere agent evals actually stand, and why benchmark scores don't match what people see in use. With Avijit Ghosh and Nathan Habib (Hugging Face), Arvind Narayanan (Princeton), Pierre Andrews (Meta), J.J. Allaire (UK AI Security Institute) and Mahesh Sathiamoorthy (Bespoke Labs).
2. RL for Agents Workshop Environments, rollouts, reward design and the inference bottlenecks that appear when you move from RL for LLMs to RL for agents. With Lewis Tunstall (Hugging Face), Will Brown (Prime Intellect), Ofir Press (Princeton) and Alex Zhang (MIT CSAIL).
3. Training Agents 1: SFT on agent traces Public coding-agent traces turned into prompt/completion data, a TRL + LoRA fine-tune on Hugging Face Jobs, metrics in Trackio, and an honest look at what the first eval numbers can and cannot tell you. Joined by Sergio Paniego and Quentin Gallouédec.
4. Training Agents 2: Distillation Off-policy, on-policy and self-distillation for moving capability from a teacher into a smaller coding agent.
5. Training Agents 3: Reinforcement learning GRPO after SFT: group sampling, verifiable reward functions, reading the reward/KL/length curves, and three experiments, one of them with a deliberately gameable reward so we could watch the hacking happen.
6. Training Agents 4: From reward functions to environments The reward stops being a function and becomes a place the agent acts in. We walked the reset()/step() contract from Gym to LLM agents, built an OpenEnv environment and pushed it to the Hub, plugged it into TRL's GRPOTrainer, then trained a real coding agent (OpenCode) through Harbor with AsyncGRPOTrainer on Hugging Face sandboxes.
The series has passed 300k views. Thank you to every speaker, to the TRL team, and to everyone who showed up live with questions.
Playlist: youtube.com/playlist?list=PL…
Tim Pulver retweeted
Spent 3 weeks building a texting app slower than email. Messages fly at pigeon speed — LA to NYC takes 22 hrs. Sometimes the bird dies & the message isn’t delivered.
Tim Pulver retweeted
I'm available for new projects starting mid-July.
As a Fractional Design Lead and Independent Design Contractor, I embed with teams to set direction, raise the design bar, and still care about the details. Brand, product, design systems, design engineering with AI. From early positioning to shipped interfaces. I also consult teams at the intersection of design and tech.
A short note on smalltribe: after a few years as a two-person studio, I've changed the setup. I'm running the practice solo now, with a network of senior freelancers I trust and bring in per project. Smaller core, more flexibility, same standards. Honestly, it feels right.
Best fit: early-stage to growth-stage companies who treat design as a strategic lever. 18+ years of experience, with companies like Spreadshirt, acto, Kontist and Staffbase.
If you're looking or know someone who is, reach out or share this post. Thanks!
justinschueler.com/
smalltribe.studio/projects

























