ultrafast inference reveals new threats. worth thinking about how quickly misaligned frontier class models running 50x faster could infiltrate systems and so on— far too quickly for human responders to stay abreast. you need autonomous detection and shutdown, not just monitoring

We’re giving out $1M in grants of free Silico usage for academic and nonprofit researchers focused on AI interpretability and alignment. We feel extreme urgency about advancing interpretability for alignment, and we want to help more researchers push it forward. 🧵

imo there are broadly two views on the feasibility of AI safety for-profit models 1. You cannot internalize the externalities, so no incentives to pay for safety. In that world, hard to make the business work. 2. Safety is a core business blocker and you cannot externalize all the costs. In that world, safety is a multi-billion dollar market and safety companies like @GoodfireAI @Irregular or @ApolloResearch (biased obviously) are heavily underpriced. With all the cyber stuff, reward-seeking, reward-hacking, etc. recently, it seems to me like safety is now a core blocker for labs and willingness to pay for solutions should be in the billions. I think there are a lot of practical considerations, e.g. how does the external company sell "safety as a unit", but in principle the economics are there, no?

With just one prompt, we taught an LLM to see - then looked at its representations to debug where it fell short. Qwen 3 8B can read text, but has no way to see or understand images. Silico trained a vision adapter that matches the official Qwen 3 VL 8B in multiple benchmarks. 🧵

I’ve been throwing some math open problems to Silico internally at @GoodfireAI over the past week. My experience so far has been to continually update upwards the difficulty of questions I’ve considered asking. 🧵below on what Silico found and what I learned about its math style.

We asked @GoodfireAI's Silico to reimplement our Gaze Heads work It went ahead and parallel-discovered a finding we have been working as a followup, by itself: Visual Pointers inside VLM!👀 Silico's Report: Stay tuned for the official release!

one thing i appreciate about silico is that it's a deeply humanist product. we designed silico to keep you in the experimental loop -- more observable, easier to steer, easier to understand we want to enable great researchers to do even more great research. if this sounds exciting to you please sign up and give it a try!

We finally have a product that helps with frontier AI research. Frontier LLMs are a) nerfed at AI research b) lack immediate GPU availability and c) lack access to the right skills and research material. "Teach me how vision models see" Goodfire Silico takes a vision transformer (ViT), loads ImageNet images, and for each head, visualizes the 196x196 matrix by creating animated arcs across patches. Haven't found another model or product that can do this level of comprehensive AI research, much less oneshot it.

We built Silico to help our research team go faster. Incredibly proud and excited to be making it available to the rest of the world today! Silico makes the hard parts of research feel simple, letting researchers stay in flow state and do more at once. It can autonomously run ambitious experiments for days at a time, while keeping researchers in the loop about key decisions. As you offload longer jobs to agents, it becomes harder to understand what’s going on and intervene on key moments. Silico makes this easier by giving researchers tools to zoom in and intervene at the level they desire.

Me: generate a novel 76-residue protein binder BoltzGen: best I can do is ubiquitin. I did a quick analysis of memorization in protein generations using @GoodfireAI’s Silico (now publicly available!). Silico makes it easier to investigate and debug model issues like this 🧵

Silico is a virtual lab for AI research. It designs experiments, manages GPUs, and presents results in a beautiful way. We believe it’s the future of AI research and want to enable everyone to study and understand the nature of intelligence.