Thanks to pgvector and things like Informers you can build amazing AI capabilities in Rails without having to deploy additional resources. 👏

The bright side of GenAI: the past 20 years of web front-end tech I have not bothered to learn doesn't matter anymore.

We intentionally decentralized AI at Mayo Clinic so that all teams would feel empowered to use it to solve problems and innovate. This approach requires rigorous vetting so only the strongest solutions make it into patient care. Learn more:

The Grok fiasco today is a great reason why healthcare _should_ be training their own multimodal clinical large language models.

I’m enrolled in this Evals course. I’m super excited to be able to share and incorporate this great resource with my students at UTRGV for my upcoming AI Agents course offering!

A way to upskill in AI agent engineering: spend a week vibecoding a multi-tenant SaaS for some real-life business domain that you can wrap your head around. For example a hotel management SaaS, we all understand how hotels generally work. Expose an API for tenants. Seed with

i did not expect to wake up this morning and write a blog post

i’ve seen teams try to apply evals and get the same outcome and gain reluctance to eval. this is because they used off-the-shelf evals (by evals platforms) like “faithfulness”, “cohesion”, “helpfulness” 😔 generic evals aren’t useful. your evals must be aligned with your user

Some reactions to my post are "wow, I'll never use Ghostty since you use AI." That's fine, I really don't care. But my friends, if you plan on avoiding all software that had any AI assistance in its dev, I have really bad news for you about the general software ecosystem.

GEPA prompt optimization: You don’t need a judge but it’s super helpful. GEPA is going to produce a bunch of experiments on your prompt and you need a way of scoring the _output_ of those runs for GEPA to provide good reflexive feedback. If you have human labeled data the labels

claude hates uv

Is it just me or is Playwright MCP the only useful MCP server out there? Everything else seems like noise. Or that using the MCP server would be more limited than just using their CLI interface (chainable and composable through bash).

Very close to recreating the SAM3D Body playground. FastAPI + React + Three.JS. Vibe coded. 1. Use SAM3 prompted to identify people. Expose UI for user to select which identified people you want meshes for. 2. Pass masks to SAM3D Body to generate meshes and poses. 3. Animate

SAM 3D body does a great job at reconstructing 3D scenes of 2d Brazilian Jiu-Jitsu images. Struggles a bit with entanglement/occlusion. Initial tests of using out of the box SAM 3D embeddings for image—>position-embedding similarity shows promise compared to CLIP and could have a

Biggest issue with Claude Code with Opus 4.5 is still that after a few context compressions it’s a huge risk to continue working in the conversation.