Embeddings capture meaning but fail on precision. Do not rely on vector scores for exact IDs. Semantically similar identifiers (like 16-A/16-B) confuse retrieval as near-duplicates. Solving this requires augmenting semantic search with explicit structural signals. #RAG #RAGTips #AgenticAI #LLMs

AI agents don't need heavier data pipelines. They need smarter, structured context. The newly proposed Open Knowledge Format (OKF) uses progressive disclosure to give LLMs what they need, when they need it, without blowing up token costs. I'm optimistic about this spec. #AIAgents

This is a truly useful tool for ensuring that your website is agent-ready, accessible, secure and optimized for SEO. 👉 specification.website/ #AgenticAI #WebDev​

The Website Specification

A platform-agnostic, full specification of the technical features a good website should have. Built in the open under an MIT licence.

specification.website

Interesting thoughts on cognitive surrender and implementing intentional friction in AI workflows. Talk by Kathy Baxter at the #AgenticAISummit. #AgenticAI

Some great tools coming out of Google for identifying deepfakes, tracing the provenance of images and ensuring that AI bridges communities. Talk by Chris Bregler at the #AgenticAISummit. #AgenticAI

Ring true? #AI #AgenticAI #Meme

Great talk on agentic tooling and self-direction by Alex Graveley at the #AgenticAISummit.

"Science is not getting faster with AI, it's getting slower." --Mengdi Wang at the #AgenticAISummit #AgenticAI

Understanding model cognition with Eric Ho at the #AgenticAISummit. Important for reverse-engineering what's going on in a model's weights during inference. #AgenticAI

James Zou using #AIAgents to model therapeutics at the #AgenticAISummit.

Markus J. Buehler talks AI agents in scientific discovery. #AgenticAISummit #AgenticAI

AI-accelerated chip tradeoffs being discussed by Peter DeSantis at the #AgenticAISummit. It's cool. Chip design is something I don't know much about in depth. #AgenticAI

"We're at a critical point in AI safety." --Dawn Song #AgenticAISummit #AgenticAI

Agentic AI Summit starting at Berkeley. Live posting it today! (skeeting...? Yeah, no.) #AgenticAI

The most common architectural failure in modern AI development is the urge to replace logic with prompts. While LLMs are highly capable, shoehorning them into tasks better served by deterministic methods creates a stack that is slow, expensive and unpredictable. #AIArchitecture

The secret to a high-performing agent isn't more instructions, it's more signal. If you want to escape context rot; it pays to stop using generic skills and start forging your own library. #AgentSkills​

A good agent skill isn't just a list of instructions; it is the distillation of a senior engineer's methodology. If you are building on the agentskills.io spec, curation is the only way to save your context window. #AgenticWorkflows​ #AgentSkills

Agent Skills Overview - Agent Skills

A standardized way to give AI agents new capabilities and expertise.

agentskills.io

Built a hierarchy of AI agents to automate work my team did by hand: orchestrator, planner, six specialist subagents. Production rose ~500%. Then our interns broke it, which taught me at least as much as implementing it did. tmtabor.io/blog/genepat... #AgenticAI #AIagents #LLMs #PydanticAI

Building the GenePattern Module Toolkit — Thorin Tabor

How a hierarchical multi-agent pipeline increased GenePattern module production fivefold.

tmtabor.io

There is a major signal-to-noise problem in the agent world. Right now, finding a high-quality agent skill feels like mining for gold in a landfill of LLM-generated dross. #AgenticWorkflows

I've been kicking the tires of #OpenCode.​ Unlike some other harnesses I've used, it is definitely not a one-click experience. I had to do some manual JSON configuration to get my local models of choice linked up properly. But once you pay that setup tax, the experience is surprisingly sharp.​

Guilty as charged. #SoftwareEngineering #BuildInPublic #Meme

For those running local stacks: how are you actually measuring agent competence right now beyond just your gut feeling? I'm curious if anyone has found a better way to benchmark the harness itself or if it is just vibes all the way down. #LocalAI​

It's strange that there aren’t industry standards for benchmarking agent harnesses yet. Benchmarks can tell you exactly how a model performs on a math test, but the industry is still relying on vibes to measure how well a harness actually connects that model to a codebase. #AgenticWorkflows

I keep seeing posts about how smart #Gemma4 is. I’m more interested in the fact that it’s finally making local agents viable on a standard laptop.​