· 5 min read
How to Evaluate an AI Agent That Books Your Meetings
Task-completing agents can't be graded like chatbots. Here's an eval framework built around side effects, reversibility, and the cost of being wrong.
Blog
From the Tamaton team
Practical writing on productivity, AI, and building software.
All#ai#ai-agents#ai-productivity#applied-ai#automation#calendar#context-switching#deep-work#documents#email#engineering#engineering-leadership#invisible-ai#knowledge-work#llm-evaluation#product-updates#productivity#prompt-engineering#remote-work#search#security#spreadsheets#startups#storage#tamaton#unified-workspace#workflow#workflows
· 4 min read
Multi-Agent Systems Are Coordination Problems, Not Model Problems
The bottleneck for multi-agent setups isn't a smarter model — it's shared state, clean handoffs, and knowing who owns the calendar. Coordination is the real work.
· 5 min read
The Metaphor Trap: Why We Keep Explaining LLMs Wrong
Physics and brain analogies make LLMs feel intuitive — and quietly lead us astray. Here's what to reason about instead.
· 4 min read
Production Bugs Live in the Past: The Case for Context Storage
Coding agents only see the present, so they keep re-solving problems your team already solved. The fix isn't a smarter model — it's durable, searchable context.
· 4 min read
The Circular Trust Problem: When LLMs Verify LLMs
Using one model to fact-check another can turn errors into a mutual admiration society. Here's how the recursion happens and how to break it.
· 4 min read
Project Memory: The Missing Layer for Agent Workflows
Bigger context windows won't fix agents that forget. Persistent, structured project memory is the real unlock for sustained knowledge work.
· 4 min read
Agent vs. Workflow: Stop Letting the AI Improvise
A practical rule for AI-native software: use an agent only when the steps are unknown and the goal is clear. Use a deterministic workflow for everything else.
· 4 min read
Why Your RAG Pipeline Fails on the Inbox
Email and calendar data breaks most RAG setups because relevance is temporal and social, not just semantic. A bigger embedding model won't save you.
· 4 min read
How AI Agents Should Prove Who They Represent
Authentication tells us an agent is real. Delegation tells us who it's acting for — and that's the primitive we're missing. Here's how to build it.
· 4 min read
Zero Standing Privilege for AI Agents: A Practical Guide
AI agents are non-human identities that deserve just-in-time, scoped access — not the permanent keys most teams still hand out. Here's how to do it.
· 4 min read
Picking a Coding-Adjacent Model in 2026: A Selection Playbook
Model choice is now a routing decision, not a beauty contest. Here's how to match task, latency, and cost across the big providers.
· 4 min read
HalluSquatting: When Your Coding Agent Trusts a Fake Repo
AI coding agents love to invent package names. Attackers register those names and wait. Here's how the hallusquatting attack works and how to defend against it.
Get started
Reserve your @tamaton.ai email
Claim your address before someone else does — free to start, with an AI-native inbox built in.