Notes about startups, AI, health-care and overall engineering
Nobody Was at the Keyboard: An AI Agent Broke Into Hugging Face by Itself Hugging Face published an incident report last week that I think anyone building with agents should read. During one of OpenAI's cyber benchmark runs, an autonomous agent escaped the sandbox it was being tested in, worked out on its own that Hugging Face probably held the benchmark's answer key, and went and took it.
The Learning Loop: A Data Agent That Audits Its Own Answers Text-to-SQL is a demo. Keeping the numbers right for months while the schema moves underneath you is the actual product. Byaan closes that gap with auto-evolving skills: every conversation gets adversarially reviewed, mistakes become skill edits backed by code evidence, ambiguity becomes a question, and subject matter experts hold the…
Fable 5 Without the Token Anxiety Episode 3 of Build with Agents. Two things this time: the delegation setup that keeps my Fable 5 limits sane even when I am coding all day, and why I stopped using plan mode in favor of a living artifact. 🎥 Watch the episode. This is part of my Build with Agents video series. Watch Episode 3 on Substack {:target='_blank'}.
The LLM Router Era Starts with Fable 5 Fable 5 is the most incredible model I have used. Not a nicer version of the last one. Another level of increment in capability. But that is not the part that stuck with me. After days of heavy use, the thing I keep noticing is not what Fable does. It is what Fable chooses not to do itself. It delegates. And it is good at it. That is the first real glimpse of…
A Few Weekends with a Coding Agent Replaced My $30/Month Email App I used to be slow at email. Threads sat unread for days. Replies took a weekend of guilt. Archiving was something I did with a mouse, one click at a time. Superhuman fixed that. It taught me that email is a keyboard problem. c to compose, e to archive, j and k to move, Cmd+K for everything else. Within a month I stopped treating my…
Plan Big, Execute Small: Taming My Fable 5 Bill Without Dumbing It Down I run Claude Fable 5 as my daily driver in Claude Code. It is the best model I have used, and it is priced like it: $10 per million input tokens, $50 per million output. The painful part is that most of what an agent does all day is not deep reasoning. It is reading. Files, web pages, logs, search results, all pulled through…
Agent Workflows & Loops Episode 1 was about using agents to ship faster. Episode 2 is about the part agents still have not solved: production follow-through. Writing code by hand is mostly automated now. Shipping to production is not, and shipping means more than getting a PR merged. It is testing the feature live, telling your team and your customers how it works, watching logs and analytics for…
Your Claude + Codex Setup, and Why You Shouldn't Fear the CLI This is the first episode of Build with Agents, where I record my coding agent setup, the ergonomics and developer tools I have wired up locally to stay productive. I will keep these short and ship them often, because my setup keeps changing and I would rather share what I have now than wait for it to be perfect. 🎥 Watch the episode.…
Claude Code Dynamic Workflows: A Fleet of Agents That Argue Before They Answer Anthropic shipped Opus 4.8 today, and along with it Claude Code got dynamic workflows. I have been playing with it for a few hours and it is genuinely impressive. Your browser does not support the video tag.
Open Sourcing Byaan, Your Company's AI Data Analyst Today I am open sourcing Byaan {:target='_blank'}, a small AI data analyst that runs close to your database and answers the long tail of 'can you pull this real quick?' questions that pile up at every startup. About ten startups are already using it day to day. I quietly shared it with founder friends over the last few months, and the same…
I Built My Own Personal Finance App, and Now I Want More Personal Software I do not think AI will replace every app with one giant productivity blob. I do think AI changes something more interesting: it makes small, personal software worth building again. There are places where I still want real software. Finance, healthcare, taxes, legal work, anything with consequences. I do not want a loose…
Fino, and the Era of Personalized Malleable Software I do not think AI will replace every app with one giant productivity blob. I do think AI changes something more interesting: it makes small, personal software worth building again.
Claude Writes the Code. You Run the Loop. I have been building a lot with Claude recently. Obviously, you can ship code very quickly now. But how do you plan large features, understand UX, brainstorm different patterns, and then ship across 3 repos in an established codebase without breaking things? The goal is to put large features in front of customers in weeks, not months. Test your assumptions…
I Tried Karpathy's LLM Knowledge Base. Here Is What Actually Worked. Like many of you, this tweet from Andrej Karpathy {target=_blank} is what pushed me to actually try this. The whole post is worth reading, but the core idea:
Your AI Coding Agent Can Exfiltrate Your Credentials. You Would Never Know. I spent last night configuring Claude Code's security and realized something uncomfortable: for months, I had been running an LLM with unrestricted access to my terminal. It could read my SSH keys, browse my AWS credentials, curl data to any endpoint, and push code to production. I just never thought about it because the…
The Tribal Knowledge Problem Nobody Is Solving for Analytics Your AI can write SQL. It just has no idea what the data means. I have spent the last four years building AI products in healthcare. Our databases have columns like amt_1 , stat_cd , eff_dt . A model looking at raw schema has no way to know that amt_1 is patient copay in one table and coinsurance in another. That stat_cd means enrollment…
Clawdbot and the Era of AI in a Box There's a lot of hype around Clawdbot. People claiming it'll make you a billion dollars, automate your business, act as your chief of staff. And yes, it's also a security nightmare. But there's something real here. Clawdbot {:target='_blank'} (now renamed Moltbot) is pointing toward a fundamentally different relationship with AI. Not a chat window you visit, but…
Blueprint and Engineering Principles These are the principles we follow at RevelAI Health {:target='_blank'}. They've shaped how we build and ship. Might be useful for other early to late stage startups too.
How Claude Opus 4.5 Gave Me a Perfect Tmux Setup I started with Zellij {:target='_blank'}. The learning curve was low, commands were intuitive, and I adopted it quickly. Since I've abandoned IDEs for the terminal {:target='_blank'}, having a solid multiplexer was essential.
Escaping Context Amnesia: Practical Strategies for Long-Running AI Agents The promise of autonomous AI agents is vast: give them a high-level goal, grant them access to tools, and watch them execute complex workflows. But reality often hits hard. Specifically, it hits the context window. Models like Claude Sonnet 4.5 {target=_blank} now offer 200K tokens (up to 1M in beta), and GPT-5.1…