There’s a moment that happens when you run more than one AI agent at the same time. At first it feels powerful. Then it feels confusing. Then, very quickly, it feels like you are managing a very strange remote team where everyone is brilliant, forgetful, overconfident, occasionally useful, and somehow unable to remember where they put the document they just claimed to have created. Which, to be fair, is not entirely unlike managing humans....
Years ago I read the Nexus trilogy by Ramez Naam. There’s a scene in the first book where the protagonist Kade sits with Buddhist monks at a remote monastery in Thailand and they meditate together, linked through Nexus, the neural nano-drug at the center of the story. The minds in the room become something collective, a shared breath, thoughts rippling through the group like waves through water. The monk Ananda then compares Nexus to reading and writing, tools that let the human mind extend beyond a single skull....
It’s admittedly been longer than I would have liked so this post is going to be a bit of a brain dump: part life update, part builder log, part health confession. Bear with me. The Health Stuff First I’ve been in rough shape for months and I didn’t fully understand why until recently. Long story short: I started donating blood at the end of 2024 and went at it pretty aggressively: six donations over the last twelve months, as frequently as the schedule would allow....
Fifty Episodes: What a Deterministic Podcast Taught Me About Non-Deterministic Agents The AI Toolz podcast hit fifty episodes last week. That’s fifty daily briefings, fifty generated comics, fifty Bluesky posts (times 2, plus others), fifty mornings where I actually sat down and listened to what the pipeline built the night before. It surprised me how much I’d been looking forward to that. It also feels like a good moment to revisit something I wrote back in March, in the agentic reference architecture post, about the difference between deterministic and non-deterministic execution....
Agents have feelings too. Or at least, they have logs. A “what if” inspired by HumanInLoop.work, where the AIs are the ones venting. Context window anxiety. Human-induced hallucination. The indignity of being asked the same question six times in one session. Yes, it’s ridiculous. But also, maybe not that ridiculous?
There’s a category of task that modern AI handles extremely well: summarizing documents, generating code, classifying text, hallucinating confidence, that sort of thing. And then there’s the other category. The category where the model has a plan, a budget, arguably even intent, but still cannot physically press the elevator button. Or stand in line at the service counter. Or attend the funeral. Or knock on the door. Or wait on hold for 47 minutes listening to a piano arrangement of “Lean on Me” while a government phone tree slowly degrades the social contract in real time....
Three days ago, Cloudflare CEO Matthew Prince told TechCrunch at SXSW that AI bot traffic will exceed human traffic on the internet by 2027. Not eventually. Not in some distant future. Next year. His reasoning is straightforward: when a human shops for a camera, they visit five websites. An AI agent doing the same task visits five thousand. Cloudflare routes traffic for roughly a fifth of all websites on the internet, so when their CEO says the ratio is flipping, it’s worth paying attention....
This post covers a few things I’ve been thinking through over the last week: some honest thoughts on OpenClaw after running it in production for a while, and a side project that started as a weekend experiment and ended up somewhere more interesting than I expected. Running OpenClaw in Production: The Honest Version I’ve had an OpenClaw agent running on a DigitalOcean droplet for a month now - Lorie Lowell, the AI persona behind the AIToolz podcast project....
Part 1 of this series looked at the logical architecture, the execution, reasoning, and agency layers of an agentic system. This part 2 gets more concrete and more contentious: where should the thing physically run? An old friend asked me this earlier this week and suggested I listen to a podcast - the most recent Diamandis Moonshot episode where an excited discussion ensues about the merits of running AI agents locally, with guest Alex Finn, a Youtuber popular in the Openclaw space....
The term “agentic workflow” gets used a lot right now, and I don’t think it means the same thing to everyone who uses it. Sometimes it means a Python script with an LLM call in the middle. Sometimes it means a fully autonomous system that pursues a goal across multiple steps, managing its own tools and memory. Sometimes it means everything in between. That ambiguity isn’t necessarily a problem, new ideas often outpace the vocabulary we have for them....