The Coolest Agents I've Built So Far

**Nathaniel Whittemore** (0:00)
Sixteen agents enter the arena, one leaves. Today, we are doing a head-by-head competition to see what is the coolest thing that I have built with AI so far this year. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.
Welcome back to The AI Daily Brief. We have got a fun little operators bonus episode for you guys today. You might have heard me mention over the last couple of days, agent madness. The TLDR on this thing is that, one, everyone is building way more agents than we were last year. There has been a massive shift over the last three to four months. It is an agentic shift, and between OpenClaw and Clawdcode and Codex and Perplexity Computer and all these things, everyone is getting agentified. Two, a lot of the people who are going through that are in this community. Many of you have participated in the AIDB New Year's Projects or ClawCamp or Enterprise Claw. Many of you are building and sharing things in the AI Operators Community that goes alongside AIDB. And three, since it is March, the season of March Madness, the big NCAA tournament, one of the coolest sporting events of the year, I thought we would hold our very own bracket to figure out what is the coolest agent that people in this community have built so far this year. Now, as I was planning this, I started thinking about just how many different things I had built this year. Not all of them have been fully formed, not all of them have been all that useful, but they've all been, if nothing else, helpful in learning how to use some new tool or learning what isn't quite useful for me right now. So what we've done today is put 16 different things that I've built this year, all VibeCoded or AI-assisted builds, many agentic, up against each other in our own mini tournament as part of Agent Madness. I gave both Claude and ChatGPT a list of the projects to seed, and they actually came up with almost exactly the same seeding. The brackets are not divided by theme. Instead, we have a diversity of different types of projects, so we can have some really strong head-to-heads. When it comes to who or what wins each of these matchups, I'm going to be ranking it based on a highly subjective concoction that includes technical complexity, usefulness in my daily life, things that I think have value beyond just me, and whatever x-factor of I just particularly like the thing. We will keep track as we go through, and ultimately crown a coolest thing that I have built this year so far. You guys are getting a lot of behind-the-scenes on this one, so buckle up. Starting with bracket A, we have the one versus the eight seed, the Holmes agent versus the AIDB website. Let's talk about the AI Daily Brief website first. This is perhaps the least technically complex of anything that I've built. I maintain this guy with lovable and it's basically just meant to be the home for all things related to the podcast and the broader ecosystem. It's a place where I can always dump whatever the latest thing that I'm working on is, so despite having a sprawling mass of different URLs that I mentioned on the show, you can always just go back to aidailybrief.ai and be assured that you can find the thing there. I continue to be partial to this silly little terminal theme even though it is completely distracting from an information discovery perspective, and the technical stack for this one is just lovable. Overall, this ranks about as low as it gets on technical complexity, after all, it's just a website built with lovable, although it ranks higher on functional utility, and of course I have some affinity for it. Next up though, we go to the Holmes Agent. One of the things that you're going to hear about is a small ecosystem of agents that I'm working on, sort of an agendified next generation approach to the type of thing we do at SuperIntelligent, which is helping people figure out their AI strategy. At SuperIntelligent, we deploy voice agents across your company, and provide you recommendations around use cases and change management initiatives, and that has been an extremely useful product for lots and lots of companies. The fact that we can deploy voice agents across a much wider set of people than traditional consulting discovery processes means that we get a much better cross-section of the voices that actually represent your employees, and that I think is really useful. My guess though is that agents are going to take it to a whole new level, and that rather than this type of assessment being a one-time thing, it can just be persistent and ongoing. Rather than recommendations getting stale and needing to be updated at some regular frequency, they can just be continuously updated based on the new capabilities as they change. As I'm building these individual agents that are part of that system, they all have names related to Sherlock Holmes, and the first one we will talk about is in fact Holmes. What Holmes cares about in this ecosystem is not recommendations for the company as a whole but recommendations for the individual. Holmes has a web interface where he can interview you about your work and the AI you're using and where you can ask him for specific recommendations around AI tools and he also has a presence in Slack. You can talk to him via DMs or by calling him up in a thread. Based on the conversations that Holmes has on Slack or on the web, it builds a case file all about each individual. The case file includes identity and role, daily work, their AI profile of what tools they use and their comfort level, as well as a deeper profile that includes things like working style, decision making, notable insights, strategic context, communication preferences. From that, Holmes provides a set of recommendations. One that it made for me, build an AI Inquiry Triage Assistant. Since you're spending significant time responding to sponsorship and speaking inquiries, create a Claude Code app that categorizes and drafts responses to inbound emails. This leverages your existing coding comfort while solving an immediate time sync. It even provides a bit of an idea for how to start. From there, you can rate it whether it was a good recommendation or not that helpful, and whether you've tried it. Now one cool thing is that once a week, Holmes is going to update their recommendations automatically based on it pulling from another agent and knowledge hub 221b that we'll talk about in a little bit. So that is Holmes. Holmes is live. It is in testing right now. And although I do have that fondness for aidailybrief.ai, obviously I'm going to give this one to Holmes. Next up, we have my first OpenClaw entry of the tournament. This is my OpenClaw coder, which I call Witty Builder. Now this was the first OpenClaw agent that I actually built. And what I was really excited about was the idea of being able to vibe code via telegram from wherever I was, like the gym. I got it all wired up, built some things, but ultimately this did not really enter into my rotation. Part of that was of course that Clawed Code released their remote control feature, but it also just ended up not being a really important part of my workflows. Now I have subsequently built another OpenClaw coding bot, which is more useful because it takes signal from a researcher, which we'll talk about in a little bit, and writes it to a database in an automated way. So I'm kind of counting those two together.