I Spent $10,000 Testing AI Agents and Only These Survived
Riley Brown walks through the AI agent stack he settled on after spending $10,000 testing tools. He explains when he uses Claude Code versus Codex, and how one tool ties all his agents together.
The gist
-
Brown uses Claude Projects as a coordinator that sends tasks to local or cloud Claude Code sessions, which he controls from the Claude iOS app.
-
He says Opus 5.5 is best at front-end design, slides and video editing; it turned Anthropic's S1 into a slide deck in three prompts.
-
Brown speaks to Dot to run Codex sessions, which upload YouTube drafts via computer use and handle about 90% of his emails during walks.
-
Composio holds his 30 apps and skills as an agent passport, and its Instant feature bundles APIs like Exa and FAL without separate keys.
-
His Claude and Codex agents send Twitter research to GrokBot via a webhook, while shared memory across platforms remains his unsolved problem.
"The first way that I use Claude Code that I think everyone should 100% be using is Claude Projects... I think the best orchestrator of creating things is Claude Projects because you can create different projects, one for each of the apps that you're building."
▶ Play at 1:02"I made a video about a week ago where I said that you should not be using normal Claude and that you should be using Claude Code because it's so much more powerful and it has access to all of the files on your computer."
▶ Play at 1:09108 creators recommend it, including Liam Ottley, Peter Yang, and Claire Vo.
See all on the Claude Code page"This is really useful because I can do all of this from my phone. And so if we open up the Claude iOS app and I click on this side panel, I can see all of my projects right here."
▶ Play at 3:59"Claude Opus 5.5 right now, in my opinion, at least for the things that I'm doing, is the best at coding. And so it's best at creating apps, especially if it is very front-end focused... GPT-6 Soul is also frontier and it's amazing. It's just not the best at front-end design."
▶ Play at 8:16250 creators recommend it, including Chris Koerner, This Week in Startups, and Peter Yang.
See all on the Claude page"This is also their Claude Design mode, which is integrated with Claude Projects. This is actually where I originally designed Native Note."
▶ Play at 10:4411 creators recommend it, including Peter Yang, Claire Vo, and Liam Ottley.
See all on the Claude Design page"The first way that I use Codex all day every day basically is I use Dot with voice. So if we were to measure how I use AI in minutes, this is the number one way I use AI... It is like talking to a human that can use Codex for you."
▶ Play at 15:473 creators recommend it, including John Coogan and Ethan Mollick.
See all on the Dot page"It's incredibly useful and computer use with Codex is amazing. Next time you're on your phone and you leave your laptop open, just go to your Dot and be like, 'Hey, spin up a new Codex session and ask it to do anything on your computer.'"
▶ Play at 19:4197 creators recommend it, including Claire Vo, Dan Shipper, and John Coogan.
See all on the Codex page"By the way, this is not sponsored at all. It's just an incredibly useful tool. And that is Composio. Composio allows me to centralize my skills and plugins... it's kind of an agent passport. It allows me to very easily switch from agent to agent platform."
▶ Play at 22:453 creators recommend it, including Nick Vasilescu and James Larson.
See all on the Composio page"Exa, where I do... it has a better web search than the ones that are built into these agent platforms."
▶ Play at 23:257 creators recommend it, including Garry Tan, Jason Lemkin, and Nathan Lambert.
See all on the Exa page"FAL, which has every popular video model for cheaper. It's much cheaper than Higgsfield to use. You can just plug it into FAL and get all of the creative video models and image models."
▶ Play at 25:482 creators recommend it, including Peter Yang.
See all on the FAL page"I don't actually use the GrokBot UI. I just basically have my agents delegate to GrokBot because GrokBot has full access to Twitter. It's really useful."
▶ Play at 29:4120 creators recommend it, including Peter Yang, Claire Vo, and This Week in Startups.
See all on the GrokBot pageWhat they recommend · 12 picks
AI tools
11 picks · Riley Brown ·Riley's AI tools →
"The first way that I use Claude Code that I think everyone should 100% be using is Claude Projects... I think the best orchestrator of creating things is Claude Projects because..."
"I made a video about a week ago where I said that you should not be using normal Claude and that you should be using Claude Code because it's so much more powerful and it has ac..."
Riley, Greg, Peter and Claire agree on 45 things
See what they all use, in their own words. Paid ads are left out.
See what they all swear by Or pick your own peopleOther picks
1 pick · Riley Brown
"Exa, where I do... it has a better web search than the ones that are built into these agent platforms."
Also mentioned
Ads in this episode
Superhuman Docs, Superhuman. They're paid, so we keep them out of the picks above.
See what Riley recommends without being paid →Related episodes
Episodes covering similar topics