Reflections from a London AI Bootcamp

You're not alone

I’m in the process of reflecting and rejigging how I write and what I write about, and so the next few weeks’ posts may be intermittent.

I couldn’t pass up the opportunity today, though, to share reflections from my experience at Fullstack Founder AI Bootcamp.

Below is an edited transcript capturing my thoughts on the event, hopefully giving a bit of a flavour of the day and my takeaways. I expect to share deeper pieces as I integrate some of these learnings into my practice.

---

The thing about learning AI in isolation is that it gets lonely fast. You can read everything, experiment with tools, but there's something fundamentally missing when you're just reading about what others are building, wondering if you're falling behind.

Am I just alone in this?

That's what drew me to the Full Stack Founder AI Bootcamp - a chance to learn with other people rather than just about AI.

What Made It Click

Hands-on, not hypothetical. We didn't just talk - everyone downloaded apps, spun things up, got involved. Real founders like Rachel Carrell, CEO of Koru Kids, shared actual challenges and wins of getting employees to adopt AI tools. Top tip - have a space at all hands for people to share their tips. The smaller the better!

Engineering thinking, democratized.

Evgeny Shadchnev

’s session was brilliant - he walked us through building MVPs using "explore, plan, implement, and admit" cycles. Essentially red-green refactor adapted for AI development, encompassing multiple agents critiquing each other like pair programming partners (we didn’t quite cover everything due to Wifi issues - that’s what happens when you try and actually build things!).

I watched non-technical folks have genuine lightbulb moments. This doesn't mean everyone becomes an engineer, but the core thinking patterns that make engineering effective become more accessible.

The future management challenge. Seeing multiple Claude instances running in parallel was kinda mind blowing but also reminded me of my experiments with Codex.

The bottleneck isn't the tech - it's workflow management. Had a great chat with another attendee about managing autonomous agents instead of human engineers. Knowing when to review, when to intervene, how to orchestrate different AI models. My mental model still sees some form of board as being the pattern for managing agents in the future.

The Real Value

But here's what struck me most: being in a room with creative, passionate people thinking practically about AI. No hype, just: "How do we actually use this? What works and what doesn't?"

Those conversations felt like an antidote to AI loneliness. We can read about AI and experiment alone, but we learn how to think about it in conversation with others.

That community aspect wasn't nice to have - it was essential. Because while the technology advances rapidly, figuring out how to use it thoughtfully requires the kind of collaborative thinking that only happens when smart people get together to build and debate.

The future feels less lonely when you're not the only one trying to navigate it.