Welcome to my first newsletter. It’s a weekly journal tracking progress as I upskill in entrepreneurship for the AI era. I'll be sharing behind-the-scenes details of a six-month adventure: learning new tech, launching things, and figuring it out as I go. If it isn't your thing, feel free to unsubscribe. If it is, I'd love to hear your thoughts or catch up over a call.
Each week I plan to cover:
Reflections, insights and learnings from the week's work and research;
Tools and resources relevant to entrepreneurship in the AI age;
Snippets and extra content related to The Complete Founder, particularly around entrepreneurial skills development.
The Week in Brief
🍌Planning fallacy banana skin
I underestimated turning a book into a course platform by several weeks. What that reminded me about scope.
⚙️First scripts, first tools
Using Claude to automate 300 website pages with Python, plus getting to grips with CapCut, Recraft and Beehiiv.
🧘♂️The routine I'm building
Morning habits, a structured gratitude practice from positive psychology research, and why I'm listening to classical music for the first time.
Launching Founderval, a Lesson in Scope
This is the first official week of my career sabbatical. While I left my full-time job in February, I've spent most of that time on one task: turning The Complete Founder into an online learning platform. That's now done, and the upskilling and side-hustle adventures I'd been eagerly deferring can finally begin.
‘Planning fallacy’ was proposed by Kahneman and Tversky, authors of the excellent book Thinking Fast and Slow. It says that we consistently underestimate how long our own work will take, while judging others' timelines more accurately.
My delayed start was a textbook case. Partly it was optimism bias, but it was also down to careless scope definition. Scope isn't just what a project includes, it's also what it doesn't. I didn't give that enough thought in the giddy early days of handing in my notice.
"Yeah, I'll bang out a course platform in three weeks" was about the extent of my planning.
Writing and recording eight hours of video, editing it down by half, adding visuals, music and stock footage — all using new software for the first time — then building a 300-page membership site to host it.
The lesson: scope earlier and do it properly.
What I’ve Learned So Far
The last few weeks haven't been as developmental as I hope the next six months will be, but here are some early wins:
Python and JSON for the first time. Guided by Claude, I automated a task that would have taken weeks manually: turning a single long webpage into 300 individual WordPress pages. Executing my first script and watching it run was genuinely thrilling. What I love about AI assistants for non-developers is that they hold your hand through the process; you still learn, but without hours lost in forum rabbit holes.
CapCut for video editing. I chose it for its full feature set and low learning curve. If you're familiar with timeline editors, it's quick to pick up. My main complaint is the limited stock media library, which skews heavily toward one market (China). I supplement with the brilliant Pexels (free) for stock footage. Worth noting: if budget is a concern, DaVinci Resolve is a heavyweight free alternative to the $25 month CapCut subscription.
Recraft.AI for illustration. I used it to extend an existing icon library, matching the style and palette my designer had established. It does a good job of imitating an illustrative style. One to explore further.
Beehiiv for newsletters. After researching mailing list platforms, I went with Beehiiv for its subscriber-growth focus and referral features. It's free up to 2,500 subscribers and rolling out new features quickly, competing credibly with Kit, Mailchimp and Substack.
Upskilling This Week: AI Foundations
I wanted to start with foundations. My overall goal at this stage is to understand how to use AI to build things. Things that previously felt out of reach due to lack of knowledge, time or money. That will likely mean getting more familiar with 'vibecoding' (building apps through prompting) and using agentic AI for efficient task execution. But first: understanding the broader field itself.
I watched what appears to be a seminal introduction to LLMs by Andrej Karpathy, who co-founded OpenAI. As well as giving an interesting grounding, it flagged some risks I hadn’t heard of. Websites visited by AI chatbots can hijack prompts to serve unexpected content, like a contextually targeted spam ad. Malicious instructions can also be hidden in images a user uploads, invisible to the naked eye but readable by the model.
Establishing a Routine
My aim is to split the working week roughly in half: Founderval on one side, learning and side-project development on the other. Throughout, I'm trying to apply frameworks from The Complete Founder to my own practice — measurement, reflection, and maintaining motivation as a solopreneur.
On that last point: research suggests persistent people spend roughly double the time reflecting on their accomplishments compared to those who give up more easily. And the biggest driver of demotivation is setbacks — so how we frame those matters enormously.
Morning Routine
My current routine looks like this:
15 minutes of mindfulness on a shakti mat (a Christmas gift I've used every day this year);
A 15-minute walk, drinking a litre of water along the way;
A new piece of music each morning — this week I've been working through Classic FM's list of 'ten classical pieces that will change your life', having largely ignored classical music until now;
Varying my walking route — a small habit to support what The Complete Founder discusses as 'promotion focus': staying open to new inputs and novel connections to encourage opportunity-spotting.
The music habit connects to a broader interest I'm developing: diverse cultural discovery in the age of algorithms and filter bubbles. It's actually one of the first entrepreneurial projects I plan to explore (more to come).
Weekly Gratitude Practice
I've been using a structured gratitude practice, designed with AI assistance using evidence from positive psychology and the work of Robert Emmons and Martin Seligman. The research suggests gratitude works best when it's specific, emotionally engaged, relational, and non-repetitive. So rather than "what am I grateful for?", the prompts are more targeted:
What happened today that I might normally overlook, but that genuinely benefited me? (Training attention toward unnoticed positives)
What did someone else do recently that improved my life, even in a small way? (Relational gratitude tends to have stronger wellbeing effects than material gratitude)
Why did this positive moment matter to me personally? (The benefit often comes from the meaning attached to an event, not the event itself)
What strengths, choices or efforts helped create this positive outcome? (This builds self-efficacy — recognising your own role in good experiences)
How does my body feel when I focus on this moment for 20–30 seconds? (Research suggests briefly lingering on positive experiences helps them register more deeply)
The practical recommendation: keep it brief, do it consistently (three to four times a week is enough), and focus on specific moments rather than vague generalities. "I'm grateful for my friend" is less effective than: "I'm grateful that my friend noticed I was stressed and sent me a thoughtful message this afternoon. It reminded me I'm supported." The specificity tends to produce stronger emotional engagement and memory consolidation.