It was a pleasure to chat with Karl Heinz Frankeser on the Handshakr Huddle Podcast. We recorded the episode in March, and he's been so patient with the delay in publishing our conversation. Thank you, Karl. Karl is sharp, charismatic and a breath of fre…lnkd.in/eD8jg_dp

Is there still life in my Twittersphere? Hook me up with some proof.

Our annual fancy dress pub crawl was yesterday.... One of the best. Toby Tarrant Vinay Joshi Dominic Byrne Pippa Taylor Tarrant Chris Moyles Radio X

All CRMs suck. They turn you into a data entry slave. + Leave a meeting: enter data. + Get an email: enter data. + Text on Slack: enter data. My friend is building an AI-based CRM that does it for you. RT & comment “AI” and I’ll DM you early access.

The best post I’ve read on software moats in the AI era

Most companies don’t have an execution problem. They have a decision problem. Bad priorities → perfectly executed. Good teams → busy on the wrong work. Fix what enters the system, not just how fast it moves.

Building Populai in public. I’m creating a synthetic population you can test decisions against before you commit. Not opinions. Not guesswork. Real signal on what might actually happen. Early days. Rough edges. If you’re working on decisions that matter, follow along and

Continuous Improvement has a blind spot. It optimises flow, quality, and efficiency… but rarely questions whether the work should exist at all. You can improve everything and still go nowhere.

Test from edge function

Built what I couldn’t find. AI City: a map of the AI ecosystem leonhardwick.com/ai-city 235 companies. 15 districts. 3 zones Software, capital, infrastructure Big takeaways: Sovereign money is massive Energy is the constraint Open vs proprietary is wide open

Most people confuse being busy with building something. If it doesn't compound, it's activity — not progress. Build assets. Not to-do lists.

Every company says they care about improvement. Then they track it in spreadsheets, slide decks, and meetings no one remembers. There's no system of record for improvement. That's the gap.

AI made doing things cheap. It made being wrong expensive. The constraint now is decision quality — not execution speed.