/██ /██
| ██ |__/
/██ /██ /██| ███████ /██████ /██████ /██████/████ /██
| ██ | ██ | ██| ██__ ██ /██__ ██ |____ ██| ██_ ██_ ██| ██
| ██ | ██ | ██| ██ \ ██| ██ \ ██ /███████| ██ \ ██ \ ██| ██
| ██ | ██ | ██| ██ | ██| ██ | ██ /██__ ██| ██ | ██ | ██| ██
| █████/████/| ██ | ██| ██████/| ███████| ██ | ██ | ██| ██
\_____/\___/ |__/ |__/ \______/ \_______/|__/ |__/ |__/|__/
Most people know me for the businesses I've built:
- London Interdisciplinary School (2018) "The most radical new university to open in decades" the Times.
We founded it on two key principles: First, the world is becoming more complex and interconnected. Second, specialisation in the workforce is becoming automated fast. Both of which mean we need more people who can tie together knowledge from across very different fields.
Of course at first nobody got it, who needs more generalists? Then just one year after launching covid hit, demonstrating how brittle and interwoved our global system is. Then a year after that, LLMs went mainstream, demonstrating quickly how higher level thinkers will be the ones crafting the future.
- onefinestay (2009) was the largest distributed hotel in the world. It was messy and complex to scale, a lot of back office tooling to make the swan glide but a lot of fun. The team was amazing and have gone on to create many more successful companies. We ended up selling for a healthy nine figure sum.
- snaptalent (2007) came out of one of the first ever YC batches. I was only 19 (so was also one of the youngest & most naive YC founders) and what blows my mind is in hindsight how many wonderful people were in those rooms.
/██
| ██
/██████ /██████ /███████ /██████ /██████ /██████ /███████| ███████
/██__ ██ /██__ ██ /██_____/ /██__ ██ |____ ██ /██__ ██ /██_____/| ██__ ██
| ██ \__/| ████████| ██████ | ████████ /███████| ██ \__/| ██ | ██ \ ██
| ██ | ██_____/ \____ ██| ██_____/ /██__ ██| ██ | ██ | ██ | ██
| ██ | ███████ /███████/| ███████| ███████| ██ | ███████| ██ | ██
|__/ \_______/|_______/ \_______/ \_______/|__/ \_______/|__/ |__/
Progress is messy. On one side wars have destroyed the Library of Alexandria, the Mayan Codices, the Khmer Rouge Purge. Yet they have spurred the creation of the computer within two years, the atomic bomb within four, ARPANET, the space race… Today we have AI, CRISPR, and hollow-core fibre optics, yet climate change continues, populism rises, inequality expands…
We have the tools and resources to fix these issues, but we lack alignment and good coordination. About 7% of all humans who have ever lived are alive today and we’re about to give birth to countless AGI. Alignment and organisation is undoubtedly the most powerful tool we have. Yet it kind of feels patchy and mystical like 1600s medicine. There's only so far leeches, blood letting and treating OKRs like voodoo can take you. So my focus is joining the effort to create structure & quantifications around these holistically. Not tinctures, but instruments.
LIS is an exercise in creating the highest agency humans possible. We think about it this way, are rats or pandas going to survive climate change? So are specialists or interdisciplinarians going to thrive alongside AGI? Taking multiple perspectives is the best route to solving the elephant and the blind men problem.
Through LIS, we then did work on getting concrete about problem categorisation. This allowed us to map how each is best understood and so how to create expertise. Experts have historically been one of the guiding organising forces Unfortunately, neo-liberalism has shown a real failure here, yet wisdom of the populist crowd is clearly not the answer either.
Most recently I’ve been working on using information theory to define measures of complexity. So we can quantitatively define when we can rely on traditional frequentist statistics, p-values or spot when black swans and fat-tails matter more.
Then a second measure for self-organisation and emergence. Truly scaled systems have to be distributed by definition. So understanding how to harness autonomous, self-organised alignment is crucial for both human collaboration and now agentic too.
/██ /██
| ██ | ██
/███████ /██ /██ /██████ /██████ /██████ /███████ /██████ | ██ /██ /██
/██_____/| ██ | ██ /██__ ██ /██__ ██ /██__ ██| ██__ ██|_ ██_/ | ██| ██ | ██
| ██ | ██ | ██| ██ \__/| ██ \__/| ████████| ██ \ ██ | ██ | ██| ██ | ██
| ██ | ██ | ██| ██ | ██ | ██_____/| ██ | ██ | ██ /██| ██| ██ | ██
| ███████| ██████/| ██ | ██ | ███████| ██ | ██ | ████/| ██| ███████
\_______/ \______/ |__/ |__/ \_______/|__/ |__/ \___/ |__/ \____ ██
/██ | ██
| ██████/
\______/
I'm continuing to explore the space of alignment. I'm currently working on ideas around the great flattening and enabling collaboration at scale. Such as the Goodhart problem of tokenmaxxing.
Working with good people matters most to me, so if you're a founder or researcher working in adjacent spaces like security, truth, metrics, complexity; definitely get in touch.