Hi, I'm Tomiwa ("Toh-mee-wah")

Current work

Everything I do centres on human agency and potential, and the ideas, institutions and technologies that expand both. I founded Logarithms, a holding company that builds and acquires products, to put that belief into practice.

I focus on:

• Spatial reasoning. Building AI models that can reason about space, geometry and the physical world. I'm interested in taking AI beyond language and into the world we inhabit: helping people create, travel, socialise and tell stories, while giving robots the understanding they need to assist us.
• Personal intelligence. Bringing state-of-the-art models onto your devices, so your intelligence and private data remain yours.
• Permissionless ownership. Giving creators direct ways to register, exchange and monetise their intellectual property without relying on central intermediaries.
• Expert human data. Working with domain experts to create the tasks, rubrics and reference answers that teach frontier models consequential work and let us test whether their judgement can be trusted.
• Short films. Stories about technological futures worth building.

My background is in aerospace, computer science and machine learning. I also write and invest.

Things I'm interested in:

• Democratising access. How can we make valuable products, services and expertise available to people who were previously priced out or excluded?
• Human potential. How can we make what people achieve less dependent on where they start?
• Ownership. How can people retain more control over their work, data and intellectual property while still allowing others to build with it?
• Humanism in the age of AI. As machines become more capable, how do we use them to expand human agency rather than diminish it?
• Creativity. What happens to taste, originality and creative expression when almost anyone can generate technically competent work?
• Deconstructing how things work. I like taking difficult skills and systems apart, testing their assumptions and rebuilding them as practical handbooks.
• Flying. I learnt to fly before I learnt to drive. I'm turning the lessons into a course notebook for anyone curious about what it takes to pilot an aircraft.
• Films. I make short films as another way to explore mythology, culture, technology and human choices.
• Philosophy. I'm interested in the ideas beneath how we live: agency, responsibility, meaning, knowledge and progress.
• Reading. I read across technology, economics, philosophy, history, fiction and science. My bookshelf is an evolving record, not a list of prescriptions.
• Economics. Why do some systems create mobility, abundance and participation while others concentrate opportunity?
• Investing. I invest in ambitious companies and people working on consequential problems.
• Small teams and breakthrough ideas. How did Orville and Wilbur Wright solve powered flight while better-funded researchers failed, and what might that teach us about the path to AGI?
• How intelligence is taught. What human knowledge, judgement and experience must be captured to make general models reliable in specialised domains?
• Theatre. I'm learning how performance, voice and presence change the way a story reaches an audience.
• The new internet economy. How can more people participate in, contribute to and own the networks they help create?
• Tennis. A uniquely demanding combination of endurance, strategy, technique and emotional control.
• Manga. One of my favourite forms of long-term storytelling. I'm compiling a list of the series I think are worth reading.
• Negroni Sbagliato. The recipe I return to.

The Mantras I live by

• Zig and zag. When everyone zigs, you zag. If you're doing what everybody else is doing, you will not succeed.
• It's not about more how-to information; it's about why-to. Before you read your next book, ask whether you have put anything from the last one into practice. It's not enough to ingest more pages.
• Don't follow footsteps; follow what they sought. Those footsteps rolled up behind them.
• The Pareto principle.
• Do things that scare you.
• Create value for others.
• Document everything you learn.
• Start from the end.
• Don't undervalue yourself or your time.
• Automate as much as you can.
• Expect disaster.

Work with me and my team

• To build, sell or join a company, visit Logarithms.
• For product, design and engineering, visit Gum.tech.
• For research, model evaluation and human data, visit Logarithms Labs.