Link Search Menu Expand Document

Rethinking thinking.

I’m building tools which augment the mind, enabling people to think in powerful new ways.

View Tools Support My Work


Notions like learning, attention, and memory are now central to AI research. However, AI research is far from being central to human learning, attention, or memory. I was intrigued by this largely one-way flow of insight from cognitive science to machine learning, and spent quite some time thinking about means of going the other way. How could breakthroughs in artificial intelligence be used to understand and amplify human intelligence? This curiosity materialized into a growing collection of tools for thought, some more radical than others, designed to augment the human intellect in ever more exciting ways.

Current Focus


Guiding Principles

1. Tools must augment the mind.

Just like there are sound frequencies we cannot hear and wavelengths of light we cannot see, there are thoughts we cannot think.1 However, just like intricate sensors help us detect sound and light beyond our senses, there are transformative tools which enable us to think beyond what was previously thinkable.2

As an example, consider the period just before the chart was invented. Data was mostly expressed as tables of numbers. The chart enabled us to repurpose our potent perceptual machinery in order to digest data in a powerful new way.3 Today, charts are an integral part of our science and engineering.

There are countless other examples of such transformative tools. Writing enabled us to compactly encode thoughts into sequences of symbols, making them observable and supporting critical reflection.4 Hindu-Arabic numerals enabled us to perform arithmetic quickly and accurately, promoting computational thinking.5 Maps enabled us to embed entire continents into handheld objects, supporting spatial reasoning.6

Such tools are victims of their own success. Their effectiveness prompted us to internalize them so deeply that most of the time we aren’t even aware of them.

Today, charts evolved into interactive animations.7 Search engines now enable us to seamlessly navigate humanity’s collective knowledge. We solved arithmetic by outsourcing it to computers. Machines help us manipulate abstract representations.8 Digital technology shows great promise in creating our most powerful tools for thought yet.

However, it’s not all roses. Social media platforms are intellectually debilitating for their users.9 Outsourcing decision-making to recommendation systems leads to users being optimized into mindless consumers.10 The collective consciousness of entire nations is shaped by a handful of opaque algorithms, resulting in digital colonialism.11

A tool is never a neutral artifact, but a faithful depiction of its creator’s beliefs.12 The tools I’m building are deeply informed by the belief that tools must first and foremost augment the minds of their users. To this end, I am:

  • Developing tools for thought.
  • Developing neural interfaces.

2. Knowledge must be distributed.

Knowledge is power. Many believe that power shouldn’t disproportionately reside in the hands of a select few. Therefore, knowledge should also be distributed. In my view, knowledge extends beyond literature, and includes diverse flavors such as multimedia content, data, source code, hardware schematics, and others.

The internet aspired to provide just that – a digital ecosystem of distributed knowledge. Unfortunately, its then frail infrastructure couldn’t fulfill this ambition. However, distributed technologies are maturing. Peer-to-peer protocols support communication.13 Distributed hash tables support storage.14 Distributed ledgers support currency.15 Version control systems support collaboration.16 The ambitions of the early internet are within our technological reach.

My intended contribution to this space is to lead by example. To this end, I am:

  • Building open source software.
  • Building on open source software.
  • Building with open source software.
  • Practicing open science.

3. The future is malleable.

Technological determinism refers to the seemingly predictable development of technology across measures like compute costs.17 Based on such observations, it’s tempting to blindly subscribe to the idea that course-correcting our technological trajectory is impossible. Indeed, this trajectory is shaped by commercial, ideological, and social imperatives, but these aren’t immutable either. They can also be shaped.

Not only is challenging our technological trajectory possible, but it’s also useful. The practice of design fiction offers us a genuine framework for peeking into possible futures, shedding light on both opportunities and pitfalls.18 By shaking off our disbelief about change, speculative practices remind us that the future is malleable, and in doing so empower us.19

To this end, I am:

  • Creating works of design fiction.