AI has recently crossed a utility threshold, where cutting-edge models such as GPT-3, Codex, and DALL-E 2 are actually useful and can perform tasks computers cannot do any other way. The act of producing these models is an exploration of a new frontier, with the discovery of unknown capabilities, scientific progress, and incredible product applications as the rewards. And perhaps most exciting for me personally, because the field is fundamentally about creating and studying software systems, great engineers are able to contribute at the same level as great researchers to future progress.

“A self-learning AI system.” by DALL-E 2.

I first got into software engineering because I wanted to build large-scale systems that could have a direct impact on people’s lives. I attended a math research summer program shortly after I started programming, and my favorite result of the summer was a...

For the first three years of OpenAI, I dreamed of becoming a machine learning expert but made little progress towards that goal. Over the past nine months, I’ve finally made the transition to being a machine learning practitioner. It was hard but not impossible, and I think most people who are good programmers and know (or are willing to learn) the math can do it too. There are many online courses to self-study the technical side, and what turned out to be my biggest blocker was a mental barrier — getting ok with being a beginner again.

Studying machine learning during the 2018 holiday season.

Early days

A founding principle of OpenAI is that we value research and engineering equally — our goal is to build working systems that solve previously impossible tasks, so we need both. (In fact, our team is comprised of 25% people primarily using software skills, 25% primarily using machine...

The text of my speech introducing OpenAI Five at Saturday’s OpenAI Five Finals event, where our AI beat the world champions at Dota 2:

“Welcome everyone. This is an exciting day.

First, this is an historic moment: this will be the first time that an AI has even attempted to play the world champions in an esports game. OG is simply on another level relative to other teams we’ve played. So we don’t know what’s going to happen, but win or lose, these will be games to remember.

And you know, OpenAI Five and DeepMind’s very impressive StarCraft bot
both have beaten good pros privately but no one has seen it it happen live — so if we win, it’ll be a first on that front too.

This event is really about something bigger than who wins or loses: letting people connect with the strange, exotic, yet tangible intelligences produced by today’s rapidly progressing AI technology.

We’re all used to...

This post is co-written by Greg Brockman (left) and Ilya Sutskever (right).

We’ve been working on OpenAI for the past three years. Our mission is to ensure that artificial general intelligence (AGI) — which we define as automated systems that outperform humans at most economically valuable work — benefits all of humanity. Today we announced a new legal structure for OpenAI, called OpenAI LP, to better pursue this mission — in particular to raise more capital as we attempt to build safe AGI and distribute its benefits.

In this post, we’d like to help others understand how we think about this mission.

Why now?

The founding vision of the field of AI was “… to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it”, and to eventually build a machine...

The text of my speech introducing OpenAI Five at yesterday’s Benchmark event:

“We’re here to watch humans and AI play Dota, but today’s match will have implications for the world. OpenAI’s mission is to ensure that when we can build machines as smart as humans, they will benefit all of humanity. That means both pushing the limits of what’s possible and ensuring future systems are safe and aligned with human values.

We work on Dota because it is a great training ground for AI: it is one of the most complicated games, involving teamwork, real time strategy, imperfect information, and an astronomical combinations of heroes and items.

We can’t program a solution, so Five learns by playing 180 years of games against itself every day — sadly that means we can’t learn from the players up here unless they played for a few decades. It’s powered by 5 artificial neural networks which act like an...

It’s been two years since I wrote define CTO, in which I documented my quest for a role where I could have scalable impact by writing code. I’ve finally found that role, though not by seeking it — instead, I sought out a problem more important to me than my role within it, brought together the right people, and found that I can best make them effective by writing code.

Formation

In August 2015, OpenAI was just an idea articulated over a dinner with Elon Musk, Sam Altman, Ilya Sutskever, me, and a handful of others. We’d each come to the dinner with our own ideas, but Elon and Sam had a crisp vision of building safe AI in a project dedicated to benefiting humanity. I wanted to contribute however I could. Sam and I started rallying a team to turn this idea into reality.

We were missing a core ingredient: we needed an AI technical visionary, someone whose intuition and ideas we could...

I started programming seriously during a gap year after high school. I’d read Turing’s Computing Machinery and Intelligence, and was inspired by the notion of writing code that could understand something that I, as the code’s author, did not. I started writing a chatbot — how hard could it possibly be?

I managed to build something that could talk about the weather very convincingly. But no matter where I looked, it seemed that no one had any techniques that could make my bot really work.

I soon shelved my chatbot pursuits. I decided to focus on creating systems that can have a real impact, and have been doing that ever since.

College

In college, I found a field that captured what drew me to AI: programming languages. I was thrilled that a compiler or static analyzer could “understand” a program in a way that I couldn’t, and then apply that understanding to do something very useful...

I’ve been advising Stellar since Stripe helped it launch about a year ago. Today I’m joining their board.

Digital currencies are still nascent, and my hopes for them remain unchanged. Particularly, we need digital currency protocols like Stellar that work with the existing financial system and focus on a seamless user experience.

I’ve always been impressed with Joyce, Jed, and their team’s approach to the space. Stellar is a non-profit entity producing open-source software, with a strong emphasis on financial inclusion. From launch day, they’ve been hyperfocused on the core technology and community. In the past year, they’ve made strong progress, including a provably-correct consensus algorithm (complete with explanatory graphic novel), a redesign and from-scratch implementation of the technology, and the groundwork for a pilot in South Africa.

Building anything valuable takes time...

Coding requires collaboration. As Andrew Bosworth said recently: doing anything meaningful past a certain point requires more than one person. So if you want to build, it’s important to do so as part of a welcoming, collaborative environment.

One environment I’ve long admired is that of the Recurse Center (formerly known as Hacker School). They’ve been unusually thoughtful about the dynamics of their culture. I’ve always thought that if I had three months to spare, I would attend a batch to experience the community directly (and hopefully contribute back however I can).

And, well, now I have that kind of time.

I was initially surprised by how many experienced engineers told me that they too would attend a Recurse Center batch, if only they could make the timing and logistics work. But I think it shows that no matter how long you’ve been coding, there will always be areas of...