President & Co-Founder @OpenAI
Joined July 2010
- Tweets 4,815
- Following 65
- Followers 497,251
- Likes 1,386
From Richard Sutton (incompleteideas.net/), an essay on the repeated historical finding that computational scale has always beaten cleverness in AI (and some commentary on why this is such a hard-to-accept fact): incompleteideas.net/IncIdeas…
12
82
8
274
Need to drop 4 pounds before tomorrow's weigh-in. On restricted water and going to visit sauna tonight. Wish me luck!
Excited to compete in my first strength competition on March 16th in Santa Clara, in an expo with over 3k athletes and 10k audience members: facebook.com/events/21722260…! Been training hard, just like our reinforcement learning agents.
9
44
Really impressed by the latest class of OpenAI Scholars. Excited to follow their work over the next few months!
Meet the Spring 2019 class of OpenAI Scholars — who come from fields like economics, quantum physics and philosophy: openai.com/blog/openai-schol…
1
4
2
60
Thanks Michaël for the thoughtful commentary:
I wrote a blogpost on the recent @OpenAI LP controversy and how the ML community should grow from it.
bit.ly/2Hdi7R2
4
1
19
Good WSJ article on OpenAI’s new structure: wsj.com/articles/nonprofit-a…
2
5
25
Agree with Miles — one of the best parts of OpenAI is working with people who care so deeply about the mission.
When I first heard about OpenAI LP, I was concerned, as OAI's non-profit status was part of what attracted me. But I've since come around about the idea for two reasons: 1. As I dug into the details, I realized many steps had been taken to ensure the org remains mission-centric.
27
Per our Charter (openai.com/charter/) our primary means of accomplishing the mission is to build safe AGI directly. As you say it's going to take a lot of compute!
(Interestingly, our safety research looks similar in profile to other ML research and is also scaling fast.)
3
1
15
All transactions between the Nonprofit and the LP were approved by a subset of the Board — only directors without a financial interest in the LP. We had the fair market value of anything transferred from the nonprofit to the LP determined by an outside firm.
1
7
We will need to raise billions of dollars. If we do not, the Nonprofit would fail at its mission.
We then needed a structure that let us custom-write (very unusual!) rules like the following:
- Fiduciary duty to the charter
- Cap returns
- Full control to OpenAI Nonprofit
2
5
1
21
Thanks Catherine!
I'm very excited about this idea in general - a credible, binding structure that ensures that massive windfall profits get directed back to the common good.
I also am not a lawyer, and would love to see independent analysis of this structure (see also nitter.vloup.ch/gwern/status/110…)
7
Wrote a blog post together with my co-founder @ilyasut on the OpenAI Mission: blog.gregbrockman.com/the-op…
6
17
3
79
The board of the Nonprofit has full control, and can only takes actions it believes will further the mission.
1
1
1
12
Announcing a new structure for OpenAI which we've been working on for the past two years:
We’ve created OpenAI LP, a new “capped-profit” company that allows us to rapidly increase our investments in compute and talent while including checks and balances to actualize our mission. openai.com/blog/openai-lp/
5
10
2
109
Cool article on GPT-2, complete with a fun write up of the modern history of NLP results. towardsdatascience.com/opena…
17
86
Cool writeup in @WIRED about the technique: "Shark or Baseball? Inside the ‘Black Box’ of a Neural Network" wired.com/story/inside-black…
1
9
1
41
A new technique for analyzing how neural nets make decisions. Not necessarily incomprehensible black boxes after all!
In collaboration with Google, we're releasing Activation Atlases: a new technique for visualizing what interactions between neurons can represent.
💻Blog: blog.openai.com/introducing-…
📝Paper: distill.pub/2019/activation-…
🔤Code: github.com/tensorflow/lucid/…
🗺️Demo: distill.pub/2019/activation-…
4
83
4
318