Ten lessons from OpenAI’s five-day rollercoaster

Brett A. Hurt
9 min readNov 23, 2023

--

Created by DALL-E 3, which is now integrated into OpenAI’s multi-modal GPT-4

I originally published this yesterday morning on LinkedIn.

I was asked by Juan Sequeda to speak on Board governance given what happened at OpenAI. That internal Lunch & Learn at data.world is coming up post Thanksgiving. To organize my thoughts, here are ten lessons from OpenAI’s five-day rollercoaster:

1. Their weird structure as a non-profit originally dedicated to developing AGI out in the open and then becoming a for-profit entity, but with keeping the original non-profit Board governance, led to this. It is still unclear what happened in that Board room coup but the truth will come out over time, as it always does and perhaps quite soon. They should have restructured the entire Board governance structure when they decided to become a for-profit. Perhaps the for-profit should have been a different company, like a public benefit corporation (Inflection AI made that move and we are proudly one at data.world too), with some kind of IP or AI governance oversight by OpenAI, the non-profit. This is why Elon Musk has poked fun at them for becoming “ClosedAI” when they switched to a for-profit focus, which I don’t believe is very fair. I actually very much believe in Sam Altman as a leader from what I’ve seen — I think he’s extraordinarily altruistic. Listen to this interview with Sam on the Hard Fork podcast (you could skip the first 30 minutes to get right to it). This was recorded days before he was fired and is a long-form get by the co-hosts, Kevin Roose and Casey Newton. Decide for yourself what kind of leader he is after actually listening to him answer their difficult questions — I personally think he was fantastic in this and showed a lot of his cards. In any case, the non-profit/corporate governance structure that they chose was way too problematic, as was clearly on display over these past five days.

2. Speaking of Sam Altman’s altruism, he didn’t own any equity in OpenAI prior to being fired. David Sacks had a conspiracy theory on the All-In Podcast that he somehow did through the non-profit legacy, but I think that has now been proven to be not true. What was most telling is how 95% of OpenAI’s employees signed a letter of resignation unless Sam returned as CEO. That is pretty incredible and shows the power of human capital and effective CEO leadership. I think it proves how much Sam truly cares about the mission and the people of OpenAI, and it’s an incredible display of the culture he helped foster. And Sam’s very positive energy was clear at OpenAI’s DevDay, which was just a week before all of this went down. If you haven’t watched his keynote, along with Satya Nadella, it is worth watching (only 45-minutes long).

3. It is interesting that there are no founders on the new Board, nor are there women or people of color. I’m all about a meritocracy personally, as I wrote in Chapter 23 of my book. In that chapter, I write about how the importance of your recruiting pipeline development is critical if you want your company to be like the best of America, and I give advice on how to be very intentional about that, as was covered on this YPO podcast. There are so many well-qualified women and people of color they could have picked for OpenAI’s new Board. For example, Fei-Fei Li would have been an incredible choice and her new book on AI is fantastic.

To be clear, I’m a huge fan of Larry Summers and Bret Taylor. I’ve gotten to know a lot about Larry through No Labels, where he’s been involved and has spoken quite a few times. I’ve heard a lot about Bret Taylor, although I don’t know him personally. Both of them are among the greatest leaders of our time. I do wonder if OpenAI will focus on diversity, again combined with meritocracy, next and adding someone like Fei-Fei Li if they go back to a five-person Board. I’m also surprised they kept Adam D’Angelo on the Board given that he was part of the Board coup and the worst Board decision we’ve seen since Steve Jobs ouster. Even Ilya Sutskever signed that open letter of solidarity and made a deep public apology for his role in the Board coup.

Getting back to no founders on the new Board, but yet now with the founders, including Sam and Greg Brockman, being back at the helm of running OpenAI, it’s particularly unusual and in my opinion not a good idea. At data.world both Matt Laessig and I serve on our seven-person Board, which has two amazing independents in Sally Jenkins and Anita Lynch. The positive argument for it would be that it is much better than the super-voting structure that some founders, like Mark Zuckerberg, have on their Boards as it would lead to better governance. But Sam and Greg were on the OpenAI Board before as co-founders (Greg was even Chairman of the Board right before Sam was fired when he resigned in solidarity with Sam on Friday night.)

As far as managing your Board, Snowflake CEO Frank Slootman’s book, “Amp It Up”, has a good chapter on that (Chapter 18) — you should read it. I’ve lived all of the lessons he writes about in that chapter and couldn’t agree with him more. Parker Klein has a good summary of it on Medium.

4. As I said above, this was the worst Board decision since the ouster of Steve Jobs from Apple. And it was followed by the worst communication and crisis strategy I’ve ever seen. Satya Nadella is their biggest partner, by far, and was not even consulted — the OpenAI Board told him minutes before the press release went out. It shows no consideration for investors, partners, or even customers. I mean, they had just had their first-ever DevDay a week before with Sam prominently on stage leading it as CEO! The Board’s communication about his ouster suggested that Sam had no integrity. That “assessment” was clearly not felt by employees, again 95% of whom signed the letter that they would quit OpenAI and join Sam and Greg at Microsoft. And the OpenAI Board still never said what actually happened inside that Board room, so we are all just left guessing for now, which is already leading to some pretty wild conspiracies. Plus, all of this happened right before a massive secondary stock sale that was going to enrich employees that had been working incredibly hard to build the world’s leading AI company.

For more context on the horrible Board handling on this, OpenAI is already well over a billion in revenue less than one year after the launch of ChatGPT (Nov. 30th will be the one-year anniversary!) It hit over 100 million users just months after it’s launch and was the fastest application to ever do so and with no inherent user base advantage like Google had with Google+ (RIP) or recently Meta had when they launched Threads.

5. Speaking of that massive secondary stock sale, it valued the stock at $86 billion. It is now likely back on the table with Sam, Greg, and the team deciding to stay at OpenAI with the new Board. I was an early pioneer of secondary sales at Bazaarvoice and I can tell you that they make a huge difference in employee morale. It helped us create over a billion-dollar outcome and retain our best people. When a secondary stock sale is about to happen, it really impacts the way people start thinking about their future, most importantly taking care of their family. Selling stock along the way is incredibly motivating to employees that don’t already have the $500–700m of wealth that Sam reportedly has from his past endeavors. So, the OpenAI secondary stock sale that was blown up by the Board’s rash actions may have been a factor here. It now is probably going to be back on, and maybe even more so by the incredible employee display of unity to Sam, Greg, and the unifying company mission. There is now no doubt in the fervor OpenAI employees have to continue to build something “insanely great”.

6. IP rights matter, a lot. Satya Nadella held a lot of the cards on this rollercoaster ride because Microsoft had negotiated the IP rights to control their destiny with OpenAI. That played a big factor in the negotiation here, including with Sam and Greg temporarily agreeing to join Microsoft and head up a new advanced AI Lab. Kara Swisher even theorized that Sam Altman would be Satya’s successor as CEO of Microsoft in the most recent episode of the Pivot podcast. Kara’s scoop of interviewing Satya Nadella on Monday during the most heated moments of this Board coup was very telling on the IP rights front and how Satya was happy either way, whether Sam and Greg returned to OpenAI or not (although he made it clear that it would have to be with a different Board).

7. X, formerly known as Twitter, became the king of the real-time news again, even with as much trouble as they’ve had this year. Kara Swisher, the queen of tech reporting, had some slightly different and more real-time updates on Meta’s Threads, but at the end defaulted back to X as most of us did. That surprised me a bit — it was like our old habits came back in a crisis mode. I turned to X first myself to get the latest, and Elon posted/tweeted about their engagement being way up. At the beginning of this year, I wrote about Elon Musk and his wizardry, and it’s been a heck of a rollercoaster year for him too.

8. The fact this all happened so quickly is a sign of the social media times. Back when Steve Jobs was ousted, it took him 12 years to return to Apple. Matt Turck at FirstMark had the winning post/tweet on this, “Sam Altman returning to OpenAI after a day is like Steve Jobs returning to Apple after 12 years, but for the TikTok generation.” As was explained in the great documentary “The Social Dilemma” featuring Tristan Harris, social media is both a utopia and a dystopia. The fact this all happened so quickly and ended up with Sam and Greg back at the helm of OpenAI with a new Board of Directors falls into the utopia camp for me.

9. Satya Nadella’s incredible handling of all of this cements his status as the most important CEO in tech right now. Microsoft is now at a $2.81 trillion market cap as compared to the king Apple at a $2.97 trillion value with Google at a $1.74 trillion value. Indian Americans are the fastest growing minority group in America by wealth, and they are at the helm of many of the most important companies, including Google. One of the leadership practices I’ve embraced since 2013 is the study of Vedanta, an ancient Indian philosophy and practice on living, which I wrote about on my blog, Lucky7. From my view of Satya, he embraces the principles of Vedanta and is extremely in control of his emotions. Clearly the prior OpenAI Board was not and it led to the chaos we’ve seen over the fast five days. Satya was the wisest adult in the room. He also held all of the cards due to the compute that OpenAI needed, his investment stake, and his IP rights over OpenAI.

10. How much we all paid attention to this is a massive sign of OpenAI’s success! As Ethan Mollick has pointed out multiple times here on LinkedIn (you should follow him), no competitor has gotten close to GPT-4 this year, including Google who invented the base technology for it in the first place (and open-sourced it). In short, it shows that OpenAI is truly out in the lead and we all care about the fate of their company because of how much we love it.

2023 will go down as the year that the productivity boost of OpenAI and LLMs in general was undeniable and now we are down to the how of execution across the enterprise. Last Wednesday in a groundbreaking research report, we showed the power of LLMs + Knowledge Graphs to dramatically increase the accuracy, explainability, and governance of enterprise data insights.

--

--

Brett A. Hurt

CEO and Co-founder, data.world; Co-owner, Hurt Family Investments; Founder, Bazaarvoice and Coremetrics; Henry Crown Fellow; TEDster; Dad + Husband