OpenAI Saga: The Best Links and Memes

PLUS: Ridley Scott's directing secret sauce.

Nov 24, 2023

Thanks for subscribing to SatPost.

Today, we will talk about the 5 best Thanksgiving side dishes. Nah, I’m kidding, we’re obviously talking about the OpenAI vs. Sam Altman saga.

Also this week: Ridley Scott’s directing secret sauce.

You probably saw the wild tech news from last week.

OpenAI fired CEO Sam Altman and re-hired him 4 days later.

When a wild tech story pops off — like the FTX and SBF implosion or the Silicon Valley Bank (SVB) fiasco — X/Twitter is still the the best place to consume breaking news and get some very hot takes.

Turner Novak’s summary of the OpenAI saga is spot on:

I was unhealthily glued to my phone for a solid 120 hours. I’m pretty sure I destroyed millions of brain cells trying to chase engagement (I’m not proud of this). On the plus side, I bookmarked some incredible posts and thoughtful article links to share with you all.

Here is a brief OpenAI timeline to provide context for the drama:

OpenAI is founded in 2015 as a non-profit with the mission to “ensure that artificial general intelligence (AGI) — AI systems that are generally smarter than humans — benefits all of humanity.”
Co-founders include Elon Musk, Sam Altman, Reid Hoffman, Jessica Livingston, Andrej Karpathy, Greg Brockman, Ilya Sutskever. The organization receives $1B in commitments (but only ever gets $130m of that amount).
Elon is the largest donor and contributes $40m+. His motivation is to create a counterbalance to Google’s for-profit AI efforts.
In 2018, Elon leaves OpenAI (citing a conflict of interest with Tesla’s AI efforts).
In 2019, Altman becomes OpenAI CEO. With Elon no longer funding the organization, Altman transitions OpenAI from a non-profit to a “capped” for-profit structure that allows for outside investment (but caps an investor’s return to a maximum of 100x and any funds above that cut-off goes back to the non-profit). The governance structure now has a board of directors overseeing a non-profit entity, which oversees the for-profit entity.
Also in 2019, Microsoft agrees to invest $1B into OpenAI (largely in the form of cloud computing credits). Altman takes zero equity in the for-profit entity.
In November 2022, OpenAI releases ChatGPT and it becomes the fastest consumer software product to get to 100m users.
In January 2023, Microsoft agrees to invest another $10B+ into OpenAI for a 49% ownership stake in the for-profit entity (it is reported that the company’s total prior investment amounted to $3B).
In the first week of November 2023, OpenAI hosts its first developer day and reports show that the business is booming with an annual run rate of $1B+.

That takes us to November 17th, 2023. The most important AI startup — now valued at ~$90B — is firing on all cylinders. But there is a corporate governance time bomb waiting to explode as commercialization efforts conflict with the non-profit’s original mission to safely create AGI.

Below is a day-by-day breakdown of the OpenAI saga with accompanying X posts:

***

On Friday, November 17th:

The OpenAI board fires Sam Altman and removes president Greg Brockman as board chair (they ask Brockman to stay in his management role).
Prior to Altman’s firing, the board was composed of Altman, Brockman and 4 other people: Helen Toner (Director of Strategy, Georgetown's Center for Security and Emerging Technology), Tasha McCauley (RAND scientist and Joseph Gordon-Levitt’s wife), Adam D’Angelo (Founder, Quora) and Ilya Sutskever (OpenAI co-founder and Chief Scientist).
OpenAI CTO Mira Murati is named the interim CEO.
The 4-member board group that ousted Altman releases a letter announcing a “leadership transition” and it says that Altman had not been “consistently candid in his communications with the board”.
The board’s letter is so vague, that people assume the worst. Why was Altman fired so abruptly? Was it an egregious personal transgression? Was it financial malfeasance? Was it a major safety or privacy snafu?
Microsoft and its CEO Satya Nadella are told about Altman’s firing only 5 minutes before the OpenAI board shares the letter on X. Nadella is reportedly “furious” about the board’s poor communication (fair enough: Nadella is the hall-of-fame CEO of a $2.8T company and only found out about Altman’s dismissal 5 minutes before me, the CEO of literally nothing of note ever).
Brockman posts his resignation on X, thus showing public support for Altman (which makes it much less likely that Altman committed serious malfeasance).
Sutskever holds a tense all-hands meeting with OpenAI employees. Without further clarification from the board, many employees believe that Sutskever has enacted a coup. A number of senior employees resign.
Brockman and Altman release a joint message sharing a timeline of how they were fired. Each was invited into a separate Google Meet chat and told about the board’s decision (I repeat: the OpenAI board used Google Meet even though Microsoft is its largest investor).

***

On Saturday, November 18th:

OpenAI COO Brad Lightcap sends a memo to OpenAI employees saying that Altman’s ouster was due to a “breakdown in communication” with board members and his dismissal did not involve financial malfeasance or security/privacy practices.
The OpenAI board reaches out to tech execs for the CEO job including former Github CEO Nat Friedman and Dario Amodei (CEO of AI firm Anthropic, which was launched by OpenAI defectors in 2021).
With the board not sharing any more details on Altman’s firing, major OpenAI investors begin agitating for Altman’s return.
Altman tweets “I love the openai team so much” and dozens of employees show their support by re-posting or replying to the post with a ❤️ emoji.

***

On Sunday, November 19th:

Microsoft CEO Satya Nadella works directly with interim CEO Mira Murati to negotiate Altman’s return to OpenAI.
A key detail of Microsoft’s $10B commitment is that “only a fraction” of the funds “has been wired to the startup, while a significant portion of the funding, divided into tranches, is in the form of cloud compute purchases instead of cash, according to people familiar with their agreement” per Semafor. This gives Microsoft significant leverage.
Altman goes to OpenAI’s office in the afternoon and posts a photo of himself holding a guest tag with the caption “first and last time i ever wear one of these” referring to the tag.
Negotiations fall apart and the OpenAI board hires Emmett Shear as the new interim CEO (Shear is a co-founder of streaming service Twitch and was in the same Y Combinator class as Altman in 2005).

NOTE: Before the board announced Shear as the new interim OpenAI CEO, one of the best running jokes was a meme about fake CEO candidates.

***

On Monday, November 20th:

Around midnight Sunday (so Monday AM), Nadella posts on X that he has hired Altman and Brockman to lead a new AI research unit within Microsoft (Nadella’s post says Microsoft will also continue its partnership with OpenAI).

A few hours later — circa 2-3am — OpenAI employees share a letter criticizing the board’s decision to fire Altman and threaten to join the new Microsoft AI unit.
Chief scientist Ilya Sutskever pulls a 180 degree turn and signs the employee letter against the board and — at ~5am — says on X that “I deeply regret my participation in the board’s actions. I never intended to harm OpenAI. I love everything we’ve built together and I will do everything I can to reunite the company”.
Dozens of OpenAI employees post this message on X: “OpenAI is nothing without its employees”.
Within 12 hours, over 700 of the company’s 750+ employees sign the letter against the OpenAI board and signal that they are ready to join Altman at Microsoft’s new AI unit.
As part of Microsoft’s offer to hire OpenAI employees, it cleared out a whole floor of the SF Linkedin office. And assured them they “wouldn’t even have to use Microsoft Teams” per WSJ.
In the evening, Nadella goes on CNBC and Bloomberg to say that he is committed to work with Altman, whether he is at OpenAI or Microsoft (either way, he expects significant governance changes at OpenAI).
During the entire previous 72-hour period, executives from Facebook, Google, Salesforce, Nvidia, xAI and other Big Tech firms are all trying to poach OpenAI employees.

***

On Tuesday, November 21st

Negotiations between the OpenAI board, Altman, investors and Microsoft are ongoing: interim OpenAI CEO Shear represents the board while Bret Taylor — former Salesforce co-CEO and former Twitter board chair — plays mediator.
Shear asks the OpenAI board for a written explanation on why it fired Altman and says he will resign if the document isn’t forthcoming.
Amidst the uncertainty, OpenAI ships an AI voice tool and ChatGPT briefly goes down.
A deal is struck by 10pm with OpenAI’s X account announcing “We have reached an agreement in principle for Sam Altman to return to OpenAI as CEO with a new initial board of Bret Taylor (Chair), Larry Summers [former US Treasury Secretary], and Adam D’Angelo.”
As part of the deal, Altman agrees to an internal investigation on his management conduct and won’t — for the time being — take a board seat. The board will be expanded to at least 9 members in the coming months with Microsoft likely getting a seat and potential veto rights on future commercial deals.
OpenAI celebrates the return of Altman and Brockman by getting their drink on.

***

On Wednesday, November 22nd

One day prior to American Thanksgiving, Reuters reports that the OpenAI board may have fired Altman due to disagreements over a new breakthrough called Q*:

Ahead of OpenAI CEO Sam Altman’s four days in exile, several staff researchers wrote a letter to the board of directors warning of a powerful artificial intelligence discovery that they said could threaten humanity, two people familiar with the matter told Reuters.
The previously unreported letter and AI algorithm were key developments before the board's ouster of Altman, the poster child of generative AI, the two sources said. […]
After being contacted by Reuters, OpenAI, which declined to comment, acknowledged in an internal message to staffers a project called Q* and a letter to the board before the weekend's events, one of the people said. An OpenAI spokesperson said that the message, sent by long-time executive Mira Murati, alerted staff to certain media stories without commenting on their accuracy.
Some at OpenAI believe Q* (pronounced Q-Star) could be a breakthrough in the startup's search for what's known as artificial general intelligence (AGI), one of the people told Reuters. OpenAI defines AGI as autonomous systems that surpass humans in most economically valuable tasks.
Given vast computing resources, the new model was able to solve certain mathematical problems, the person said on condition of anonymity because the individual was not authorized to speak on behalf of the company. Though only performing math on the level of grade-school students, acing such tests made researchers very optimistic about Q*’s future success, the source said.
In their letter to the board, researchers flagged AI’s prowess and potential danger, the sources said without specifying the exact safety concerns noted in the letter. There has long been discussion among computer scientists about the danger posed by highly intelligent machines, for instance if they might decide that the destruction of humanity was in their interest.

If you want a technical explanation on the topic, Nvidia AI scientist Jim Fan has a good post.

My woefully inadequate summary: do you remember that time Google’s DeepMind created AlphaGo Zero? It was that AI program that taught itself how to play Go — without leaning from any human games — and then it beat the Go world champion. Q* seems to be a similar breakthrough for general reasoning.

Fan — who has a great last name — concludes by saying, “Note that what I described is just about reasoning. Nothing says Q* will be more creative in writing poetry, telling jokes [xAI’s GroK], or role playing. Improving creativity is a fundamentally human thing, so I believe natural data will still outperform synthetic ones.”

***

On Thursday, November 23rd:

Sam Altman had Thanksgiving dinner with OpenAI board member Adam D’Angelo. He posted about it and D’Angelo reposted the post. And I am re-posting that re-post below.

I’m guessing the OpenAI saga still has some chapters to come.

OpenAI Saga Links

If you want a podcast deep-dive on the OpenAI saga, the Not Investment Advice (NIA) team cooked for 75-minutes on the topic.

Otherwise, here are some notable links and other random posts I bookmarked:

The history of OpenAI’s board: A very good breakdown of how the board’s complicated structure and turnover over the years created a window of opportunity for the coup. (Fabricated Knowledge)
Effective Altruism (EA) takes another L: Three of the OpenAI board members that moved to oust Altman — Helen Toner, Tasha McCauley, Adam D’Angelo — all have links to the EA movement. This movement was infamously associated with FTX founder Sam Bankman-Fried and is very concerned with AI safety, at the expense of AI innovation. (Pirate Wires)

Thoughts of the new OpenAI board:: Former Facebook exec and current investor Sam Lessin argues that “With OpenAI’s new non-profit board, they just retired their ‘cult’ founding identity in favor of being a ‘business’. Now [OpenAI] needs a proper employment contract with its CEO, to actually pay him, drop the religious AGI crap, and move on.” (X/Twitter)
Why was Altman fired from Y Combinator? Paul Graham founded startup incubator Y Combinator. He tapped Altman as his successor in 2014 but then fired him in 2019. Details of that decision — including concerns over Altman’s side-dealings and focus on personal power attainment at the expense of the organization — show parallels with the OpenAI drama. (Washington Post)
Altman was trying to raise billions for an AI chips startup: To do so, he was leveraging the OpenAI brand. If Altman is able to build chips that are competitive with Nvidia and custom for OpenAI, he would clearly have leverage over the non-profit OpenAI board. This venture may have been one of the reasons that the board said he was not “consistently candid in his communications”. (Bloomberg)
Altman vs. Toner: Prior to his ouster, Altman tried to remove Helen Toner from the OpenAI board. His reason: she had written a research paper critical of OpenAI’s approach to AI safety. The board had shrunk from 9 to 6 people over the course of 2023. The plan was to add new board members but the board couldn’t agree on commercialization vs. AI safety. Altman may have made a move to oust Toner in an effort to re-stack the board with more allies. Then, the Board struck back. (NYT)
Did Altman plan the entire OpenAI saga to gain more power? After everything, it looks like Altman will return to OpenAI with an aligned board and 90%+ of the company having publicly pledged support for him. Was this all a 5D chess move to be in a much stronger position? Some believe that Altman previously pulled off a 5D chess move in the re-organization of the company Reddit (although Reddit co-founder Alexis Ohanian says otherwise).
Altman and Grand Strategy: Packy McCormick draws lessons from great historical figures (Napoleon, Augustus Ceasar) to better understand OpenAI’s board struggle and how Altman wields power. (Not Boring)
AI model agnosticism: Box CEO Aaron Levie says the OpenAI shakeup will lead to more companies becoming “AI model agnostic”, which means that they aren’t married to a single AI provider. While OpenAI is ahead on performance and price, the other competitors (Anthropic, Meta, Google, Stability AI) are close enough in model quality that companies can hedge by using multiple AI models. (Big Technology podcast)
What is the AGI threat? Thomas Pueyo writes “We don’t know how to contain or align a FOOMing AGI”. (Uncharted Territories)
Did Microsoft win? On Monday, it looked like Satya Nadella was able to acquire OpenAI for $0 (obviously, he’d have to pay the salary of 700+ employees). Most considered this a W but Cloudflare CEO Matthew Prince took the opposite position. He said that prior to the board coup, Microsoft had all the upside from OpenAI with an arms-length relationship that protected it from regulatory scrutiny and potential lawsuits around LLM AI hallucinations. This is something to watch as OpenAI expands the board to 9 or more members. For its $10B+ investment, Microsoft should reasonably want a board seat and veto rights on OpenAI commercial deals. But it may also want to maintain deniablity.
Does Meta win? There is clearly risk of betting all your chips on a single centralized AI provider. In that case, open-source AI models look like a winner. And Meta has the most notable open-source AI language model (Llama). Meta’s head of AI Yann LeCun posted this during the OpenAI saga:

Here is a 90-second recap using the legendary El Risitas meme template:

Here’s a baller thread comparing Game of Thrones characters to OpenAI characters:

Huge props to billionaire Salesforce founder Marc Benioff for getting his hustle on and trying to recruit OpenAI employees on X:

Today’s SatPost is brought to you by Bearly.AI

Why are you seeing this ad?

Because I co-founded an AI-powered research app called Bearly AI. And I really like putting blue buttons in this email.

If you press this blue button below, you can save hours of work with AI-powered tools for reading (instant summaries), writing (Claude, ChatGPT), text-to-image art (literally type some text and get a wild image) and speech-to-text transcriptions.

Available as a desktop or iPhone app.

Try Bearly AI for FREE

SatPost by Trung Phan

Discussion about this post

SatPost by Trung Phan

OpenAI Saga: The Best Links and Memes

PLUS: Ridley Scott's directing secret sauce.

OpenAI Saga Links

Other Links

Discussion about this post