An unprepared board’s battle against a CEO with enormous soft power – yes, it is about OpenAI
As I wrote about the Board vs CEO struggle at OpenAI I kept ending up here: it’s all about soft/indirect power and understanding how it is wielded most efficiently.
This is too big a topic for one post, and I’ll focus on these points here: Why it matters; Who are these people anyway; A strange governance structure; Why you need experienced generalists on boards; Power has many guises; Hijacked by ideologues?; Links to: good LLM, AGI, and general AI explainers and OpenAI drama background.
Why it matters
You realise how much the Board of a business matters when things don’t go to plan. The bigger the problem - the worst impact a badly constituted or weak board will have.
It isn’t just about technical competence but business savvy, strategic thinking, experience, fiduciary care, understanding governance and whose interest they need to act in. Which is not always easy or straightforward: being able to navigate complex situations takes on a whole different meaning when it comes to problems that reach a company’s board. Takeovers, especially the hostile kind. Scandals. Replacing a CEO.
I read somewhere that being a non-executive board director is like “Being a critical friend”.
Boards have several functions, many of which fall under strategy and governance, and board members also take on serious responsibilities, including fiduciary. (You need D&O insurance to manage this risk.)
Being a critical friend is an interesting summing up of how independent directors should think and act. You need to suss out and call out critical issues – but you are not there to attack and dethrone the CEO. The way you act as an independent director should be in the interest of the company and the shareholders (hence the fiduciary part). In theory, in my view, that also means you shouldn’t put your ideological beliefs ahead of what is best for the organisation and the shareholders. This is one reason why one needs to be careful what board appointment one takes – you don’t want to find yourself in a situation where your personal values and the needs of the business clash.
The drama two weeks ago at OpenAI, which saw the removal of Sam Altman, the CEO, the naming of two CEOs, the reinstating of the original CEO and the removal of the board in the space of a weekend and a couple of days, really caught the public imagination. Dizzying, I know.
Part of the reason for the flurry of attention was that everything from the first public statement that kicked off a furious, fast-moving soap opera, all the way until Altman was officially reinstated two days ago, hinted at a clash of values and duties and interests – and some very bare bones power struggles.
Ironically, it doesn’t get much more human than that.
Who are these people anyway
If you are not familiar, still, with OpenAI, it is the organisation that brought us DALL-E, the AI that creates images based on simple worded commands and also ChatGPT that some folks already use to write their blogposts and college kids to write their essays. The former is also the AI that was trained on all those images we’ve all been posting all over the internet. The latter was trained on all the words, and the best of books, and is being used as the basis for more specialist GPTs, like ones that some think can replace lawyers and others living off their brains and ability to write.
If you would like a brief review of what OpenAI does and why many of us consider it one of the most important companies in the world right now, you could start with this Bloomberg explainer video
It is worth noting that to this day no one has been able to ascertain what if anything Altman did wrong other than having ways and ideas that conflict with those four people who fired him.
A strange governance structure
OpenAI has many quirks. I won’t go into too much detail as much has been written about it even in daily newspapers, but what is important for our purpose is that the founders have innovated the governance structure, combining non-profit and for-profit elements – and ended up with a kind of ownership and board structure that doesn’t seem fit for purpose. At least not in this form.
For example, Microsoft invested billions into the for-profit businesses and has a 49% stake in it - but got no board seat on the non-profit board sitting on top of it all. (This is changing now.)
Normally, in a business, you would expect to find representatives of the largest shareholders on the board. They are there at least partly to protect the investment, i.e. the value of the business. Even the independent directors would be obliged to consider the interests of shareholders. Being bound by fiduciary duty is pretty universal.
It is the board’s job to hold the CEO to account. To put it crudely, to make sure they don’t just do whatever their own personal interests dictate.
Also important is to act as a sounding board to the CEO et al. To debate, to deliberate, which by nature includes asking questions and getting answers – not throwing toys out of the pram (like the OpenAI ex-board) and not just sitting there letting the executives do whatever.
To be clear, even though top executives normally sit on the board, the board does not execute. It does not run the business. That’s the CEO’s job.
Anyways, I am not here to lecture you on boards I just wanted to put into context why what happened at OpenAI is interesting and important.
So important, in fact, that I did what I never do and real life posted through it on Threads. (You can find me at juditpetho. Those in the EU should be able to get on Threads soon, but I’ll post some screenshots at the end of this post.)