AI: Humpty Dumpty still on the wall

...the challenges for OpenAI to regain its heights

For all the charm of the tale of Humpty Dumpty on a Wall, we never quite had the scenario of ‘All the King’s Horses and Men’ preventing Humpty from quite falling off the wall. OpenAI for now has avoided falling off the wall, thanks to Herculean behind the scenes actions and efforts by its sometimes feuding founders, a slate of advisors on either side, and of course the heart-felt petitioning of most its 800 strong employee base. Not to mention the backstop promises by its biggest partner and supporter, Microsoft and its proactive CEO Satya Nadella.

All of that enabled founders Sam Altman and Greg Brockman to return as CEO and President, with Mira Murati back as CTO. Founder Ilya Sustkever and former and presumably current Chief Scientist, who instigated the surprise firing of CEO Sam Altman on the friday afternoon of November 17, may or not return in his former role and capacity.

And of course, all the senior executives of OpenAI are for now OFF the board of the still non-profit OpenAI, as the new three member board puts together a fuller, more permanent board to figure out what happened and why, while also putting together a more stable governance structure and expanded Board, that threads the needle of AI Safety that was paramount for the non-profit board, with the commercial needs of the for profit business. And that includes tens of billions of further investments from Microsoft and other investors to continue its mission of providing Foundation LLM AI models that will make the current GPT-4 Turbo offerings seem ‘quaint’ in comparison, in not too many years.

I recount all this since it bears underlining the unusual forces ahead for OpenAI as it remains balanced on the Wall, to resume on its path of immense promise amidst the cacophony of perils that have been making OpenAI wobble on the Wall. That task remains ahead for its founders and other AI luminaries, not to mention regulators and other interested parties. All those parties and more remain laser-focused on those Safety issues even after a year that OpenAi’s ChatGPT electrified the world on the promise of AI. The AI Tech Wave as described in these pages, needs OpenAI to play ot its full promise, with as few distractions as possible. And yet the distractions in front of the company are many beyond the herculean nuts and bolts of running the AI rocket ship that is OpenAI.

Axios summarizes the situation aptly in a piece titled “OpenAI's board reset could alter nonprofit status”:

“Reinstated OpenAI CEO Sam Altman told Axios the "highest priority" for the company's new board is exploring a range of changes to "improve the governance structure" — including whether OpenAI remains controlled by a nonprofit entity, Ina reports.”

“Driving the news: Altman's comments came Wednesday as the company announced it has finalized arrangements to bring him back as CEO and give key partner and investor Microsoft a non-voting observer seat on the company's board.”

“Why it matters: Under the company's current structure, for-profit work by a subsidiary is overseen by a nonprofit board whose mission is to ensure that ultra-powerful artificial intelligence "benefits all of humanity."

“Altman told Axios that OpenAI's effort to reorganize itself will "take real time and consideration ... We're going to need some patience there."

• “Altman added that he hopes an independent review of events that was a condition of his return will shed light on why he was ousted in the first place. "I'd like to understand it better," he said.”

As Sam Altman himself outlines his key priorities in the weeks ahead, they include:

“I have a mountain of very important and urgent work in front of me," he said. "I imagine plenty of scenarios where I'm not [on the board], even though I think CEOs usually are."

“Altman's memo said OpenAI is focused on three tasks: Ensuring safety, continuing to build products and expanding its board.”

Most industry leading companies led by their visionary founders in earlier tech cycles have mostly had the not to easy task of just building and scaling no. 2, “build products”. Presumably that matter to partners, customers and users of all types, that ensure the leadership of the company in question.

In OpenAI’s case, Humpty Dumpty has to make sure that it’s also done without ‘existential risks for all Humanity’, as per its core non-profit charter, and navigate a path to an expanded board that continues to thread the needle between safety and speed. At scale. Heavy responsibilities indeed for Humpty Dumpty going forward on that Wall. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)