The news: OpenAI’s board, which includes the company’s chief scientist Ilya Sutskever, Quora chief executive Adam D’Angelo, entrepreneur Tasha McCauley and Toner, decided to oust Altman at the weekend. They said he had not been “consistently candid”. Company chairman and Altman backer Greg Brockman was not involved in the discussions. He was demoted and then resigned. Sutskever, an OpenAI co-founder who fronted the push to oust Altman, then apologised. Altman appeared likely to return with the backing of investors but the board instead chose the co-founder of video streaming platform Twitch, Emmett Shear, as chief executive. Microsoft hired Altman and Brockman. Almost all OpenAI staff signed a letter threatening to defect to Microsoft unless the board resigned. Chaos reigned.
Who is Toner? Toner graduated from the University of Melbourne in 2014, where peers remembered her as smart and principled. She was “not a dud”, one said, which is a compliment in the language of people who were involved with Toner at UN Youth, an organisation where nerdy students play-act as countries at international forums. Another said she was just part of the crowd of the organisation, where most are high-achievers.
By the time Toner graduated she was already firmly enmeshed in Australia’s effective altruism scene. At its simplest, the movement is about ensuring individuals or charities do the most good with limited resources, by investing in antimalarial mosquito nets, for example.
Others, such as the FTX crypto fraudster Sam Bankman-Fried, can use it to rationalise or provide cover for doing immoral things or working immoral jobs in service of a noble goal. But there is another strand in the movement, which is popular with technology workers: preventing AI domination and using it for good. It’s that broad strand that Toner worked in.
After graduating Toner worked at effective altruist organisations GiveWell and Open Philanthropy – a project from Facebook co-founder Dustin Moskovitz and his wife – before moving to China to study that country’s AI industry. Then she joined the Georgetown Centre for Security and Emerging Technology as a director of strategy, working on AI safety. In late 2021, Toner joined the AI board.
“I strongly believe in the organisation’s aim of building AI for the benefit of all, and am honoured to have this opportunity to contribute to that mission,” she said at the time.
Our take: OpenAI is a strangely structured company with a flawed mission. It started out as a charity and is still overseen by the non-profit board that counts Toner among its members. That structure allowed the directors to blindside investors like Microsoft. It is clearly an unstable way of running a start-up, but investors knew that when they put money in. Sympathy for them should be low, especially given Microsoft is on the cusp of playing a blinder and recruiting hundreds of talented AI workers.
The same goes for OpenAI itself. Its approach to building a very powerful AI for good is too similar to those who insist the best approach to gun violence is to equip more “good people” with firearms.
As for Toner and her fellow board members, they appear to have been the rare directors to oust a powerful chief executive without the threat of public anger to compel them. In Silicon Valley especially, that could prove costly.