[Column] Michiel Frackers: Five conclusions after the chaos at OpenAI

A few days after the royal drama at OpenAI, let's try to look past the ruins of this company whose brains, Chief Scientist Ilya Sutskever , have said it could herald the downfall of the world. Because it is too often forgotten that Sutskever and his colleague Jan Leike, also not a pancake , published this text in July on the official blog of OpenAI:

Ilya Sutskever and Jan Leike, OpenAI: “Superintelligence will be the most impactful technology humanity has ever invented and could help us solve many of the world's most important problems. But the immense power of superintelligence can also be very dangerous and could lead to the powerlessness of humanity or even to human extinction.”

And oh yeah, they have over $10 billion in the bank to figure out whether they're going to save the world or kill it. Yet few serious attempts are made to seriously monitor, let alone regulate, OpenAI and its competitors.

Imagine that Boeing develops a new plane with a similar promotional text: 'this super-fast plane will fly on dirt-cheap organic pea soup and could help humanity make aviation accessible to everyone, but it could also fail and crash and blow up the earth .' Then the chance of a permit does not seem very great. This is different with AI; the tech bros just put the website live and see how it goes.

No problem, great turkey

What is the media reporting about at OpenAI now? About the Thanksgiving dinner at which reinstated CEO Sam Altman sat down with Adam D'Angelo, one of the commissioners who had fired him six days earlier. Both tweeted afterwards that they had had a great time together. The treasures.

Despite the tendency of the media to quickly fall into choosing a hero and a villain in conflicts, more and more frayed edges are observed in the acclaimed Sam Altman. According to the Washington Post, Altman's firing had little to do with a disagreement over AI safety, as first reported, but more to do with his tendency to tell only part of the truth while trying to line his pockets left and right.

Even though they are apple green

Altman is now back with a new board about which there are many doubts . Christopher Manning, the director of Stanford's AI Lab, noted that none of the board members have knowledge of AI: “OpenAI's newly formed board is likely not yet complete. Nevertheless, the current composition of the board, without someone with in-depth knowledge about responsible use of AI in human society and consisting only of white men, is not a promising start for such an important and influential AI company.”

It doesn't matter to me what color or gender they are, even if they are apple green with three knuckles, but I do like it if they understand the matter that, according to our own experts, has the potential to put humanity to the sword.

Five conclusions after the chaos

The AI ​​war has been won by America.

After a week of craziness and fuss, look at OpenAI and we see that Microsoft, the old board and the new board are all Americans. The competitors? Amazon, Google, Meta, Anthropic, you name them: Americans. The rest of the world is watching and holding rallies and speeches, but the race is running.

Good management is nice, but bad management is disastrous.

By that I don't mean that the people who fired Altman were right or wrong, because no one still knows; the core of their argument was that Altman had not provided full disclosure and if that is true, that remains a cardinal sin.

But the root of the problem was deeper. The OpenAI board was appointed to ensure the mission of the OpenAI foundation, which, in short, was to develop AI to create a better world. Not to create maximum shareholder value, as has now become the goal. The problem arose from those conflicting objectives.

Twitter, or X as it is now known, remains the only relevant social network in a crisis.

Elon Musk went on a rampage again last week and it apparently cost him $75 billion in sales, but Altman and everyone else involved still chose X as a platform to tell their story. Not Threads or TikTok – although I would have liked to see this mud fight for power depicted in dance.

Microsoft wint.

I already thought Microsoft was a funny name under Bill Gates, because the company was neither small nor weak, but in the almost 10 years under CEO Satya Nadella, Microsoft has grown into a dominant party in all kinds of markets.

While Amazon, Google, Meta and also Apple struggle to formulate an AI strategy, Microsoft seems to have found a winning formula: it is investing heavily in OpenAI, which uses the Microsoft Azure cloud, allowing a large part of the investment flows back to Microsoft. In the meantime, Microsoft is enjoying the increase in value through its 49% interest in OpenAI.

AI needs to be tested and probably regulated

Precisely because companies such as Microsoft, Google, Meta and Amazon also dominate in the field of AI, the development of AI must be carefully monitored by governments. The years of privacy violations, disinformation and abuse of power that take place via social media, for example, show that these companies cannot regulate themselves.

The tech bros' motto remains unchanged: move fast & break things. But let them do that with their own planet, not with the current one. The potential impact of AI on the world is simply too great to let the often socially limited minds that run tech companies make the choices for society.

An initiative such as the AI-Verify Foundation can be a vehicle for achieving the responsible introduction of AI applications. I'll close with the same quote as last week from OpenAI's Chief Scientist Ilya Sutskever, which shows that world leaders in AI almost hope that future AI systems will have compassion for humanity:

“The bottom line is that eventually AI systems will become very, very, very capable and powerful. We will not be able to understand them. They will be much smarter than us. By then it is absolutely crucial that they feel towards us the way we feel towards our babies.”

Michiel Frackers is the Chairman of Bluenote and Chairman of Blue City Solutions

www.bluenote.world
www.bluecity.solutions

[You can receive this newsletter free of charge in your mailbox every Sunday morning, sign up here for it]