I don’t mind having more of the bed to be on when Dawn is gone, and I’ve stopped feeling guilty about it. I spread out like a starfish and sink into a deep sleep. But when the clocks toll downstairs I wake and my mind gets to thinking.

I really shouldn’t care about AI ruining humanity. I mean I should, but what’s the point of that? It’s been almost a year since OpenAI released ChatGPT and funny, all this drama about their CEO Sam Altman being fired. It’s probably good timing for the media since they needed a boost on the AI storyline.

Microsoft, which owns a 49% share in the company, got one minute’s notice before the board ousted Altman last Friday. One minute! Satya Nadella, Microsoft’s CEO, had just shared the stage with Altman a week prior. Microsoft’s stock fell with the news of Altman’s ousting. And now there is talk that it may have backfired, and instead Altman could return as CEO and fire the board.

While the board of directors claimed “he was not consistently candid in his communications,” the real reason for his ousting is not clear. The company’s unusual structure gives its board full governance rights over investors, including Microsoft, and it’s been speculated the reason for the rift is ideological between board members and Altman. It’s possible they don’t see eye to eye over the risk of releasing more powerful versions of ChatGPT.

But the thing is, Microsoft has invested billions of dollars in the company and has used OpenAI’s technology (and PR) to launch new products this year, including Microsoft Copilot, which has somehow managed to make Microsoft cool again. Some have speculated adding ChatGPT to Microsoft’s virtually unknown search engine Bing could be the beginning of the end for Google’s hold on the search market, one that Microsoft owns a measly sliver of.

So you can bet that Friday was not a good day over at Microsoft with the news of OpenAI’s drama. And if Altman is reinstalled, this won’t happen again: because Microsoft will insist that OpenAI get a handle on its board.

I went down the AI rabbit hole hard this past year. ChatGPT was released on my birthday last year, late November, just as the big tech companies had started a round of layoffs that continued well into the spring. It was hard not to associate the layoffs with fear over AI replacing the workforce, and a bit of existential fog over what to do with my life.

I read books and white papers on AI, took online AI literacy courses, listened to podcasts. One of the most interesting books I read was called The Alignment Problem, which follows the history of AI from the 1950s to the present. If you worry about artificial general intelligence, the sci-fi type of AI worry, you fear the AI becoming smarter than humans and wanting to destroy us. If you follow that line of reasoning, the way to avoid that is by giving AI human values so that it won’t harm us. You align the AI’s values to ours.

Now if you really want to go down a rabbit hole, consider how you would train an AI on human values: and whose values you’d pick, and then how you’d give it the ability to adjust, and be nuanced about its values as we humans are. I still don’t understand how large language models or neural networks work, despite how hard I’ve tried. I sure as hell don’t understand how assigning human values would.

If we’re fortunate, we get to decide how we’re going to feel about the world. It’s a kind of privilege to educate oneself and dither over the fate of humanity. I spread out in our clean, wide bed and drift in and out of consciousness. Today I’ll kick off a new project with Microsoft, that’s based on OpenAI’s generative AI. I am well paid and gainfully employed, and the world will not end today. I’ll make a good buck by the end of it.

Categories: Technology, writing

Tags: Creative Nonfiction, existentialism