2023

By the time spring started I hadn’t worked for four months. There’d been a short gig in December but prior to that, nothing since June. The agency I contract with slows down in the summer as their primary client Microsoft slows down too, having run so hard for so long they just need a break. It normally picks up in the fall but that year the economy teetered, ChatGPT appeared, big tech companies issued mass layoffs, yet everyone claimed the layoffs and ChatGPT release were unrelated. For the first time it looked like technology could replace white collar workers, even artists.

Looking back I kind of freaked out, judging by my writing and frenetic behavior. I was set on this idea I could write for The Atlantic, for example. That could be my ticket out. Or that I could become fluent in AI if I just read enough books and white papers. So I consumed every bit of news and information I could, even fancied myself a kind of machine. I actually corrected a guy online about the difference between foundational and NLP models (and I was right). Worst of all, I read a book about existential threats to humanity and got sucked into the singularity, the AGI debate. A dark place of eyeless creatures. I got a virtual ticket to see the technologist Geoffrey Hinton give his first talk after leaving Google on grounds he was concerned for the fate of humanity (Hinton, one of the three “godfathers of AI”). In the end I understood very little and worried quite a lot.

I wrote my AI manifesto and nobody gave a damn. I found an AI mentor on Vashon Island and ferried over on Saturdays to talk shop. He was older and worked in tech, a coder geek, long ponytail, thinning hair. Sober 40 years. Lived alone and built analogue synthesizer systems with complex wiring and displays of twinkling lights and sounds, indefinite loop. He said you really don’t need to worry about ChatGPT being sentient, there’s no “there” there. And then he gave me some sources I could trust in the untrustworthy Silicon Valley.

In May I got more work again, to write sales decks for Microsoft on generative AI. How it could help replace people in call centers (without saying that), how it could help white collar workers save time. I used the tech myself to fill in content gaps on slides or to summarize transcripts. It worked perfectly.

In the last two years the AI narrative has shifted from generative to agentic AI and now everyone is trying to untangle what’s meant by agents and what it will mean for people. I’m actually getting paid to do that right now, to market agents: the spectrum of agents, from task retrieval to semi- or fully autonomous. I started down the path of building my own agents but didn’t have anything for them to do other than generate content but the thing is, I like generating content. It’s one of the things I like about being human.

I got a gig ghostwriting for Microsoft, a series on how generative and agentic AI could help retail executives. I used Copilot to help me write one of the blogs and yes, it saved time but no, I’m not proud of the output. By contrast, I labored my way through the last blog and it’s my best.

I think we often use tech to help remove friction or inconveniences but there are times the friction yields a better result. My work (content marketing) is on the chopping block. Could there really be a utopic future where people don’t need to work and use machines to serve them? Is that a real utopia?

Categories: Corporate America, Technology

Tags: Creative Nonfiction, existentialism, Technology