15 AI Ongoings Unfolding Right Now
A five-minute update from Scotland
As many of you probably suspect, I love writing essays. I feel I’m not a good writer yet so I try to overlap my efforts to improve with my aim to provide you valuable insights. Sometimes, however, I step too deep into self-indulgence.
I write about things that interest me that don’t fulfill the promise of the newsletter, or go cryptic with acronyms and statistics new readers won’t follow, or turn up too much my critique-o-meter, or choose a style that’s not easily consumable in the age of goldfish attention spans (I speak for myself).
This isn’t an apology—I like to write for myself because otherwise I wouldn’t write at all—but a thank you for bearing with me through my more selfish times.
So instead of an essay that’s hard to parse (or too contrarian), this is something I think we can all find valuable—a short, mostly neutral list of things I see unfolding in AI right now (no hype or anti-hype).
It makes up a comprehensive bird’s eye view at multiple levels: industry trends, practical applications, current research, regulatory developments, and behind-the-scenes insights that aren’t widely publicized (If you want me to dive deeper on any of these, let me know in the comments).
No more ado.
Soon everyone will be using voice assistants, even those who profoundly dislike generative AI (with perhaps the exception of privacy obsessives). People resist new tech because it deforms the status quo (see Ted Chiang’s New Yorker essay) but eventually adapt to it free of emotion—what’s useful stays and AI voice assistants are hella useful.
Generative AI hype in the tech space is over. Generative AI hype in investor circles is over. This doesn’t mean generative AI is over. Bubbles pop, investors lose money but the world grows as a result. What comes next is the silent grinding phase; people making practical sense of this tech we’ve built. Enterprises will slowly integrate generative AI into their work processes (if they manage to do so reliably). This phase will be decisive in answering the big question: Can generative AI grow the economy and increase productivity across jobs?
Don’t listen to AI CEOs when they make promises that sound too good to be true. Not because they’re lying to you but rather because you’re not the target audience of their pitch decks disguised as interviews and predictions. No one knows what’s coming with certainty but they have to sell a specific story merely to have a chance to make it happen. Don’t buy what’s not being offered to you.
AI agents are coming. One year. Maybe five. Forget everything you think you know about agents if your sources of knowledge are the past two years of attempts with babyAGI, autoGPT, Devin, and the recently announced Sakana AI scientist (they don’t quite work). AI agents needed new technology (or rather new techniques to make the tech work) and research seems to have borne fruit. I can’t say more yet.