Writing about AI again? I wrote about it a few weeks ago. (Chat GPT And Change) Why write about it again so soon? Things are developing that quickly.
I follow lots of topics. AI is one of them.
AI, now, feels like the internet in 1995. I’d been aware of Arpanet/Internet since 1983, then watched it boom and bust with everyone else around 2000. Yet, it didn’t really bust. Many companies were broken and bankrupted, but our usage of it grew from a curiosity to part of our lives to where some people spend most of their days for work, school, news, and socializing.
The adoption of the internet looks obvious in retrospect. But there was time when we had to Capitalize Internet because it was that novel. Now, the internet is lower case ‘i’.
Computers went through a military and industrial phase, but for most people they were a curiosity. PCs? Who needs one? They’re only good at home for managing recipes. Now, computers are so ubiquitous that they’re in our pockets. So much for those tower stations built into special rooms in upscale suburban houses.
AI is something that’s intrigued me for decades. I’m not an expert, but I know some.
That previous post, Chat GPT And Change, chronicled much of that history. Since then I’ve been respecting the possibility of even more rapid changes by more reading and researching. That work continues because 1) the technology continues to evolve, and 2) users are becoming aware of real rather than speculated realizations. I realized it made sense for me to update my thoughts because, as I read the work of others I realized I ignore source material that is more than three weeks old.
Three weeks, and that news from a month ago is already out of date.
Once upon a time there was a PBS show called Wall Street Week with Louis Rukeyser. Watching it after the market closed on Fridays was a regular event. After years of the familiar pace of topics and discussions, things started to change. Topics made no sense to some of the commentators. Guests were exasperated as they tried to explain the internet, which had been around for years, to financial experts. Guests talked about eyeballs, a measure of internet traffic and a land rush as people bought up domain names. It was obvious that some commentators thought the guests sounded silly. Selling books online? Why?
AI? Who needs it? Chat GPT? It’s only mimicry. Right? Sure, it is artificial; but is it intelligent?
Companies are already ‘hiring’ AIs to do simple things like write short articles (something I did six times a day for a national real estate news site, some of which could be handled automated), or generate stock images for ads, or…? The legal profession is realizing it can use AIs to sift through far more documents – and possibly summarize the results. These are a few of the real actions I’ve seen mentioned. There are undoubtedly far more.
PCs do more than sort recipes. The internet does more than deliver email. AI does more than play chess.
A random note I wrote to myself a week ago;
“Tonight’s ongoing thought exercise:
AI prior to 2022, a maturing curiosity
AI in 2022, surprisingly capable
AI in 2023, already being hired and used for real
AI in 2024,…?
Will we even recognize 2025’s AI?”
This week I finished the tenth book in my five-year photo essay of Whidbey Island’s natural shorelines, Twelve Months at Fort Ebey.
That seemingly disconnected accomplishment made me ask myself whether that kind of art would be usurped. Stock photos, particularly for generic topics, can be produced more readily, quickly, and cheaply with image-related AIs. My books are site-specific, and chronicle particular seasons; so they might be defended against such pressures. Fort Ebey is Fort Ebey. Faking it can be done, but fakes can’t honestly chronicle a literal snapshot of reality. Locals can appreciate that. But many of my favorite photos are abstracts. A landscape is a landmark, but a photo of the undulating colors of a sunset or a closeup of a flower aren’t as tied to a place; and could be indistinguishable from something artificially generated.
The larger effect is that AI has allowed recent hardware to out-perform my already limited camera from 15 years ago. Smartphone photos can be so much better because that phone really is much smarter than most folks realize. I pay a professional to polish my photos by removing dust spots, adjusting horizons so they are level, and color balancing and enhancing to compensate for sensor (and photographer) limitations. My old camera is now obsolete, at least compared to my phone’s camera.
Sitting beside me, waiting for me to finish writing this (after having lunch) is the first draft of the sequel to my sci-fi novel, Firewatcher. Ironically, the precipitating event in the book is humanity’s response to an accelerated growth in AI, and I started writing it several yeas ago. I guessed that the Digital Singularity might happen as early as 2040. Surprise! I’m confident that novel writing will remain humanly creative; but, friends are already using AI to generate characters, backstories, and to suggest things like plot lines. Of course, maybe AI won’t make great incursions into novels because novel writing is usually not profitable enough.
Writing short articles, or business correspondence, or ads – Sure, why not try AI? It doesn’t cost much for that first draft, and that first draft may be good enough. That first draft can also be ready in seconds, rather than an hour or longer.
I’m re-evaluating my career and business options and finding that my expectations are already changing. Fortunately, I enjoy meeting and being with people rather than paperwork and bureaucracy; but I’m not going to take any of it for granted. People like dealing with nice people; but AI robots are being designed for emotional responses, too.
A thought just came to mind. I’ve avoided talking about some topics because I don’t want to be seen as sensationalizing certain topics; but…
(saved to a separate file because the various social media AIs may pass judgment on the implications for a rather…personal industry that has been known for being an early adopter of new technologies.)
…”Honey, there’s a robot in a bikini here to see you. Did you order something extra with your pizza?”
And then there’s Chat-GPT 5, due soon. Ah, but there’s talk of a pause in AI development by CEOs and policy makers. And there’s scoffing because the people actually working in AI know that decisions made by people in offices mean nothing to curious developers. Ban it in one place? Don’t be surprised that some new business or mistake is unleashed in some basement or garage.
One of the fundamentals of the Digital Singularity is a increase in the increase in the speed of change.
“Tonight’s ongoing thought exercise:
AI prior to 2022, a maturing curiosity
AI in 2022, surprisingly capable
AI in 2023, already being hired and used for real
AI in 2024,…?
Will we even recognize 2025’s AI?”
Will we? Will we even recognize anything else in 2025? The only constant is change.
PS A flash: As I finished typing this, LinkedIn flashed an ad and a news item, “Recruiters are joining the AI race.” They’re already reading an estimated 75% of resumes from job applicants. How else are they going to use these AIs?
AI has a long history of “AI winters” when progress temporarily grinds to a crawl and investment $ drop precipitously. This time feels different. But what if it isn’t. What if development hits a computational or ethical limit just short of being generally useful rather than narrowly useful? If so, can ChatGPT (or any of a handful of other LLM implementations) devise a path forward?
Where is my flying car?
(sigh)