Sam Altman’s AI Hype Is Familiar

I think Sam Altman is too young to have been influenced by Rolling Stone magazine, but I feel like he’s learned something about retconning previously expressed hype in order to make room for fresh amazement from how Rolling Stone treated the new albums of the 1980s and 1990s by the Rolling Stones.

By that time, the Stones had established themselves as permanent rock royalty, but their music was undeniably less vital than their late-’60s, early-’70s heyday that produced all-time great work. Music tastes had changed, Mick and Keith were less interested in and less capable of breaking new ground, and so the work understandably suffered next to albums like Exile on Main Street or Sticky Fingers.

Not according to Rolling Stone, which could be relied on to wax rhapsodically about whatever the boys had produced upon the album’s release, declaring it a return to greatness after a previous fallow period. Unfortunately, you can only return to greatness once, so when the next album would arrive, they had to retroactively downgrade the previous album that had been dubbed a near masterpiece.

In 1983, Kurt Loder declared that Undercover “reassembles, in the manner of mature masters of every art, familiar elements into exciting new forms,” giving the album four and a half stars.

Undercover had one minor hit, “Undercover of the Night,” which sounds like second-rate Duran Duran, and has Charlie Watts playing electronic drums, an absolute offense against all that is good and holy. The idea that it is a near-perfect album is, literally, insane.

We move forward to 1989 and the Steel Wheels album, also given four and a half stars, this time by Anthony DeCurtis. The review opens with “Nothing reinvigorates Sixties icons like having something to prove. In the past few years the reverence typically shown both the Rolling Stones and Bob Dylan has worn perilously thin. The Stones’ last two albums, Undercover and Dirty Work—not to mention Mick Jagger’s solo recordings—ranged from bad to ordinary” (emphasis mine).

It gets better. In 1994, Barbara O’Dair declared in her review of Voodoo Lounge, “Gone are the smooth moves, trend nods and lackluster songcraft of Dirty Work and Steel Wheels, the Rolling Stones’ last two studio discs. The band’s new album, Voodoo Lounge, is ragged and glorious, reveling in the quintessential rock & roll the Stones marked as their own some 30 years ago.”

But this time it’s true, the Rolling Stones really are back!

The popular explanation for all these rave reviews upon a new album’s release is that Rolling Stone editor Jann Wenner and Stones lead singer Mick Jagger were close friends. But even that friendship could not stand up against the fact that over time, it became clear that these albums were duds, and so each review had to retroactively throw the previous effort under the bus.

Recently, on the platform I will only ever call Twitter, Sam Altman declared, “We trained a new model that is good at creative writing (not sure yet how/when it will get released). This is the first time I have been really struck by something written by AI.”

This is a strange statement, given that Altman has been relentlessly hyping this technology since its first public appearance in 2022, expressing personal marvel at its smarts, its empathy and now its creativity. One would think he’s been struck repeatedly by what his models produce, but apparently not—this is the first time.

Note that this model is not yet available for public consumption, so we cannot judge for ourselves if it is “good” at creative writing, except I am totally going to judge whether or not it is good at creative writing and say it isn’t.

Despite being well established in the skeptic camp about this technology, I think anyone who reads More Than Words: How to Think About Writing in the Age of AI would come away seeing that I am quite open to experimentation and exploration of this technology where it has the potential to enhance, as opposed to substitute for, human capacities.

But “creative writing” is clearly not an area in which large language models will excel, because I will go to my grave believing that the whole point of writing creatively is to attempt to capture the artistic intention of a single unique intelligence and then to share that intention with other unique intelligences. This challenge, which I have wrestled with over many, many hours of my life, is difficult, fascinating and very much worth doing even if the product of that wrestling never sees the light of day beyond the audience of the original author, which is something I’ve experienced rather often in my career.

Large language models are not unique intelligences. They are highly sophisticated, technologically amazing pattern-matching machines that generate syntax as their outputs. There is no intention behind this generation, therefore there is no creativity at work. It is not writing, not as I understand it, and not as I value it.

I know lots of people who are willing to argue about these things who will say that we’re in the midst of a “new” intelligence, blah blah blah. I’m happy for other people to wrestle with these thought experiments, but I know for a fact that the human experiences of reading and writing the creative work of other unique human intelligences is worth doing no matter what this technology—that cannot and never will work from an intentional place—is capable of.

Look, I imagine some of my frustration is starting to leak through, and I do not wish to outright dismiss those with other perspectives, though I wonder about folks who are not capable of seeing past Altman’s relentless hucksterism by now.

The thing is, thanks to More Than Words being in the world and having the opportunity to talk to lots of different people in lots of different contexts about what I have to say about writing in a world where large language models exist, it’s increasingly clear to me that in many cases, no one is asking for this stuff.

If no one is asking for it, we certainly have no responsibility to give it the time of day when it does arrive just because it’s shiny, new or amazing at the surface level.

The future is ours, not AI’s.