AI Hype and the End of Innovation

The following story is co-published with Freddie deBoer’s Substack.

At this point, my feelings on AI, or “AI,” are pretty plain. I think that these LLM systems will have some meaningful economic consequences, almost all bad, as well as social consequences, universally bad; some industries will prove to be susceptible to automation even if doing so entails people in power ignoring obvious inefficiencies and problems that come with turning to AI, and a lot of people are going to have whatever remaining ability they have to form meaningful human relationships destroyed. It’s not like there won’t be victims. But in general, I’m quite confident that the impact of these systems will fall vastly short of the relentless hype that our media simply will not stop engaging in, we will not see any of the repetitively predicted major revolutions in human existence (whether good or bad), and in the long run this type of AI technology will have significantly less impact on human life than the rise of the internet, which itself has not prompted anything like the change to ordinary human life that we’ve seen with advances like electrification, the internal combustion engine or germ theory.

As I’ve long argued, mundanity will win. Business as usual will win. We will not escape the bardo that contemporary late capitalism represents, not with technology anyway. Maybe with political revolution. But probably not.

Today, I have a pretty limited claim to make: this period of absurd AI hype has not emerged purely or even primarily from organic reactions to technological development but instead is largely the product of a deepening need, among our most prominent corporations, for exciting new products and services. In the early 21st century, the American economy came to be defined by its tech sector, in much the same way that the automotive industry had been a symbol of our greater success in the mid-20th century, as the finance sector became in the 1980s. “Silicon Valley” was a place but more than a place; its borders did indeed extend around large parts of the San Francisco Bay area but also up to Seattle and into Texas and sundry other locales, as a shared set of values — and valuations — trumped mere physical geography. Since the birth of the iPhone in 2007, which I would use to define the beginning of the modern era of consumer technology, the world of smartphones, laptops, apps and online networks have carried the torch for American innovation, at least in the popular imagination.

But progress has slowed. You won’t have to look far to find arguments that the madcap progress of the late 2000s to late 2010s has slowed dramatically in the tech industry. The smartphone, long a symbol of everything that made these companies hugely valuable, has entered into a commodity phase, with fewer and fewer obvious routes to meaningful development appearing open to manufacturers. Modern flagship smartphones now have ridiculously high-resolution, high refresh rate screens with searingly bright colors and remarkable contrast; their cameras have massive sensors, multiple purpose-built lenses, impressive low-light capabilities, and aggressive software processing to fix any lower-quality photos that are somehow still captured; processors and RAM combine with ever-more-efficient operating systems to give users buttery-smooth user experiences and as much power for apps as they could possibly need. While budget and mid-range phones lag behind, top-end development lifts all boats, such that you can buy a $250ish dollar Android phone that will give you 80% of the experience that a $1,000 iPhone can. Unsurprisingly, global sales have slowed.

This might sound all good for consumers, but many dedicated techies now routinely complain that there’s nothing fun in smartphones, nothing to look forward to. Given a choice between a camera with even more megapixels or one without, or a processor that’s got a few more flops squeezed out of it rather than less, of course you’d choose the more advance gizmo. But these things simply don’t inspire the same kind of breathless anticipation that once attended every annual release of the latest flagship phones. Meanwhile, the more that phone manufacturers struggle to differentiate between the last year’s model and the latest, the more consumers feel comfortable hanging on to their old phones, buying used, or picking up a mid-range model for a fraction of a price. This sounds healthy to me — I think my family had the same rotary home phone from the late 1970s until we finally got a cordless in 1990 or so — but it’s bad news for companies that have grown used to massive revenues and which have immense expenses that are not easily reduced. The iPhone has been as close to a license to print money as a tech company has ever enjoyed, allowing Apple to live like oil companies do for several decades. But what happens when it gets harder and harder to keep a straight face while you say that the latest model is even better?

This is a general conundrum for our tech industry. Folding phones are an interesting gimmick, but they are a gimmick, and I say that as someone who bought a Samsung foldable last year and mostly thinks it’s fine. I’ve reflected before on the fact that the iPhone was, at one point, sold as a potential “everything device,” and predictions that we would someday do all of our computing on our smartphones became unavoidable early last decade; amusingly, that imagined future gave way to a present in which the tech companies introduced the tablet and smartwatch form factors into our lives, in an effort to expand their product lines. (The Samsung Dex-style vision of plugging your phone into a monitor and keyboard, etc., at work and then at home, while still sort of cool, is less relevant in an era of near-limitless cloud storage.) The relentless, quixotic effort to make consumer VR a mass-market interest reflects the same fundamental anxiety. Ford introduced model years when it found that it had saturated the market with Model Ts such that nobody was buying any new ones; the gambit worked, but only to a point, and the industry has to proceed with the assumption that most people aren’t going to buy a new car more than once every decade or more. What does Silicon Valley look like when this becomes the norm for smartphones?

Well, among other things, it looks like Apple sweatily trying to sell its $3,500 VR glasses, which debuted last year to mountains of YouTuber and gadget blog hype and anemic sales. I will continue to argue that most human beings have no interest in wearing computers on their face, but either way, the promise of virtual reality as the next great moneymaker for the tech industry has been made for decades with precious little to show for it. Facebook Meta has staked its existence on the promise that we’ll all spend our lives in VR someday. But based on what actual consumer behavior?

There are plenty of other sectors of tech that we could consider. For a long time, web hosting served as the secret financial rocket fuel of the industry, with companies like Microsoft and Amazon seeing far better profit margins in that domain than in their more traditional offerings. That’s still a very good business to be in, but it’s now subject to intense competition, a saturated market and little possibility of the “hockey stick growth” that remains the elusive trophy for investors. The app economy is stagnant, as every possible idea for how to monetize software on a smartphone has been explored for more than a decade and with the big incumbent tech players demanding extortionate cuts of any income. Widespread disgust with social media has grown and grown, with the migraine-inducing experience of looking at Instagram for five minutes a good indicator of why — it’s impossible to see anything that you actually want to see, as viral bilge and AI slop is forced into your feed while the accounts you follow are almost impossible to find. Self-driving cars remain the future, but the market is broken up, the short-term profitability unclear and severe problems with serving bad-weather areas ongoing. The challenges to tech are considerable, and that’s just restricting ourselves to technological progress and profitability; ongoing public disillusionment with the industry has become as big of a story as the products it sells, and many of the field’s biggest names have responded to the existence of even moderate criticism by withdrawing into a reactionary shell.

And so now you’ve got AI, which is a story that has been as relentlessly, shamelessly and irresponsibly hyped as any media narrative has been in my lifetime, with the exception of the threat of terrorism following 9/11. Tech needs AI to be everything that the press is credulously, uncritically insisting it will be. Apple, which is facing its most serious problems since bankruptcy was a serious concern decades ago, needs AI to sell iPhones, which is particularly concerning given that their own efforts in that domain have badly lagged behind those of competitors. Again, you already know how I feel; these AI affordances are incredibly lame when compared to the hype, automating certain trivial tasks out from under adults who could already do them, saving spare minutes here and there that users will then later spend watching endless hours of TikTok alone in their sad, dark rooms. But even if you have a far less jaundiced view of this technology than I do, the profitability question looms large. What exactly is the product that’s being sold here? Who will pay for it when so much of it is already available for free? Can any company survive the squabble for market share? And what guarantee is there that any of this will actually make sound financial sense, given the absolute ludicrous amount of computing power and electricity that are required to keep modern LLMs going?

Of course, the stock market has never demanded much in the way of basic fiscal sense from tech companies, and Wall Street is sure to float the major players in this war for a very long time, in search of transcendent profitability. Tech has for decades essentially operated in a … postmodern space when it comes to revenue, expenses and profits. Tesla has long been the poster child of a company with a paper valuation that has little at all to do with its financial fundamentals. Still, you would like to believe that eventually there’s going to have be some sort of come-to-Jesus moment with this technology, that at some point an ambitious hedge fund is going to start shorting these companies in a bid to force them to prove their value. When and whether that happens, though, this is the point: the hype is a phenomenon driven by needs that are fundamentally financial in origin. The tech companies need a new suite of products that can restore their eroding profitability and inspire the public the way that the public was inspired in the late 2000s and early 2010s; the financial sector and investors need the tech companies to be the unicorn stocks that they once were. As usual with speculative capitalism, the tail is wagging the dog. When hockey stick growth does not emerge naturally from reality, it will be invented.

Media hype about AI, the almost literal absence of any countervailing narrative, the relentless way that the stodgiest publications inflate the threats/hopes that have been invested in these technologies … it’s all downstream of the desire to feel about tech the way people did in the Obama era and the demand from the moneyed to have the right lottery tickets to buy. A lot of people who are used to getting what they want are looking to let the good times roll again, and “AI” is precisely the kind of vague instrument into which they can throw those hopes. Anyone who points out that the emperor has no clothes is simply told that they don’t understand the technology, and every year that goes by that human life is not seriously disrupted by this technology is just a minor delay. Rinse and repeat.

When you look at your average media AI enthusiast who spools out hype and refuses to engage with skepticism — which is to say, almost everyone in media — it’s not like they’re taking part in what I’m describing intentionally. Instead, they’re in a profession where there is no greater fear than the fear of being owned, and this story has contrived to make them feel like they’ll be owned if they undersell their predictions; on this issue there’s every incentive to hype and no incentive to be careful. They are moved by investor demand the way most people in our culture are, as if by animal spirits. And they’re largely millennials or Gen Xers who were adults for the heady Obama tech days, who remember the fun of joining Instagram when it actually showed you pictures of your friends, the excitement of waiting in line for the iPhone 3G, the rush of discovering a cool app that your friends hadn’t yet and then bragging about it on Facebook. They want those days back again. And with everyone and their brother promising that AI will rescue us from our mundane disappointments forever, well, why not give in?