The AI Promise Gap

Why, exactly, are chatbots doing sycophancy, shopping, and social media?

God has let me down lately. In his creation, the earth filled with the pleasant smells of rotting mulch and joyful sounds of baby giggles, there is beauty and majesty and potential. Then, the bills come. Boston Children’s Hospital—$14,500. 60,000 mile car service—$958. Groceries are up 23.6 percent in price from 2020-2024. There is a gap, an ugly crevice, separating the promise of the world and the commercial reality of the world as I experience it.

So too, there is a promise gap between the creation and commercialization of technology. The launch of new science is this kinetic bundle of potential. However, as that tech deploys, there are second and third order effects that never live up to that hope. This is a problem. Perhaps the problem is inherent to the system in which it exists: America’s brand of moralistic capitalism has always promised profits and moral superiority all at once. The stories we sell, the divine visions of founders we worship, tell us we can build technology to make humanity better, all while we make boatloads of money off selling it.

Take the messy, beautiful internet. You’ve likely heard—or felt—both sides of what I’m saying. Online shopping is better than going to Walmart, and customers flock to Instagram and TikTok because they’re better than anything else they could spend their time on. But also, social media has gone from connecting friends to stealing people’s time. Between all the ads, it’s hard to know what’s actually a good product in online shopping. Google search has become an information swamp where you need to wear rain boots to muck through the SEO mind mud.

The techno-optimist in me can’t help but ask: what if we just fixed it? What would it look like for the next world-changing technology to fully live up to its early promise?

I’m asking this now because of AI (duh). I’m an ardent believer in capitalism, but also, I’m not a dummy. An AI that lives up to its potential will create a beautiful, wonderful world where rote labor is automated away. But if we aren’t careful, the cutthroat nature of startup development will mean the most popular uses of the tech won’t benefit humanity. They’ll just benefit shareholders. And no one should want AI to be distributed the same way as the internet has—through venture capital dollars that inevitably go to attention marketplaces.

If we do not learn the lessons from the internet’s history, we are doomed to repeat them with AI.

But this has already begun. As I’ve written about previously, many of the most popular consumer applications are porn and artificial emotional relationships. AI is also being used as a rationale (real or not) for layoffs by corporations. This is a sinful, duplicitous, a bummer of unrighteous proportions.

It doesn’t have to be like this. There are opportunities to build startups that are just as big as the internet giants, but in a way that actually benefits humanity. And, if we do this right, there are opportunities to make large amounts of money. (As I said, I’m a capitalist baby). Here are the three large markets for disruption where I think divinely great companies can be built.