Let’s Not Let What Happened to Web3 Happen to AI
This article originally ran in Fortune.
It seems like everyone in the tech industry is marveling at the opportunity in AI. And rightfully so, the technology has a magical quality to it, surprising even the most skeptical users, with potential to reshape industries and jobs.
But this stage we’re at — one marked by effusive enthusiasm and a blistering pace of change — is also somewhat of a precarious one. The promise of the technology is undeniable, but the long-term, real-life utility is still unclear. Existing AI products, like Siri or Alexa, have failed to scale beyond simple use cases. Humans are notoriously impatient with computers when they make a mistake. And, publicly-traded AI-enhanced solutions, like StitchFix, are valued very differently than private AI startups with nascent traction.
This stage we’re at isn’t dissimilar to when crypto rebranded as Web3 and gained enormous buzz in 2020. The promise of democratized finance was (and still could be) enormous, but somewhere along the way the focus shifted from enduring applications to buzz-fueled by coin speculation. The conversation shifted from early adopter utility to early adopter riches, with too many solutions searching for a problem. We’re seeing something similar now in AI with less questions around the factual quality of results versus the human-like tone of the results. As much as I find it fascinating to read that an AI chatbot would like to be alive (and in love), I hope the capital & talent moving into the ecosystem understand that mass adoption can only be achieved first through micro-utility.
To contextualize this, the evolution of mobile technology feels like a better analogy. Early mobile apps had a similar toy-like quality where new capabilities sat at the forefront of the experience. The Lightsaber app was powered by the accelerometer, Bubbles took advantage of the touchscreen and Shazam leveraged the microphone. But, once developers really honed in on the advances of GPS and the Camera, mobile startups started offering transformational services that had historically been out of reach for consumers. The camera was obviously essential to create the social products of today, but it also is critical for some serious operational use cases, like two-factor security and document scanning. The accuracy of GPS chips ultimately gave way to some of the biggest businesses capitalizing on mobile as a platform: Uber, DoorDash, Google Maps. What will be the equivalent in AI that helps this new technology go from a toy to an essential part of daily life?
When it comes to AI, we at Forerunner believe the equivalent to GPS-chips and quality cameras will be the computer’s ability to understand nuance in intent and emotion. As more consumers are turning to digital solutions for their daily needs, understanding these needs better will be a requirement. Just like Uber enabled everyone to have a private driver, something historically too expensive for most people, AI will be able to take human services (financial, travel, mental health, and coaching) and bring these to a broader population with accessible pricing. Some of the most interesting companies to us are the ones who are offering these services today at Uber-black like prices, but are uniquely positioned to ride the cost curve down to offer more UberX like pricing. The open question is whether the service quality holds up in much more critical categories versus the quality of UberX has over time, which has declined significantly in reliability and accessibility of price of time.
Ultimately, the opportunity in AI lies in leveraging the technology to serve the customer vs. being an "AI Company" as the end goal. The elephant in the room is that people don't want to talk with a computer when they need something done. Why? Because the magic of a computer understanding your intent doesn’t eclipse the frustration of it missing the point. That won't change anytime soon. We think some of the best AI companies will be the ones where the customers don’t even know AI is being leveraged. This is already happening in the crypto world, where real businesses are leveraging blockchain technology where needed…and, dare I say, not even talking about it!
Our industry has a tendency to indulge the highs and the lows, so it’s worth recognizing the less glamorous necessities: the builders who are working through the data quality issues, integration challenges, and the unique domain knowledge requirements to make AI experiences reliable and repeatable. These are the folks who will end up building enduring, strong businesses — ones where AI is a means to an end, vs. the end point itself.
Changes won’t come overnight, but sectors core to society’s quality of life, like transportation or travel, or essential to equality, like healthcare or our legal system, have the opportunity to dramatically change. Technology has removed so much humanity out of experiences, it is somewhat ironic that AI, and its ability to interpret people’s intent and underlying feelings, might be the answer to bringing some of it back.