are we building for a future where we don't exist

Living in a man-machine interlude.

Another day, another agent. 

AI nurses, AI receptionists, AI engineers, AI doctors and even AI girlfriends. 

Every day, we hear stories of an AI-automated future that feels… human-less?

Though that future still remains quite unclear, many of the smartest minds seem to believe that we exist in a brief phase transition. This transition has been marked by an industrious frenzy of internet philosophers and startup founders looking to claim the new intellectual territory that AI has created. 

Despite being amongst the beneficiaries of this technology-driven prophecy, I resonate with the many who have a creeping feeling that we are making ourselves obsolete. 

A grim thought, but the implications have sat on my lap since “AI” has become humanity’s favorite prefix. Whether it takes the shape of fully autonomous cars or world-class mathematical reasoning ability, the general theme across all media suggests that we may just be biding time before a new generation of intelligence becomes the dominating force on the planet. AI beat top humans in chess in 1997, Go in 2016, and even poker in 2017. This quick rise to dominance, which took place in two decades, makes for a compelling argument that machines will continue to beat expert humans in new domains. 

This idea foreshadows unfamiliar territory for our species. In our entire history, humans have never experienced intelligence even remotely near our own. In fact, we have frequently differentiated ourselves from the entire rest of the animal kingdom on the premise of being a higher “intelligence”. What happens when/if we are eclipsed?

I don’t feel it to be productive to spend time estimating exactly when or if this will happen, but I feel it’s important to address the world that we live in before we get there. 

As of today, AI has proven to be fantastic at certain things like assisting programmers, but we are still plagued by many of the same problems of the past century. Despite sugar-coated TechCrunch articles and Linkedin headlines, I have yet to see major progress of “AI-agents” making meaningful progress on something other than being a receptionist. But this doesn’t stop dozens of companies from appearing out of nowhere to announce yet another anthropomorphized OpenAI (or I guess Anthropic lately) API call. 

I don’t blame them. To sell the idea that they will be able to dodge the powerful stomp of a fortune-5 backed research lab (or steam-rolled by a koenigsegg), requires them to sell an imaginative vision. It’s not original to sell software tools to doctors, people have been doing that for 25 years. What is original is to replace the Doctor outright. This wouldn’t be too much of a problem if it was only part of the VC pitch, but this promise often becomes inseparable with the mission and spirit of the company. The pitfalls of pitching with this methodology are already beginning to show and pose a major concern for many capital allocators. But I’ll leave that problem to be addressed by someone who cares more about the economics of technology investing. 

What I do care about are the things that brought us here in the first place: tools (technology) built for humans. I think it's necessary to criticize the distortive idea that we may be nearing the end of an age where human intelligence is a primary driver of innovation. This outlook is being masterfully sold by the leaders of top research labs, and has rapidly spread amongst a certain well-connected tech/startup population. While this strategy is effective for unifying engineers and researchers around a lofty mission, I fear that the trickle down effect of these ideas are harmful to the broader technological ecosystem. If companies are built predicated on the mission of replacing “X” rather than building for “X”, we may miss out on reaching the height of our human-computer-symbiosis potential.  

I will be able to rest easy, knowing the machines have taken over, when a human makes no difference in an examination of a patient, research effort, or creative act. In the same way that the top chess player in the world would offer no help to Stockfish, I believe that our inability to contribute to the work of a machine will mark a point where I will be able to raise a white flag and retire to mimosas on the beach. 

But that isn’t the case now, and I optimistically (or I suppose, pessimistically) believe that our path towards an omniscient super-intelligence will be paved by at least another hundred thousand displays of human ingenuity. 

AI has undoubtedly unlocked incredible potential for new types of human-centered tooling but we can’t let the “AI-automation” hype deter us from building the beautiful, functional tools that extend our very own capabilities. Maybe the robots will do better science, write better books and make better memes, but not today.