Embracing the AI Platform

Here are two things I believe about AI and why I am optimistic.

My underlying assumption is that AI is a platform. It is not a person. It is not itself a worker. It isn't itself an application. It is a platform ingredient and from that the world will build more tools. Some tools are already widespread including the ubiquitous chatbot. Some of those tools will be general purpose and some highly specialized. That's the spectrum of applications.

Like many 'first' tools, what we see today is not an endpoint by itself. In fact in almost all platforms the very first tools or example use case is not the one that builds an enduring paradigm shift. In the lingo of tool v. Paradigm, one might even think AI is itself more of a paradigm and the tool is mass-scale learning and models.

By analogy, the first tool for the microcomputer platform was a BASIC interpreter. Many could not see how this would bring about a future like The Jetsons or a future where we all had to type programs with line numbers. It was the programs built with BASIC. And then a lot of computer science that found better ways to program that then led to whole new uses beyond Adventure and ways to store recipes. The Internet platform followed a similar pattern. The first network based tools of FTP and email were great, but for a very small set of people. Using those tools led to HTTP/HTML. That led to a graphical browser. Then we got shopping. Then phones. From that first mail message or FTP in 1971 to the 1991 HTTP to the 1994 browser w/SSL was 25 years of excitement, hype, doubts, and a lot of predictions that turned out to be pretty wildly wrong. But here we are.

Thinking about it as a tool helps me arrive at these two thoughts:

Embracing the platform and tools is key. Once you concede that AI is a tool for humans to use it becomes much easier to see that those who embrace tools will themselves demonstrate varying levels of proficiency, productivity, and creativity. There were photographers that resisted changes like auto focus or digital, and many were fine doing that. But those that embraced the new 'technology' were also the first to capture whole new types of images. They were the first to expand what photography could do from a baseline of manual focus, silver halide image capture. Suddenly spot news, art, fashion, and more saw an explosion in new art forms, new expressions, and a wild expansion of how images could be captured and used. The same happened with something as mundane as word processing. As much as there was consternation about 'everyone has to type' or 'what about the typing pool' the reality has been a vast explosion in creativity and output by those that embraced the new way of communicating.

Explosion of apps is next. Chat and generating words is insanely cool. Hallucinations and weird grammar and more are probably not so great. If we just stop there it is easy to be critical of AI or claim 'this won't work' or worse 'jobs are replaced by computers that do a worse job.' This is the world of cable tv and telco operators explaining the first HTTP/HTML could never implement their version of the digital highway. This too is a static view of the world. It says the technology today is the endpoint. What comes next will be an explosion in creativity in applications. Some will work. Some will not. Every day companies big and small are looking at the underlying platforms of AI and thinking that there are whole new ways to solve problems that go beyond 'automating the way we currently solve them.' This is really the Cambrian moment for a new platform. The first internet way to learn about, read, or buy things looked like an 'automation' of how we'd use an almanac, read a magazine, or shop a store catalog. The innovations in all of these led not just to the kind of disruption that people today love to dwell on when thinking negative but to the wondrous innovation in how all of those work.

These points are about how new tools create new baselines of creativity and possibility. The first is about the individual potential of humans. A static view of the world only thinks about how a new tools just makes the old thing easier. So a banker who uses Excel to simply 'calculate' portfolio returns in 1983 might make marginally fewer mistakes, one that looked at a spreadsheet and instead analyzed many other allocations, sensitivities and more did a vastly improved job with much less effort. That seemed insanely cool and new and was scary to many people. Then one day, that new way is simply how every portfolio manager works and what everyone comes to expect. This new baseline is not a negative. It expands the space. It is all opportunity to be seized. The work expanded because of the new tool.

The second point is about the broader potential of the economy and systems. Many people today didn't experience or can hardly remember the 'old way' of doing things and that was just a blink of an eye ago. It is not lost on me how many people are using the internet itself with smartphones to rally to defeat new data centers or to ban the use of AI in schools. The only reason these campaigns came to be in such short order is because of data centers and the internet. There are so many broken things in so many areas from healthcare to education to just the basics of how government works where new AI applications can so obviously improve not just individual tasks and productivity but the whole system.

It is normal now when expressing any sense of optimism to also try to temper such thoughts with concerns of disruption and more. There is not a single person who lacks empathy for changes taking place, least of all those building technology. I spent my first summer with a real tech job setting up PCs and teaching civilian executives and military officers how to use the new PC that I was told was provided to them to make their jobs easier. That was a memorable and formative experience. Some were scared or threatened. But many were excited. Yet all of them knew and expressed that there was uncertainty for their specific work how the PC would change things. I was 20. I had no idea. It was weird to have that conversation.

I was just recently chatting with my first boss from that summer job setting up computers. He was a mainframe programmer who worked on the moon landing!! We reminisced about those very first 'IBM PCs' and the small group of people of various experience called to drive the change. There were seasoned full-time employees that were mainframe architects, COBOL programmers, and system operators all of whom had never seen PCs, C programming, or even held a floppy disc. It was a wild time of change.

I'm on the other side of this now. I understand it. I'm also optimistic about the technology as much as I am for how there is a path.

PS: The line is fuzzy between tools and apps, and always has been. Just as the line between platform and tool can be fuzzy. You've seen that historic caricature of Microsoft organization. Now you know why.

PPS: No AI was used in this post. I just typed it.