The AI Paradigm Shift

The 6.5 Billion Dollar Man.

If you had any doubt that “this time it’s different,” and AI’s existential threat to humanity wasn’t enough, perhaps OpenAI’s $6.5b acquisition of Jony Ive(‘s startup) might convince you. Sorry, make that $76.5b including AAPL’s stock drop.

Unlike every platform shift—PCs, Internet, Mobile—this AI thing is a true paradigm shift.

It ushers in a new era of unpredictable disruption and riches. Every boardroom is scurrying, investor is analyzing, and founder is building. Whether by blind watchmaking or visionary engineering, nobody knows how this will play out.

Here are a few observations shaping my thinking…I’d love to hear yours:

→ AI favors the incumbents. Unlike previous tech platform shifts that created walled gardens (think semis, search, social, mobile), the core AI tech is available for low to no cost via API. No excuses, dinosaurs, you have the means to adapt and the competitive advantages to win. Wake up, Apple, the call is coming from inside the house!

→ Yes, software is dead. So is digital advertising. So are websites. So is search. So is content creation. So is …  The barriers to creation are demolished. Consumer behavior is already changing (ehhem, first quarter since 2002 that Google search volume decreased on Apple devices!) Innovator‘s dilemmas abound, wince now because the bandaid is about to get ripped off.

→ Digital Deluge. Our chimp brains can’t handle the deluge of content coming. Every book becomes a movie. Every email becomes a tome. Every whim becomes an article. AI has become a runaway modem, modulating our dribble into prose. The demodulation can’t come fast enough. (No, I didn’t use AI to write this post. Any of them. Except for images… where I do have some fun with GPT and Pixelmator.)

→ Data Defensibility. There’s a reason why OpenAI’s BD team is raining money on proprietary data owners. The models are quickly commoditizing. Diminishing returns on training. Feature parity is always one release away. Firewalled enterprise data, copyrighted libraries, and personal information are the finest ingredients — it is what it eats.

→ It’s a Trap! Investing in AI is really hard. The foundation model owners have to make big app-layer grabs (heeeere’s Jony!), but incumbents won’t go gentle into the night. The abundant AI tool providers have to roll up or risk getting rolled over. Emerging AI apps risk drowning in the 500lbs gorillas’ backwash. You either have to have an incredible GTM edge or, yes, proprietary data.

Who knows to what degree, and how fast, any of this is true, but whoa, strap in and grip the bar. I personally believe it’s a matter of when, not if.