AI: Apple, OpenAI and fence-swinging strategies

...self-driving cars and AGI have similarities & a hoped for difference

The Bigger Picture, March 3 2024

New technologies always trigger large ambitions in companies big and small. Makes them truly swing for the fence. I’ve written at length about how self-driving cars may be the ‘Canary in the AI coal-mine’ in the current AI Tech Wave. I outlined how a decade plus vision of fully self-driving cars (L5 in industry parlance), took billions and years longer than originally presumed.

And it may be just as far out of reach for a while, leading Apple to shut down their self-driving car project Titan. Current ambitions by key AI players like OpenAI and others for AGI or Artificial General Intelligence, may be as far off if not longer. There are a lot of similarities, and one critical difference. That’s the ‘Bigger Picture’ in AI I’d like to unpack this Sunday.

But first a flashback to a decade ago, to 2014, when one of the pre-eminent tech companies of our time, Apple, thought they could outdo Tesla and race ahead to a full self-driving car. As Mark Gurman of Bloomberg explains in a follow-up piece to his recent piece on Apple’s decision to shut down their self-driving car project Titan explains:

“It all comes down to how the company set out to build a car. When Apple began work roughly a decade ago, there were two main schools of thought about how to proceed:

1. “Build a less ambitious electric vehicle with autonomy features in line with models from Tesla Inc. That means the car could drive itself on freeways and some roads but not everywhere.”

2. “Change the world with a full-blown self-driving vehicle, taking passengers from point A to point B with zero intervention from a driver. And make it look like nothing anyone had seen before.

   

   “Of course, Apple picked the second approach, and that was the issue. All those years ago, the company thought it could solve a problem (full self-driving) that the auto industry is still struggling to crack today — all while introducing a groundbreaking design. That challenge proved too difficult for even Apple to overcome.”

“Now, you can understand Apple’s hubris around 2014. The company had just upended smartphones, tablet computing and music playback. And it was about to launch the Apple Watch, which it thought would crush the Swiss watch market.”

If one substitutes AGI for self=driving cars today in 2024, and OpenAI’s AI industry leading positioni today, it’s not a stretch to see the comparison between Apple’s confidence in 2014, and OpenAI’s in 2024, having changed the decades-long quest for ‘AI’ with ChatGPT, GPT 3 and now 4 since November 2024. And founder/CEO Sam Altman’s confidence in talking about raising trillions for AI chip and power infrastructure.

With the exponential growth in technical capabilities for Foundation LLM AI models expected by OpenAI, Anthropic, Inflection and many top AI companies, it’s not a stretch again to see the confidence of the AI industry today to the confidence of tech companies in 2014 to achieve truly Level 5 self deriving cars in a few short years.

Those similarities are palpable. A potential difference this time with AI vs self-driving cars MAY BE that fast-improving AI GPU & data center hardware infrastructure, combined with exponential growth in LLM AI models coming next, may get AI to succeed where self-driving cars failed.

The industry currently hopes that next generation LLM AI models like OpenAI’s GPT 5 soon and presumably GPT 6 in a few short years, are expected to truly leap-frog the impressive capabilities of the AI models like OpenAI’s GPT 4, Google’s Gemini Ultra and others of today.

Elon Musk and some investors are counting on next generation versions of Tesla’s Dojo Supercomputer to make all the difference in BOTH self-driving cars and AI concurrently. So his Tesla, X/Twitter ‘Grok’, and xAI ventures.

Whether these hardware and software improvements will get us to what’s been deemed ‘AGI’, is the question of the day. My suspicion is that tech hardware and software generally always takes longer than expected. And this time may not be any different, despite the exponential leaps in performance expected in just the next three years.

That may make all the difference, or it may be a road to disappointment, like self-driving cars for now. Remains to be seen. But the top AI companies today are nevertheless swinging for the fences. That is the Bigger Picture in AI this Sunday, as we read about Apple’s ambition when feeling top of the world a decade ago. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)