Apple, of course, had a lot to show off on Monday. It was like introducing an exciting new sports car. They had to show off the style: how the revolutionary fender design cut the airflow, how the revolutionary paint protects from the sun and vastly improves the mileage. Plus, the breathtaking acceleration and speed. And don’t forget the competition-beating, impossible mileage.
There was no way the audience could absorb the miracle that hovers under the hood. This will take many more hopeful presentations.
Monday’s Apple Intelligence presentation, for example, was a vast understatement. The real under-the-hood impact is this: ChatGPT, or any new LLMs, do wonders when responding to a query prompt. However, Apple demonstrated an LLM that responded to intricate prompts encompassing many unique elements—more complicated than any average person could create on their own without much trial and error. The real Apple Intelligence in action is understanding and knowing how to do this by responding to a simple user request.
No one understood that this was beyond the capabilities of a simple ChatGPT. I don’t think a copilot is that capable. This took way more than a “frenzied jump-in.” It took collecting all of the thousands of Apple’s real intelligence workforce and their years of collective knowledge, slip-ups, and all.
But the real sleeper is this: Apple shares gained seven percent on its user AI announcement, whereas Nvidia shares gained 10 percent on its hardware AI announcement. No one seemed to have caught Apple’s mention that its own super-private, personal LLM server was built on Apple silicon!
Take that, Nvidia. What will happen to Apple shares when investors realize that Apple, known for its industry-leading silicon designs, has silicon that can out-compute Nvidia’s? What, then, would happen if Apple made these servers available to the whole world?
I’m not sure even the Apple tech followers understand the potential of Apple’s subtle announcement. Hey everyone – now is the time to be kind to Apple for once.
Please leave a comment …
Add a comment