Apple Fix: Just Buy Something That Mostly Works
July 4, 2025
No smart software involved. Just an addled dinobaby.
A year ago Apple announced AI which means, of course, Apple Intelligence. Well, Apple was “held back”. In 2025, the powerful innovation machine made the iPhone and Macs look a bit like the Windows see-through motif. Okay.
I read “Apple Reportedly Has a Secret Plan to Quickly Gain Ground in the AI Race.” I won’t point out that if information is circulating AND appears in an article, that information is not secret. It is public relations and marketing output. Second, forget the split infinitive. Since few recognize that datum is singular and data is plural or that the word none is singular, I won’t mention it. Obviously few “real” journalists care.
Now to the write up. In my opinion, the big secret revealed and analyzed is …
Sources report that the company is giving serious consideration to bidding for the startup Perplexity AI, which would allow it to transplant a chunk of expertise and ready-made technology into Apple Park and leapfrog many of the obstacles it currently faces. Perplexity runs an AI-powered search engine which can already perform the contextual tricks which Apple advertised ahead of the iPhone 16 launch but hasn’t yet managed to build into Siri.
Analysis of this “secret” is a bit underwhelming. Here’s the paragraph that is supposed to make sense of this non-secret secret:
Historically, Apple has been wary of large acquisitions, whereas rivals, such as Facebook (buying WhatsApp for $22 billion) and Google (acquiring cloud security platform Wiz for $32 billion), have spent big to scoop up companies. It could be a mark of how worried Apple is about the AI situation that it’s considering such a major and out-of-character move. But after a year of headaches and obstacles, it also could pay off in a big way.
Okay, but what about Google acquiring Motorola? What about Microsoft’s clever purchase of Nokia? And there are other examples. Big companies buying other companies can work out or fizzle. Where is Dodgeball now? Orkut?
The actual issue strikes me as Apple’s failure to recognize that smart software — whether it works particularly well or not — was a marketing pony to ride in the technical circus. Microsoft got the message, and it seems that the marketing play triggered Google. But the tie up seems to be under a bit of stress as of June 2025.
Another problem is that buying AI requires that the purchaser manage the operation, ensure continued innovation of an order slightly more demanding that imitating a Windows interface, and getting the wizard huskies to remain hooked to the dog sled.
What seems to be taking place is a division of the smart software world into three sectors:
- Companies that “do” large language models; for example, Google, OpenAI, and others
- Companies that “wrap” large language models and generate start ups that are presented as AI but are interfaces
- Companies that “integrate” or “glue on” AI to an existing service, platform, or system.
Apple failed at number 1. It hasn’t invented anything in the AI world. (I think I learned about Siri in a Stanford Research Institute presentation many, many years ago. (No, it did not work particularly well even in the demo.)
Apple is not too good at wrapping anything. Safari doesn’t wrap. Safari blazes its own weird trail which is okay for those who love Apple software. For someone like me, I find it annoying.
Apple has demonstrated that it could not “glue on” SIRI.
Okay, Apple has not scored a home run with either approach one, two, or three.
Thus, the analysis, in my opinion, is that Apple like some other outfits now realize smart software — whether it is 100 percent reliable — continues to generate buzz. The task for Apple, therefore, is to figure out how to convert whatever it does into buzz. Skip the cost of invention. Sidestep wrapping AI and look for “partners” who do what department stores in the 1950s: Wrap my holiday gifts. And, three, try to make “glue on” work.
Net net: Will Apple undertake auto de fe and see the light?
Stephen E Arnold, July 4, 2025
Comments
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
