AI, a curse or blessing for OpenAI?

Last week, it was the unveiling of the biggest rumor mill.

During Apple's Worldwide Developers Conference (WWDC), the company announced the introduction of AI to its latest devices.

In one fell swoop — your choice if you will call it genius or evil — 苹果 co-opted the term AI.

From Artificial Intelligence, it renamed it to Apple Intelligence. From now on, every corporate deck — think 谷歌 , 微软 , AWS, OpenAI and a myriad of others — will promote, for free, the new name, and every presenter will preface that when they mention AI, they are not talking about Apple AI, but, simply, AI.

Talk about going viral. It will, of course, affect every single search done using either Google or Bing. The irony is breathtaking.

Note: Apple is not the first one to do something like this. Microsoft did it with its collaboration product Teams, and now any product that includes any team capability is using the term to promote MS product.

Now, back to the rumor mill. It was expected that Apple would announce something about AI, but the question was, ‘How is Apple going to deliver on AI functionality without any known internal AI project?’?

The closest to AI functionality is Siri, which has capabilities that are frequently questioned or even ridiculed, together with Alexa by Amazon. The cliffhanger was which tech or company Apple would pick to plug the hole.

A few days before the conference, the internet started buzzing that the chosen one was OpenAI with its ChatGPT tech stack. And, sure enough, it happened.

On June 10, OpenAI announced it made a deal with Apple. I am sure that the celebration was quite exuberant at the OpenAI offices.

And now here comes the dose of reality. As Apple was announcing the new Apple Intelligence capabilities, media reported the company is not going to promote OpenAI. It is not even going to pay for the service. Apple, based on the demo, which starts around the 1:05 mark, hints at the upcoming AI features in its products.

Do they talk about OpenAI? Not really. In the press release, OpenAI is mentioned only once, at the bottom. What they are talking about is helping you with your writing, from grammar to summarization, — goodbye Grammarly, it was nice to know you — turning your mother's picture into a comic superhero portrait, or creating custom emojis called Genmoji. And then there is the integration with Siri. At least these are the plans, because the beta version is coming in fall...which means that for another year, you'll see nothing.

However, Apple did differentiate itself from the rest of the pack when it came to AI. Not only did it take over the term, but the presenter also spent a considerable amount of time talking about privacy and security. In major contrast with the Microsoft fiasco — which introduced the badly designed and hastily announced Recall — people at Apple talked about how data security is built into the product, starting with the hardware for communication and data processing in the Apple cloud.

With all the announcements, Apple was able to address the concern from the investors and market that it had no AI strategy.

Apple is not late to the AI game.

This company is still waiting for the technology to mature and is gleefully observing the blunders committed by others. It is also making sure that OpenAI won't survive and slowly bleeds to death. It will either run out of cash or will be absorbed into Microsoft.

Why?

Based on all the talk from Apple, ChatGPT is presented as one of many options, and as new products come to the market, Apple will unplug one and plug in a new one. If Apple decides that AI is really that important, it will build its own.

In the meantime, OpenAI will have to support the deployment of its technology to Apple. The AI functionality will be available only on the latest hardware with the latest OS.

To begin, we are talking upwards of 100 million users. That will not only require extra hardware, but a reliability (aka uptime), which, so far, OpenAI has had a hard time maintaining. Yes, there are times when OpenAI hits 100%, but mostly they are lucky to hit 99.9%, mostly just 99%.

I wonder if Apple will demand 99.99% or 99.999%.

That level of redundancy and operational excellence is far beyond what a startup can do and requires serious resources.

Considering that OpenAI is not profitable, has to keep building and training new models and has to prepare for its new deal with Apple, it will have to find more and more funds to support all this.

Will Microsoft keep giving OpenAI more money? How ironic that Microsoft is going to subsidize Apple on its road to build AI — Apple Intelligence.

The recurrent pattern in all this? Strategy. Without it, you just end up with the circus monkey.