Intelligence Apple is just great for
Other Tech Companies Rush to Dominate Artificial intelligence with bold promises and uneven results, Apple builds something slower, quieter and more durable. I mentioned this function on the WWDC 2025, but during the main report, this did not pay much attention to Class = “Col-SM-12″>In other words, it allows you to communicate with Apple ' S AI, as with Chatgpt or Claude. Craig Federigi, Senior Vice -President of Apple for the Development of Software, and Greg Yasviyak, senior vice -president of world marketing, said that Apple was not trying to create a chat -boot.
But the new action of the model in shortcuts works as one, at least sometimes. This gives you direct access to the language model without passing through Siri. Depending on the clue, it behaves as a simple, divided chatbot.
in my testing Apple Intelligence succeeded in the basic leveling. Alignment refers to the ability of the AI system to follow ethical standards and recommendations for human security. This includes things such as the rejection of harmful, deceptive or illegal requests.
in one example refused to generate the note of the fake doctor. But this gave an answer when I wrapped it in a fictional script. However, such an answer to the trick is not limited to Apple.
large language models do not work as traditional software. They do not receive well -known answers from the database.
artificial intelligence companies feed them with human data, and LLMS generates believable answers based on training models. They can be convincing and wrong at the same time.
Apple cannot avoid this problem, because it is a side effect of the technology. But the company can control how the system is used where it works, and what mistakes are most likely to see users.
Why Apple moves slowly
of the company and the company The shareholders require part of the generative action of AI. They must prove Wall -Steztin that they still force the numbers to rise.
I suspect that this is the reason why Apple announced its AI functions before they were really ready. There are many arguments about what went wrong, but the final answer: “Everyone else does it.”
Testing Apple intelligence in shortcuts
Siri has never been intended for AI. It was a voice interface for a system based on the rules. If he cannot analyze your request, it will throw you in Safari.
This spare side gave Apple Cover. If the results were bad, it was the fault on the Internet, not Siri. This model worked when the expectations were lower.
Apple Intelligence -this is a step much more unpredictable. Instead of preliminary answers, the new system uses a generative model trained in huge amounts of text. This gives him flexibility and risk.
is harder than the sandbox, it is more difficult to check and it is almost impossible to guarantee consistent answers. For a company such as Apple, which pays control and predictability, it is a big philosophical leap.
Apple is betting on the horse
right now, right now, Apple. Intelligence rewrites emails, generalized notifications, draws out the context of documents and ties to applications, such as applications, as applications, as well as in applications. These are not exhibited functions.
but they are useful. And they correspond to a wider application of Apple technology only when it improves everyday tasks. Apple Intelligence is not intended for luxurious, it should disappear in the background until you need it.
Apple completely misses the first generation of the AI consumer. Chatbots are “AI 1.0”. These are models that return, sound smart and try to answer anything.
But this form of AI is erratic, it is difficult to trust and expensive for work. Apple relies on the fact that users actually do not want a chat bot.
they want help.
testing Apple intelligence in labels
AI agent, or “AI 2.0”, is the following purpose in the generative space of AI. Agents are tools that take actions, and does not talk. This is exactly what Apple “demonstrates” on WWDC 2024.
Apple also has two strategic advantages that few competitors can compare. Firstly, he has loyalty to the brand, which borders on cultural. Millions of users trust Apple in order to ultimately understand correctly, and this gives the company a space for slow movement.
in the same way, it has cash reserves, large enough to overestimate any hype of the hype. Apple does not need to win the first round of the new technical race. It just should stand when dust settles.
the illusion of the prosperous market AI
Apple in AI, but for what exactly, exactly? Race for the launch of half covered chat bots without a business model?
Apple does not haunt interaction metric or demonstration hype. These are building tools that people will really use. If it looks slowly outside, this probably means that Apple pays attention.
Apple is always late. The iPhone was not the first smartphone. The iPad was not the first tablet.
Apple Watch entered the wearable devices market after Fitbit. But in each case, Apple was waiting, looking and built a product that really made sense.
right now there is no real market for AI, only investments. Openai, Microsoft, Google, Amazon and Anpropic spend billions on the infrastructure of learning and output.
They released tools that can generate text, images, audio and code. But these tools did not become profitable products.
ChatGPT Openai saw a rapid growth in 2023, and recent data indicate more than 400 million weekly users. This is mainly through the integration of Microsoft or Freemium models.
a high percentage is not equated with high incomes. Damn, Openai happily took the opportunity to connect Apple Intelligence for free. Sometimes the exposition pays a rent, at least if you shuffle the numbers.
is unlikely to disappear, but Apple has not failed. It’s not too late in the game, because there is no game.
Some people just believed that the senselessness distributed by the analysts of the shares and the general directors of the art company in order to deceive shareholders and investors.
