Skip to content
Apple · OpenAI · Apple IntelligenceEpisode 025 · 29 September 2024 · 43:41

Apple Controls the Device; OpenAI Controls the Expectation of Magic. Now They Are Moving Toward the Same Product

Central question

What happens when Apple, with its devices, and OpenAI, with the expectation of “magic,” begin building the same product?

What you take away

Test whether Apple and OpenAI become useful everyday interfaces or require constant correction. The decision requires the reader to check how many steps the interface actually removes and what dependency it creates in return.

Main threads

What to watch for

1Compare “ToTheMoon is a podcast and a hello from Silicon Valley!” with “Impressions of the o1 release”: they provide different criteria for judging the same issue.
2Test the conclusion from “The unreal quality of o1's work” in your own use case—what actually changes in the process and what remains a promise.
3Before choosing a product or approach, record the constraint identified in “The difference between Kindle Scribe and iPad”.
4Define the owner of the outcome and the quality metric for the situation described in “AI laws in California”.
Signals to track afterwards
→Watch for actions by Amazon and Apple that confirm or challenge the episode’s central claims.
→Compare new launches and policy changes with “The unreal quality of o1's work”: have access, quality, price, or constraints changed?
→Check whether the scenario in “AI laws in California” becomes repeatable practice rather than a one-off demonstration.
Most useful for
Executives and managersEntrepreneursProduct teamsAI usersProduct managersDesigners

Key takeaways

00:00Apple knows how to turn the purchase of a phone into an event

In the context of “ToTheMoon is a podcast and a hello from Silicon Valley!,” this criterion applies: Apple Intelligence still looks raw while o1 already feels like a qualitative leap; OpenAI lacks a device and a permanent place in the user's life, Apple lacks a model worth changing habits for — and the whole episode is built on that intersection.

01:18Who owns the outcome: thoughts on Apple Intelligence

The decision in “Thoughts on Apple Intelligence” depends on one criterion: the features arrive in beta, work heavily, and give no reason to say "this genuinely changed how I work" — which is especially noticeable against the ritual Apple builds around buying an iPhone.

06:11What changes in real work: using the Kindle Scribe

The discussion of “Using the Kindle Scribe” yields a practical test: one host returned the Kindle Scribe within a week — the device integrates with no outside infrastructure and locks you into Amazon's apps; another keeps it to save paper, but notes that go nowhere devalue the idea.

08:52Why context matters more than one metric: emotions from the trip to the main Apple

For the “Emotions from the trip to the main Apple store” scene, the decisive point is this: Apple turns buying a phone into an event — the trip to the flagship store, a new iPhone on launch day, half a day inside the ritual; that is a level of loyalty the company's software announcements have yet to reach.

10:31OpenAI has the opposite situation

The practical meaning of “Impressions of the o1 release” is that o1 impresses precisely through model quality, yet the interaction still starts with opening an app, choosing a mode, and phrasing a request — an impressive answer has not yet become an impressive product.

15:16The practical meaning of the issue: Jony Ive's work on a device for OpenAI

In the context of “Jony Ive's work on a device for OpenAI,” this criterion applies: Altman's work with Jony Ive on a dedicated device matters strategically more than another chat update: hardware with constant access to context would remove the need to keep opening an iPhone and give OpenAI a shot at Apple-level loyalty.

18:45That is why Sam Altman's work with Jony Ive on a dedicated device may matter strategically more than another chat update

The “The unreal quality of o1's work” issue should be assessed with one constraint in mind: on hard tasks there comes that moment when the technology feels almost unreal — yet the model does not see the user's whole life and shows almost no initiative, a limitation that further gains in answer quality cannot cure.

22:45What determines the outcome: expecting magic from OpenAI

In the context of “Expecting magic from OpenAI,” this criterion applies: the expectation of "magic" is really an expectation of initiative: users will be choosing not just a model but a new way of interacting with it, and the winner is whoever meets that expectation rather than merely improving answers.

32:20Apple still has an enormous advantage: notes, photos, email, calendars, watches, and payments are already inside its system

The “The difference between Kindle Scribe and iPad” scene leads to a working conclusion: notes, photos, mail, calendar, watch, and payments already live inside Apple's system, and even a Kindle Scribe or Evernote struggles against a tool built into familiar infrastructure — but that strength holds only while the features' quality does not fall visibly below the alternatives.

41:12The next stage of the race is the symbiosis of model and device

The “AI laws in California” topic becomes clearer once this point is included: Governor Gavin Newsom has thirty-eight AI bills awaiting signature — the signed count grew from eight to nine while the episode was being prepared: California is becoming both the most technological and the most regulated market at once.

What this episode is about

Apple Intelligence still looks unfinished, while o1 already creates the feeling of a qualitative leap. But OpenAI lacks a device and a permanent place in the user's life, while Apple lacks a model powerful enough to change behavior. That is why the idea of a dedicated device from Sam Altman and Jony Ive looks less like fantasy than the logical next step in the race.

Apple knows how to turn the purchase of a phone into an event. A person can travel to the flagship store, receive a new iPhone on launch day, buy a case, and spend half the day there—the company has built an entire ritual around the product. Apple Intelligence does not yet create the same feeling.

Features arrive in beta, work awkwardly, and provide no reason to say: this genuinely changed the way I work.

OpenAI has the opposite situation. o1 impresses through the quality of the model: on difficult tasks, there is that moment when the technology feels almost unreal. But the interaction still begins by opening an application, choosing a mode, and formulating a request. The model does not see the user's whole life and shows almost no initiative.

That is why Sam Altman's work with Jony Ive on a dedicated device may matter strategically more than another chat update. If OpenAI creates a device that has constant access to context and removes the need to keep opening an iPhone, the company has a chance to build Apple-level loyalty.

Users would be choosing not only a model, but a new way to interact with it.

Apple still has an enormous advantage: notes, photos, email, calendars, watches, and payments are already inside its system. Even Kindle Scribe or Evernote struggle to compete with a tool integrated into a familiar infrastructure. But that strength works only as long as the quality of the functions does not become visibly worse than the alternatives.

The next stage of the race is the symbiosis of model and device. Someone has to combine powerful intelligence, a convenient interface, security, and initiative.

Apple can integrate an outside model, OpenAI can build its own hardware, and Google already controls both Android and its models. The winner will not be the company with the most beautiful announcement, but the one people allow to remain beside them all the time.

The company with the most beautiful announcement is not enough. The advantage goes to the one people allow to remain beside them all the time.

Episode transcript

The episode is in Russian; below is an English reading guide to the transcript (the full EN transcript is a machine translation). Voice matching applied to 86 segments: 44 identified, 3 mixed, 26 probable, and 13 unresolved.

Read transcript on a separate page

Loading…