https://twitter.com/munster_gene/status/1786179406831988909
![]()
Gene Munster,
Investors knew…
Traders knew…




Data
ChatGPT doesn’t train on your data. Apple can offer on-the-fly training on selected data or trained your data passively to facilitate inference later (something like indexing the Mac to speed up searching later).
Agents
ChatGPT doesn’t have access to applications on iPhone. You can authorize apps and data to be accessed by Apple agents.
On prompt then act standalone apps like ChatGPT
is not an elegant implementation. What you want is while doing your task, gen Ai is activated without you knowing. That’s how ML AI and Spotlight are implemented on Macs and iOS devices. Early implementation of iTunes hide browsing activity from users… in similar vein, you won’t know gen AI is activated when using Apple apps…

I think should be deeply integrated like Spotlight, that is won’t be necessary to use “Hey Siri” to activate Siri to do something. An agent that can accept any inputs, not just voice, and can be activated by an app, not just by a human.
Q4, 2023. Trimmed 10 million.
Q1, 2024. Trimmed 13%.
![]()
An investor asks a specific question and WB talks about principles and willingness to pay tax?

Upgrade always drive sales e.g. Q1 Mac sales increased yoy. iPad is inherently robust in addition is well built by Apple. Increasingly, light usage consumers are migrating to iPad (with a keyboard cum cover) from laptops. Stable sale is ok so long can increase margin.
Contrary to AAPL haters and anti-Apple mainstream media narrative, AAPL leads in rolling 5-year return.
.
WSJ repeats MacRumors’ rumor…
Rumor below is typically a hallucination of Rivian investors.
Apple, Rivian Reportedly In Talks For Partnership
First M4 Apple product is an iPad Pro
together with the new Pencil Pro and a magic keyboard is ideal for creative professionals and mobile employees e.g. salesmen and insurance agents. Personally, I prefer a MBA, currently WMT is selling M1 MBA for $700 which is sufficient for many consumers.
Make sense to skip M3 for iPad since M4 draws less power than M3 for same performance.
Another obvious reason is M4 has an improved neural engine making it more suitable for Gen AI/LLM inference.
Wondering aloud (1): NVDA designs GPU for both graphics and AI acceleration. Why did Apple choose to make a neural engine specifically for AI acceleration?
Wondering aloud (2): Current PC design use separate chips for CPU, GPU and chipsets. While Apple silicon combines CPU, GPU, neural engine and chipsets into one silicon. Apparently is for higher performance per watt but would it mean Apple chip won’t be as high performance as NVDA for AI acceleration? AFAIK, performance of CPU of Apple silicon is higher than Intel and AMD, and performance of GPU is on par with NVDA.
Wondering aloud (3): Is obvious that Vision Pro 2 will be using M4 chip. What about the cheaper Vision? Which one will Apple release first, Vision Pro 2 or Vision?
Nvidia chips include tensor core as well.
Apple to Power AI Tools With In-House Server Chips This Year
Deploying this year? That mean has been developing for awhile.
Apple Nears Deal With OpenAI to Put ChatGPT on iPhone
Nears? Means not cast in stone. Bloomberg is so eager to create eyeballs.
AI is getting fun. Co-opetition at play. Companies are busy cooperating and competing with each other.

About $12 below Dec 2023’s ATH.
https://twitter.com/sama/status/1790068157404569608
Apple already has a damn good real time translator installed automatically for iPhone since 2020. This new ChatGPT 4o is an improvement over that. Is ChatGPT 4o using the same backend?
















