or i could go the musk route and be born into wealth and never actually make anything 🤩
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Elon Musk Threatens to Ban Apple Devices Over ChatGPT Integration
- Thread starter MacRumors
- Start date
- Sort by reaction score
What Musk is doing is known as spreading fear, uncertainty, and doubt about a near-future competitor, now that Tesla is an AI and robotics company, and not a car company. Classic.
When’s the last self landing reusable since the shuttle? First private also.
Say what you want, Elon has spearheaded some amazing stuff.
My work has started rolling out some strict AI rules. Won’t be surprised to see app controls next.
The question I keep asking myself is what is OpenAi/MS getting out of this?
It ain’t a freebie …
Yeah agreed on that point…but Tesla does run its own OS and controls what’s apps run on it. Even Elon removed Disney when he started to disagree with them.
THis coming from the guy who sucks up a ton of data, including video and audio, from your car if you buy a Tesla, and can even update/change/kill your car with automated over the air updates.
But they haven’t “baked ChatGPT in”. It is an external service connected like any web service would be with an api and contracts. It is very much like the web search deals in Safari in that regard. Easy to swap providers, allow choices, or even build your own if that makes more sense.
Starship - game changer
Falcon V - game changer. Some boosters have 20+ starts and returns
Dragon vs Starliner
And you know that NASA is outsourcing a lot of flights to SpaceX, right? Guess why
I think he pretends to but the reality is that he was right place right time to write a dinky little payment processor and lucked himself into a fortune.
He doesn't care about free speech. He bans plenty of accounts that talk **** about him. He cares about particular kinds of speech. And influencing the electorate.
ChatGPT is not really integrated into Apple OS platforms. It is an optional add-on.
Apple’s own AI provides more than enough safe assistance for most users.
The nonsense of ChatGPT and friends writing a book for users with limited writing skills is dishonesty piled on top of terrible quality.
AI research cannot be trusted. Cross checking and vetting are essential. AI writing is redundant and shallow, lacking heart. AI can proofread only if it understands your style. Otherwise it is just doing 1980s spellchecking.
ChatGPT data suffers from a mix of accuracy heavily sprinkled with error. Ask the chatbot about its data sources. It replies Internet, plus other things. Can you trust Internet sources without vetting? Nope. Press AI on topics where you are an expert and get ready for confidentially delivered mistakes.
The best use of chatbots is brainstorming. AI does well at high-level outlining, but suffers when you get to details.
Apple’s own AI provides more than enough safe assistance for most users.
The nonsense of ChatGPT and friends writing a book for users with limited writing skills is dishonesty piled on top of terrible quality.
AI research cannot be trusted. Cross checking and vetting are essential. AI writing is redundant and shallow, lacking heart. AI can proofread only if it understands your style. Otherwise it is just doing 1980s spellchecking.
ChatGPT data suffers from a mix of accuracy heavily sprinkled with error. Ask the chatbot about its data sources. It replies Internet, plus other things. Can you trust Internet sources without vetting? Nope. Press AI on topics where you are an expert and get ready for confidentially delivered mistakes.
The best use of chatbots is brainstorming. AI does well at high-level outlining, but suffers when you get to details.
I’m curious of Lord Elon’s feeling on Microsoft, and their system-level integration of CoPilot AI in their new OS and on their laptops.
I guess the pasty boy wonder will need to build his own computer network and OS completely from scratch. Or, is he just going to use XTwitter for everything?
What a tool.
Of course they're not and neither are Apple though.
It's unclear at this point who paid who though. It could be a case of OpenAI paying to be the default LLM when Gemini, Claude etc are added later on, the same way Google pays to be the search engine.
Or it could be Apple paying OpenAI so that users can get an unlimited free use of the ChatGPT 4o for questions from Apple devices (I'm not sure we're going to see the coding and image side, at least they didn't demonstrate sending images for inspection)
I suspect it'll pale in insignificance to using the ChatGPT app when the full vision model is released later in the year and you have fast natural conversations with it. I imagine OpenAIs plan is to provide more value to the premium package (which you can link to iOS) and hope to get more subscribers from it (which they probably will)
Is google search scanning every email you have sent, every text you have sent? Scanning pictures, and basically everything in your phone so AI can use it when you ask what time your mom's flight is coming in? Or what restaurant was I at when I was wearing the black leather jacket on that rainy day picture? Is google scanning your entire phone and computer so it can answer any question relative to you? Maybe it is, then I would agree, but I have trouble believing this has been the case that a search engine has the same level of access that you are giving AI.
One thing about AI algorithms is that the more they’re used, the better they perform. There may not be any payment by anyone. Apple will provide anonymized, user blinded interactions with open AI learning models. In return, open AI will likely see a tenfold or more increase in model use.
It could be that this is a symbiotic relationship with no financial component.
Last edited:
LLM's don't get better the more they're used, it's not an AI algorithm like that. Their performance is determined by the data they were trained on and the architecture of the model itself.
Unfortunately not. There are diminishing returns when it comes to machine learning.
The models have to be retrained. The algos have to be fixed manually by engineers when they reach their limits and keep making the same errors.
Software engineers find it hard to fix bugs in a 3D video game so don't ever expect some kind of consistently perfect generative AI. It would require super human levels of engineering and so much compute power that it would consume a lot of the energy of all the superpower countries if over a billion people used it every day.
You may be right but exactly how it is linked in is not shared nor why. Apple has a plan that they are not sharing. OpenAI/ChatGTP is a mess. I am surprised Apple went this route instead of doing one of their own.
I believe as part of their agreement with OpenAI that user requests are siloed in an anonymous bucket and the bucket self destructs after the user has finished. They have open sourced their methods for transparency so we will know how it works soon.
Imagine if we didn’t have SpaceX.
With the current tension between the US and Russia we would have to rely on whom? Amazon?
