I have used iOS since iPhone 4, I remember Googles voice assistant back then being much better than Siri is now, lol.
Why even call it Apple intelligence, if it’s mostly using OpenAI APIs?
Dumb Siri still can’t understand simple things.
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.
OpenAI Launches ChatGPT Agent That Can Complete Tasks For You
- Thread starter MacRumors
- Start date
- Sort by reaction score
Why even call it Apple intelligence, if it’s mostly using OpenAI APIs?
Dumb Siri still can’t understand simple things.
Yet even Siri can write grammatically.
My question is do I want this? It sounds pretty cool, but so does (a whole lot of things).
Last edited:
That’s you. There are all kinds of people. There are people with little time that could benefit from this or people that simply don’t want to bother researching, this is not that different from opening and following the instructions of a recipe book.
On win, macos or both?
OpenAI today launched ChatGPT agent, a new agentic model that is able to think proactively and complete computer-based tasks on the user's behalf.
The ChatGPT agent is in the same family as o3. It combines several existing ChatGPT features, and it can do things like research and generate reports, execute code using Terminal, generate slides and spreadsheets, and connect to external data sources and applications.
OpenAI gives several examples of how ChatGPT agent can be used:
The ChatGPT agent uses its own virtual computer, and it will navigate websites, filter results, prompt users to log into websites when needed, and deliver summaries of its findings. It is designed to seek permission before taking any "actions of consequence," and OpenAI says that users can interrupt tasks to add extra instructions, and stop tasks at any point.
- Look at my calendar and brief me on upcoming client meetings based on recent news.
- Plan and buy ingredients to make Japanese breakfast for four.
- Analyze three competitors and create a slide deck.
ChatGPT agent is rolling out starting today for Pro, Plus, and Team users. Just select "agent mode" from the dropdown menu in the composer during a conversation. ChatGPT users are able to transition between conversations and action requests within the same chat.
Pro users will get access by the end of today, while Plus and Team users will get access over the next few days. OpenAI plans to add the functionality for Enterprise and Education users in the coming weeks. Pro users have access to 400 messages per month, and other paid users will get 40 messages monthly with additional usage available through flexible credit-based options.
Article Link: OpenAI Launches ChatGPT Agent That Can Complete Tasks For You
Maybe but it's also a mouth piece for Musk and his political views taint anything it says.
AI experts shocked as Grok 4 checks Elon Musk’s posts before forming opinions
You're thinking way too narrowly. There are myriad use cases where 100% certainty isn't required.
Think about helping your kids with math homework. I have an advanced science degree, but every now and again I get foggy on a particular math concept when helping my kids (happens more than I'd like to admit 😅). Generative AI is awesome at refreshing my memory. And it helps me be a better tutor so that I'm preserving my kids' productive struggle and leading them to the answer rather than just giving to them.
Now imagine that power for parents who aren't as good a math as I am. AI has enabled them to connect with their kids in a way that no longer ends up in a frustrated, "I don't understand!"
And then realize that if a parent is working a late shift, Khan Academy's Khanmigo can do the same thing: a socratic tutoring session that doesn't let the kid of the hook and just give them the answer.
So, no. AI isn't going away. You just have to find ways for it help you achieve your unique goals.
Funny. Please would complain about how Google is mining your data. And, here ChapGPT will do the same. But, the company has already shown that it can't fully be trusted. No company that makes money can fully be trusted I guess. But, OpenAI is basically sucking all the data out there and using for monetary use like Google.
Those are all well and good if we pretend these things aren't enormous money sinkholes. There's no clear path to breaking even, let alone turning a profit, for any of these companies, which means they'll be around only as long as the people funding them can be tricked into thinking there's going to be a payoff at the end.
We have a hard push to integrate AI. But I’ll tell ya, I’ve tried and it fails most of the time. I literally just asked it to draw me a diagram comparing the sizes of double and queen size beds. 10 minutes later, 3 tries later, it errored out each time.
It hangs a ton.
I’m open to it all. But man, it needs some polish.
It hangs a ton.
I’m open to it all. But man, it needs some polish.
There is no official results for Grok 4 on Humanity's Last Exam. The reported result is 25.4% which is slightly higher than Google's Gemini (but Grok is newer). The higher Grok 4 scores are pointless. The test was designed for language models. Grok 4 Heavy is not a language model (although it includes LLM).
Which tool did you use? I just copy pasted your request to Gemin 2.5 pro and immediately got this:
For sure! ChatGPT helped me script wildcard SSL cert renewals for several TLDs I self-host on my NAS, one of which is a Lovable.dev-created site that auto-pulls updates from a private GitHub repository and deploys the updates to my live site (again, all set up by ChatGPT). It helped me configure my CloudFlare tunnel as well as Tailscale, so I can access everything securely without messing with a VPN.
I've found it incredibly useful and "super helpful" 🤣 for a ton of things, but one of the main ones has been NAS-related scripting.
Meanwhile Copilot can't create a slide deck and it's maintained by the same company that we all rely on for making slide decks
I was going to watch the video you posted, and my eye fell on one of the top comments: 'Please make clear in the video title that you didn't actually test the thing. I don't like wasting time.'
So I find it a bit ironic that you remind everyone that 'facts are stubborn things,' and then post a video without much substance, matey
Analyze three competitors and create a slide deck.
Always be careful with AI tasks like these. You never know how much it actually hallucinates and just writes out of its ass. Your documents also can't be too long, it will just stop reading it after X tockens but you won't even notice.
Ever notice how chatgpt never tells you "I don't know that"? It would rather tell you something that isn't true than nothing at all. It's dangerous for people that don't know how it actually works.
Always be careful with AI tasks like these. You never know how much it actually hallucinates and just writes out of its ass. Your documents also can't be too long, it will just stop reading it after X tockens but you won't even notice.
Ever notice how chatgpt never tells you "I don't know that"? It would rather tell you something that isn't true than nothing at all. It's dangerous for people that don't know how it actually works.
Last edited:
I too have been using Google's assistant for over 7 years and it was much better then than it is now, sadly. For some reason it is becoming increasingly dumb.
This is totally wrong.
Mathematics in education is one domain I am an expert in. It is definitely not suitable for that. It generates misleading methodologies, makes horrible mistakes and loses context very rapidly when structuring processes which require multiple steps.
The last thing you want teaching you is something that doesn’t know what it’s doing.
So please stop promoting this use case. It only looks like it works because you don’t know the domain well enough to know when it doesn’t work. And that’s dangerous - something authoritative sounding and wrong.
Very few areas of my existence require less than 100% correctness.
This is the fundamental problem. It doesn’t know when it doesn’t know something. The entire concept of an LLM is to give you a statistically probable bunch of tokens, not a reasoned or honest answer. Because this works for a lot of things people assume that it’s better than it is.
☣️ chatGPT will certainly not get access to my calendar 🔒 😝