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.
Apple Continues Losing AI Experts to Meta
- Thread starter MacRumors
- Start date
- Sort by reaction score
Must admit as a member of the public I have not seen how it has benefited me. We are all being told AI is going to be the biggest thing and that it is going to revolutionize everything. All I've seen it do at the moment is being used to let students cheat on their exams, be used to create images of non existent females so they can appear on magazines, create music from non existent musicians and be used in many aspects of the adult entertainment industry.
In addition to having a better net margin (~39% vs Apple's ~24%), Meta has fewer products over which to spread their R&D spend and doesn't rely on hardware sales for the vast bulk of their revenue.
Agreed. This is on par with Google search. Apple never built their own search engine. They built the platform to use any search engine. I suspect one of the AI companies will realize this and make a deal. The catch is how private the data will be.
That is a list that is skewed by popular media click-bait-level news. As a counter, consider that AI is already having a huge impact on all STEM fields.
Sundar Pinchai has said that 30% of new code written at Google is AI generated. Microsoft has reported similar numbers. Zuckerberg claims within a year, 50% of Meta's code will be AI generated.
AI is being used by professional Mathematicians for proof correctness verification, conjecture generation, proof strategies, counterexample generation, symbolic computation, high-order PDE solving, etc.
The creators of DeepMind won the 2024 Nobel Prize in Chemistry for revolutionizing the study of protein folding using AI. DeepMind's AlphaFold2 predicted the structure of over 200 million proteins, essentially cataloging the protein structures for all known life forms.
In medicine and pharmaceuticals, it is being used for drug discovery and medical image analysis. Over the next 10 years, the impact of AI on biology and medicine will be transformational.
I'm sure someone else can fill you in on its use in more prosaic problems in marketing and commerce.
But, yeah, you can also use it to create pron.
Please tell me you're a current Facebook employee complaining about others getting paid more, while also simultaneously wasting company time by replying on MacRumors lol that would be epic!
I asked Gemini on my Fold 7 what type of grass I have in our new rental property. She said "Send me a photo so I can take a look" sent her a photo, and she immediately described what it was, how to care for it, etc.....did the same on my iPhone.....Siri "I'm sorry I can't do that, here are a few links to what grass is"....It's absurd the biggest tech company in the world is so far behind on this technology....but wait.....for our next "innovation" we'll add another button!!!
They'll bake it into their products and their users will use it to do simple stuff like edit images in Instagram to using search queries in their main app as well as a load of other things.
AI is just software & a hardware backbone at the end of the day. At much as I hate Meta, at least they are serious about AI and their own in-house AI development. Apple are just a complete laughing stock with their woeful pathetic attempts so far. They are so far behind they really need to just partner with companies who know what they are doing because Apple certainly doesn't.
Apple have lost the plot with software in the last 10 years and are behind with almost everything they offer. This is what happens when the likes of Apple & Microsoft are run by useless bean counters with no vision.
Don't bother. The narrative has set in, and truth no longer matters.
Maybe people at Meta know more than random commenters on Macrumors...
Last edited:
No, not exactly. Apple’s internal AI testing only got to an accuracy of around 80% and they wanted it to be higher. They don’t want things to be routinely “hallucinating” (AI marketing speak for errors), so they opted not to roll out something that wasn’t quite to spec. The usual suspects all claimed Apple was doomed because they didn’t squat and squeeze out something system wide that was as unreliable as the news summaries.
I’d personally have a delayed product (or no product) than something that doesn’t work 20%+ of the time. People who want that jank are welcome to download ChatGPT.
But how is all of that going to benefit me? because the creators of AI keep telling everyone how it is going to revolutionize the world. When the word 'world' is used in ANY AI documentation, that means me, a member of the public. If AI is not benefiting the public right now then it is very disingenuous of AI creators to say it will help everyone/help the world. Is it going to help me get to work quicker? Is it going to help me cook my meals quicker? is it going to do all my shopping for me and get it sent to me safely? is it going to help me be more productive at work?
If it's use is currently limited then AI creators should restrict there optimism on where and how it will be used.
1. Yes. Machine learning in traffic control is on its way (https://www.nature.com/articles/s41598-025-00762-4; https://www.isarsoft.com/article/ai-in-traffic-management)
2. Maybe. There are appliances (microwave ovens, ovens, stoves, etc.) that have basic sensing now that is like baby machine learning. Others are in development to implement machine learning to improve cooking.
3. OpenAI's new agent is the start of that: https://openai.com/index/introducing-chatgpt-agent/. I haven't tried it out for shopping yet (not likely to in the near future), but I was using it the other day to help me update and reorganize some material for a couple of my university courses. It saved me hours of rather tedious work.
4. Will they help you? It depends on what your job is. Because of LLMs like GPT, Gemini, etc. I save dozens of hours of work every month or so. One of the ways is through statistical analyses. I do a lot of statistical analyses in my job. Rather than having to generate all code myself and dig through stack exchange and various websites to figure out how to code a specific analysis, I can use something like Copilot to generate R code for me. Is the code accurate? It takes some back and forth at times to get it to do what I need, but I have a colleague independently verify key analyses with other statistical software and the results have always been the same. Within a few hours I can generate hundreds to thousands of lines of code to run analyses, generate figures, and more. In the past that would have taken me multiple days, even weeks.
Last edited:
The term AI is really too broad. AlphaFold uses a technique colloquially called deep learning. It's based on similar concepts like LLMs, but has little to do with general chatbots like ChatGPT.
Apple has never been about elaborate answers to search questions. Apple is a device company with tightly integrated hardware and software that maximizes the user experience. Once AI has matured, become more reliable, and become commoditized, Apple can be expected to partner with an AI provider like they did Google search to provide such results.
I think there are plenty of use cases where "ai" can be very helpful. Here is a short list off the top of my head:
- Translation
- OCR
- Speech recognition/synthesis
- Image recognition to enhance search
- "AI" image manipulation for your personal photos library seems to also be quite capable already
Having said that, I'm much more sceptical about claims from top ai companies about the reasoning abilities of LLMs. I'm not convinced that larger LLMs will really give us general intelligence, however you might define it.
This is an important point. You still have to verify the solutions the coding agents spit out. And I don't think that's something that will change soon, even with more resources. A LLM will never "know" if the solution it spits out is correct. It might be, if it has "seen" the problem in its training data. But what about original and new problems?
It's funny you bring up becoming a plumber: I was having this conversation last night with someone I work with about AI eliminating jobs and he literally said he is encouraging his son to become a plumber and not go to college.
There is a lot of hyperbole about AI, but the march towards eliminating many careers is going to happen and sooner rather than later.
I am not sure I agree with that. If someone threw obscene money at them like Meta is, I'd bet the vast majority of people would jump ship.
What I want to know is when is AI going to be used to create medicines because it should make the medicine so much cheaper having one AI bot create it rather than a team of 50-100 or more human researchers and scientists. AI is already doing away with human jobs, why not medical researchers and scientists where the pharmaceutical companies say it costs them billions to research and build a new drug/medicine which they use to try and justify the very high costs of drugs/medicines. If one AI bot will be able to create a new drug/medicine that took hundreds or even thousands of people to create, surely it would mean cheaper drugs and medicines because not so many humans are needed to create them anymore. I would like to see how a pharmaceutical company will be able to justify selling a drug/medicine for tens of thousands of $$$ when it takes an AI bot a couple of days to create the drug/medicine.
Looking on that principle, companies using AI to replace humans could end up going out of business very fast because they would no longer be able to justify the high cost of their product. The public would demand reduction in prices because only one or two AI bots are being used.
How long will it take before Apple does away with human designers and get's AI bots to design all their products. Apple would no longer be a trillion $$$ company because they would no longer be able to justify the high costings of their products. Bye bye profits, you can thank AI for it.
Looking on that principle, companies using AI to replace humans could end up going out of business very fast because they would no longer be able to justify the high cost of their product. The public would demand reduction in prices because only one or two AI bots are being used.
How long will it take before Apple does away with human designers and get's AI bots to design all their products. Apple would no longer be a trillion $$$ company because they would no longer be able to justify the high costings of their products. Bye bye profits, you can thank AI for it.
The one thing I know AI will not be used for and this for things that make companies and their owners extremely wealthy. The post about DeepBlue winning the noble peace for work on proteins got me thinking. Will AI be used to find a cure for cancer, one of the most profitable areas of medicine there is. Look at all the other medical problems and disabilities that have no cure but yet make those trying to find a cure extremely rich.
Will AI be used to make buildings stronger and more energy efficient. Will AI be used to solve energy problems, like how to make it more cheaply. AI could solve much of the worlds problems but by doing so it would make those who profit from world problems go from very rich to being poor. I seriously doubt AI is going to be used towards the things we feel it should be used towards because far to many rich men and women would be very scared they would be about to lose their wealth.
Will AI be used to make buildings stronger and more energy efficient. Will AI be used to solve energy problems, like how to make it more cheaply. AI could solve much of the worlds problems but by doing so it would make those who profit from world problems go from very rich to being poor. I seriously doubt AI is going to be used towards the things we feel it should be used towards because far to many rich men and women would be very scared they would be about to lose their wealth.