4 ms·
Apple demonstrates its commitment to AI with new open source code release
- dartharva 3y agoFor a second I was excited thinking they were going to introduce their own LLMs like Meta, but of course.. Apple's gonna Apple.
- adr1an 3y agoIt's an array manipulation library for their silicon https://github.com/ml-explore/mlx https://github.com/ml-explore/mlx
- bastawhiz 3y agoIs this really a commitment to AI? MLX only enables apple silicon to do things. It feels very little like "we're committed to pushing AI forward" and much more like "we're afraid if we don't do this, frameworks will have middling support for our expensive hardware and developers won't embrace what we're doing and consumer software will have AI pushed to the cloud". Which is to say, it does not feel altruistic commitment, it feels like strategy.
- UncleEntity 3y ago> Which is to say, it does not feel altruistic commitment, it feels like strategy. ...to sell more iThings.
- zamalek 3y agoExtend...
- dangus 3y agoOn the other hand, you might as well say that about literally anything they make because they happen to be a company that makes their own OS and hardware. We might as well say that Meta isn’t committed to VR/AR because the fruit of their labor is only sold on their own platform, but I think that description wouldn’t give credit where it’s due. Not that I disagree with the idea that Apple is at risk of being left behind. They are bad at cloud like Nintendo is bad at online play.
- olliej 3y agoYes, a library made to use hardware only works with the hardware it's made to work with. This comment is like saying Nvidia isn't committed to AI because their CUDA libraries only work on Nvidia hardware.
- bastawhiz 3y agoNvidia is the single biggest hardware supplier for anything related to AI. Apple's name essentially doesn't come up in conversations about AI. If Nvidia wasn't a meaningful player in the AI space and they put out a library for their hardware, I'd also say it didn't feel like a "commitment".
- olliej 3y agoAlmost every apple product produced has included "AI" coprocessors for years at this point, and I would guess they've sold many more of those to many more people than Nvidia has. The limiting factor is the ability of 3rd parties to use that hardware, which is what it seems apple has been trying to do for the last few years.
- bilekas 3y ago> That's the power of semantics — because Apple calls it Machine Learning instead of AI, it is believed to not be doing AI enough. That's not semantics, they are different things.
- zamalek 3y agoAI is widely used in place of ML, even in academic situations. We haven't invented AI yet, so all usages of it are incorrect (sans talk about it happening in the future).
- olliej 3y ago"AI" in a fictional setting is different from ML. Modern "AI" in the setting of reality is ML with hype (and funding). That's the difference.
- CamperBob2 3y agoAI = artificial intelligence. It's "ML" when you understand how it works. It's "deep learning" when it's, uh, deep, I guess.
- gumballindie 3y agoYou have not been paying attention. AI is the new IoT and Crypto. Everything is being renamed to AI because it’s what people crave. It’s got electrolytes.
- butlike 3y agoI like Apple's name better. Why is it artificial if it's intelligent?
- mdaniel 3y agothe repo link has a slight edge on comment activity: https://news.ycombinator.com/item?id=38539153 https://news.ycombinator.com/item?id=38539153