Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apple introduces M4 chip

apple.com

Post reply on HN
Print view View original

741–750 of 1001 posts

Re: Apple introduces M4 chip

#741

Post by BitBanging »

bschne wrote: post #515
paulpan wrote: post #183

Earlier quoted context omitted.

I don't think it's strictly for price gouging/segmentation purposes. On the Macbooks (running MacOS), RAM has been used as data cache to speed up data read/write performance until the actual SSD storage operation completes. It makes sense for Apple to account for with higher RAM spec for the 1TB/2TB configurations.

Shouldn't the required cache size be dependent on throughput more so than disk size? It does not necessarily seem like you'd need a bigger write cache if the disk is bigger, people who have a 2TB drive don't read/write 2x as much in a given time as those with a 1TB drive. Or am I missing something?

IIRC SSD manufacturers are likely to store a mapping table of LBAs (logical block addresses) to PBAs (physical block addresses) in the DRAM or Host Memory Buffer.

Some calculation like:

total storage size / page size per LBA (512B or 4KiB usually) * mapping data structure size

Re: Apple introduces M4 chip

#742

Post by marinhero »

_akhe wrote: post #376
marinhero wrote: post #318

I get frustrated seeing this go into the iPad and knowing that we can't get a shell, and run our own binaries there. Not even as a VM like [UserLAnd]( https://userland.tech ). I could effectively travel with one device less in my backpack but instead I have to carry two M chips, two displays, batteries, and so on... It's great to see this tech moving forward but it's frustrating to not see it translate into a more si…

Think the play is "consumer AI". Would you really write code on an iPad? And if you do, do you use an external keyboard?

Yes. If I’m plugging it to a thunderbolt dock I’d expect it to work like a MacBook Air

Re: Apple introduces M4 chip

#744

Post by Havoc »

strangescript wrote: post #569
rsp1984 wrote: post #457

Together with next-generation ML accelerators in the CPU, the high-performance GPU, and higher-bandwidth unified memory, the Neural Engine makes M4 an outrageously powerful chip for AI. In case it is not abundantly clear by now: Apple's AI strategy is to put inference (and longer term even learning) on edge devices. This is completely coherent with their privacy-first strategy (which would be at odds with sending dat…

The fundamental problem with this strategy is model size. I want all my apps to be privacy first with local models, but there is no way they can share models in any kind of coherent way. Especially when good apps are going to fine tune their models. Every app is going to be 3GB+

You could always mix and match. Do lighter task on device and outsource to cloud if needed

Re: Apple introduces M4 chip

#745

Post by Havoc »

zachbee wrote: post #722
rsp1984 wrote: post #457

Together with next-generation ML accelerators in the CPU, the high-performance GPU, and higher-bandwidth unified memory, the Neural Engine makes M4 an outrageously powerful chip for AI. In case it is not abundantly clear by now: Apple's AI strategy is to put inference (and longer term even learning) on edge devices. This is completely coherent with their privacy-first strategy (which would be at odds with sending dat…

If they're doing inference on edge devices, one challenge I see is protecting model weights. If you want to deploy a proprietary model on an edge AI chip, the weights can get stolen via side-channel attacks [1]. Obviously this isn't a concern for open models, but I doubt Apple would go the open models route. [1] https://spectrum.ieee.org/how-prevent-ai-power-usage-secrets

Nobody is taking particular care protecting weights for edge class models

Re: Apple introduces M4 chip

#746

Post by jonplackett »

whimsicalism wrote: post #727
jonplackett wrote: post #699

Earlier quoted context omitted.

I think it will be a winning strategy. Lag is a real killer for LLMs. I think they'll have another LLM on a server (maybe a deal for openai/gemini) that the one on the device can use like ChatGPT uses plugins. But on device Apple have a gigantic advantage. Rabbit and Humane are good ideas humbled by shitty hardware that runs out of battery, gets too hot, has to connect to the internet to do literally anything. Apple…

Having used groq and other fast LLM services a fair bit, lag seems negligible. You're literally just passing text at close to the speed of light.

* when you have a good internet connection

** when you live in the USA

Re: Apple introduces M4 chip

#747

Post by Havoc »

_akhe wrote: post #730

Anyone know more about the "NPU" (neural processing unit)? Basically a GPU specialized for AI?

Yeah. Basically specialises in fast matrix operations so focused on ai and ML. Vaguely like a gpu minus the graphic output and the pipeline needed for that.

Re: Apple introduces M4 chip

#748

Post by Havoc »

dev1ycan wrote: post #736

It's hilarious how they still push the "better for the Environment" garbage when they industrially destroy old iphones.

They have been working on tech to pull them apart via robots to isolate component parts. Unsure if it’s in use at scale though

Re: Apple introduces M4 chip

#749

Post by gnabgib »

ipqk wrote: post #305
slashdev wrote: post #72

I've got a Mac Pro paperweight because the motherboard went. It's going to the landfill. I can't even sell it for parts because I can't erase the SSD. If they didn't solder everything to the board you could actually repair it. When I replace my current Dell laptop, it will be with a repairable framework laptop.

hopefully at least electronics recycling.

Depending on where, a lot of electronics "recyclers" are actually resellers. Some of them are even cheeky enough to deny electronics they know they can't resell (If they're manned.. many are cage-drops in the back of eg Staples)

Re: Apple introduces M4 chip

#750

Post by zmmmmm »

I'm most interested in what this means for the next Vision device.

Half the power budget could well translate to a very significant improvement in heat on the device, battery size and other benefits. Could they even dispense with the puck and get the battery back onto the headset for a consumer version that runs at slightly lower resolution and doesn't have the EyeSight feature?

If they could do that for $2000 I think we'd have a totally different ball game for that device.

Post reply on HN