Apple's OSes run a lot of local ML models for many tasks that aren't branded as Apple Intelligence, and they have done so for many years now.
dagmx•Mar 2, 2026
If you strip away the branding, Apple has and continues to ship a ton of algorithms that likely use the ANE and end users can use CoreML to do the same.
Just some things that people will likely take for granted that IIRC Apple have said use the ANE or at least would likely benefit from it: object recognition, subject extraction from images and video, content analysis, ARKit, spam detection, audio transcription.
sroussey•Mar 2, 2026
Don’t forget FaceID and many of the image manipulation.
And while everyone else went to more powerful giant LLMs, Apple moved most of Siri from the cloud to your device. Though they do use both (which you can see when Siri corrects itself during transcription—you get the local Siri version corrected later by the cloud version).
love2read•Mar 2, 2026
This article was clearly written by a human (and AI) but still has a few "LLMisms" such as:
- The key insight
- [CoreML] doesn't XXX. It YYY.
With that being said, this is a highly informative article that I enjoyed thoroughly! :)
We've got about a year before so many people are interacting with LLMs on a daily basis that its style starts to reverse infect human speech and writing
Angostura•Mar 2, 2026
My honest take? You're probably right
sholladay•Mar 2, 2026
You are absolutely right.
Here is why you are correct:
- I see what you did there.
- You are always right.
pixl97•Mar 2, 2026
This said, there were people that talked like this before LLMs, it didn't develop this whole cloth.
DrScientist•Mar 2, 2026
Exactly. LLM's are mimics.
People seem to be going around pointing out that people talk like parrots, when in reality it's parrots talk like people.
pixl97•Mar 2, 2026
I mean, it's both.
Did you develop your own whole language at any point to describe the entire world? No, you, me, and society mimic what is around us.
Humans have the advantage, at least at this point, of being a continuous learning device so we adapt and change with the language use around us.
pcrh•Mar 2, 2026
The article above doesn't read well, at all.
It's not my subject, but it reads as a list of things. There's little exposition.
dylan604•Mar 2, 2026
Gawd Damn LISTICLES!!!! And all of those articles that list in bullet points at the top of the article the summary of the article. And all of those people saying they don't want to read exposition, just give me the bullet points.
baxtr•Mar 2, 2026
Great insight – Would you like to try and identify some specific "AI-isms" that you've noticed creeping into your own writing or your colleagues' emails lately?
gogopromptless•Mar 2, 2026
It's already happened to me. I've started to have dreams where instead of some sort of interpersonal struggle the entire dream is just a chatbot UI viewport and I'm arguing with an LLM streaming the responses in. Which is super trippy when I become aware its a dream. In the old days I'd dream about playing chess against myself and lose which was quite bizzare feeling because my brain was running both players. But thats totally normal compared to having my brain pretend to be an LLM inside a dream.
rafram•Mar 2, 2026
Also the Prior Art section, which has telltale repetition of useless verbs like "documenting," "providing insight into," and "confirming" on each line. This was definitely AI-written, at least in part.
tzs•Mar 2, 2026
Below are the items from that section. How should they be written to not look like an AI?
> hollance/neural-engine — Matthijs Hollemans’ comprehensive community documentation of ANE behavior, performance characteristics, and supported operations. The single best existing resource on ANE.
> mdaiter/ane — Early reverse engineering with working Python and Objective-C samples, documenting the ANECompiler framework and IOKit dispatch.
> eiln/ane — A reverse-engineered Linux driver for ANE (Asahi Linux project), providing insight into the kernel-level interface.
> apple/ml-ane-transformers — Apple’s own reference implementation of transformers optimized for ANE, confirming design patterns like channel-first layout and 1×1 conv preference.
mattlangston•Mar 2, 2026
The future is bright for software engineers.
The big takeaway isn't reverse engineering the ANE per se, but what Manjeet could do with his software engineering skills when accelerated by AI.
This is a good example of the present state of software engineering. Not future state - present state.
6.6 FLOPS/W, plus the ability to completely turn off when not in use, so 0W at idle.
AceJohnny2•Mar 2, 2026
But not 38 TOPS that Apple claims, with the weak explanation of
> Apple’s “38 TOPS INT8” is computed as 19 TFLOPS FP16 × 2, following the industry convention of counting INT8 operations as 2× the FP16 rate. But the hardware doesn’t actually execute INT8 operations twice as fast.
Why would Apple follow that convention when the hardware explicitly doesn't seems like a more straight-faced lie that I expect from Apple
kamranjon•Mar 2, 2026
I have always wondered if the neural engine could be used for training - pretty excited for part 3 of this to see if the juice is actually worth the squeeze
juancn•Mar 2, 2026
In principle most if not all inference hardware should be usable for training.
Efficiency is the question.
eleventyseven•Mar 2, 2026
> Throughout this series, “we” refers to maderix (human) and Claude Opus 4.6 (by Anthropic) working as a pair. The reverse engineering, benchmarking, and training code were developed collaboratively
Sure, "collaboratively." Why would I ever trust a vibe coded analysis? How do I, a non expert in this niche, know that Opus isn't pulling a fast one on both of us? LLMs write convincing bullshit that even fools experts. Have you manually verified each fact in this piece? I doubt it. Thanks for the disclaimer, it saved me from having to read it.
withinboredom•Mar 2, 2026
Claude likes to hide bad benchmarks from you, so it will show you where you are clearly winning. You even see some weird benchmarks in the article.
Anonbrit•Mar 2, 2026
Humans also write endless amounts of convincing bullshit, and have done since time immemorial. False papers and faked results have been a growing scourge in academia before LLMs were a thing, and that's just counting the intentional fraud - the reproducibility crisis in science, especially medical and psychological science, affects even the best designed and well intentioned of studies.
Humans also make mistakes and assumptions while reverse engineering, so it will always need more engineers to go through the results, test things
LatencyKills•Mar 2, 2026
I worked on the Xcode team for years and know the lengths Apple goes to make this stuff difficult to figure out.
I just wanted to say that you’ve done an excellent job and am looking forward to the 3rd installment.
RetpolineDrama•Mar 2, 2026
>I worked on the Xcode team for years
Why did you guys remove the ability to detach the console and move it to another window?
daoistmonk•Mar 2, 2026
Tangential: Is anyone doing something similar to accelerate the support matrix of Linux on anything higher than M2?
GeekyBear•Mar 2, 2026
The recent news is that Apple is supposedly replacing the Core ML framework with an updated version that will make it easier to integrate third party LLMs into your apps.
> the company is also planning a few other software-based AI upgrades, including a new framework called Core AI. The idea is to replace the long-existing Core ML with something a bit more modern.
It's insane that the source code of ANE is not available even to the MLX team, possibly one of the reasons Awni (MLX project head) left Apple.
FL33TW00D•Mar 2, 2026
Unreadable Claude slop
mayhemducks•Mar 2, 2026
I never realized just how much hardware engineering Apple dedicated to enabling people to type faster with their thumbs!
giancarlostoro•Mar 2, 2026
Reverse Engineering with AI is only going to get better. I have seen some crazy things friends of mine have done with Claude alone. Let's just says SaaS isn't the only industry that could one day suffer.
msie•Mar 2, 2026
I remember the good old days when Apple was desperate for developers and produced great documentation and there were a lot of great 3rd-party books too. You can't just give out awards in hopes that someone will make that great app.
pstuart•Mar 2, 2026
Yeah, the Inside Macintosh guides were epic.
ericol•Mar 2, 2026
> human intuition driving the exploration
This, a thousand times this.
For me, what AI brings is augmented humans. Just as we don't calculate on paper anymore, what is the reason of doing things by hand when a machine in X times better.
Want to code by hand, as artisans of old? Suit yourself.
I, for one, love the smell of burning chrome.
pklausler•Mar 2, 2026
If "AI" were doing anything more than repeating content from the web without attribution, I might agree with you.
grey-area•Mar 2, 2026
If only they could fix the iOS autocomplete, which is getting worse with every iteration.
zozbot234•Mar 2, 2026
Much of this information we already knew the very basics of from documentation of the M1/M2 ANE as accessed via bare-metal from Asahi Linux, but it's nice to see confirmation and it being explored in further depth. Note that according to OP Parts 1/2 for very large matmuls CoreML adds little to no overhead compared to the lower-level interface, so there seems to be plenty of scope for supporting ANE for prefill in local AI frameworks. Decode is generally memory-bandwidth limited unless context is very large, and the ANE requires special handling (converting from matmul to 1x1 convolution as described here is wasteful of memory bandwidth, as is potentially dequantizing to INT8/FP16 in memory) so it's less of a clear win.
notepad0x90•Mar 2, 2026
I've been guilty of this myself, but every other comment here is like "What about <insert something unrelated to the topic but related to apple>".
blobbers•Mar 2, 2026
Can someone help me understand when these neural engines kick in in open source software?
I typically use python ML libraries like lightgbm, sklearn, xgboost etc.
I also use numpy for large correlation matrices, covariance etc.
Are these operations accelerated? Is there a simple way to benchmark?
I see a lot of benchmarks on what look like C functions, but today in my jobs I rely on higher level libraries. I don't know if they perform any better on apple HW, and unless they have a flag like use_ane I'm inclined to think they do better.
Of course chatgpt suggested I benchmark an Intel Mac vs. newer apple silicon. Thanks chatgpt, there's a reason people still hate AI.
zozbot234•Mar 2, 2026
> when these neural engines kick in in open source software?
It mostly doesn't because NPUs are bespoke and vendor-specific (which incents neglect by software devs working on open source numerics and ML/AI infrastructure), and the Apple ANE is no exception. Part of this effort is most likely about fixing that for the specific case of the Apple ANE.
blobbers•Mar 2, 2026
Part of which effort? The Reverse engineering is so it can be used blog article?
I just think: great it seems like I'm paying for a hardware accelerator that makes Siri go faster. And I use siri on my laptop exactly 0 times in the last infinite years.
19 Comments
Just some things that people will likely take for granted that IIRC Apple have said use the ANE or at least would likely benefit from it: object recognition, subject extraction from images and video, content analysis, ARKit, spam detection, audio transcription.
And while everyone else went to more powerful giant LLMs, Apple moved most of Siri from the cloud to your device. Though they do use both (which you can see when Siri corrects itself during transcription—you get the local Siri version corrected later by the cloud version).
- The key insight - [CoreML] doesn't XXX. It YYY.
With that being said, this is a highly informative article that I enjoyed thoroughly! :)
The article links to their own Github repo: https://github.com/maderix/ANE
Here is why you are correct:
- I see what you did there.
- You are always right.
People seem to be going around pointing out that people talk like parrots, when in reality it's parrots talk like people.
Did you develop your own whole language at any point to describe the entire world? No, you, me, and society mimic what is around us.
Humans have the advantage, at least at this point, of being a continuous learning device so we adapt and change with the language use around us.
It's not my subject, but it reads as a list of things. There's little exposition.
> hollance/neural-engine — Matthijs Hollemans’ comprehensive community documentation of ANE behavior, performance characteristics, and supported operations. The single best existing resource on ANE.
> mdaiter/ane — Early reverse engineering with working Python and Objective-C samples, documenting the ANECompiler framework and IOKit dispatch.
> eiln/ane — A reverse-engineered Linux driver for ANE (Asahi Linux project), providing insight into the kernel-level interface.
> apple/ml-ane-transformers — Apple’s own reference implementation of transformers optimized for ANE, confirming design patterns like channel-first layout and 1×1 conv preference.
The big takeaway isn't reverse engineering the ANE per se, but what Manjeet could do with his software engineering skills when accelerated by AI.
This is a good example of the present state of software engineering. Not future state - present state.
6.6 FLOPS/W, plus the ability to completely turn off when not in use, so 0W at idle.
> Apple’s “38 TOPS INT8” is computed as 19 TFLOPS FP16 × 2, following the industry convention of counting INT8 operations as 2× the FP16 rate. But the hardware doesn’t actually execute INT8 operations twice as fast.
Why would Apple follow that convention when the hardware explicitly doesn't seems like a more straight-faced lie that I expect from Apple
Efficiency is the question.
Sure, "collaboratively." Why would I ever trust a vibe coded analysis? How do I, a non expert in this niche, know that Opus isn't pulling a fast one on both of us? LLMs write convincing bullshit that even fools experts. Have you manually verified each fact in this piece? I doubt it. Thanks for the disclaimer, it saved me from having to read it.
Humans also make mistakes and assumptions while reverse engineering, so it will always need more engineers to go through the results, test things
I just wanted to say that you’ve done an excellent job and am looking forward to the 3rd installment.
Why did you guys remove the ability to detach the console and move it to another window?
> the company is also planning a few other software-based AI upgrades, including a new framework called Core AI. The idea is to replace the long-existing Core ML with something a bit more modern.
https://www.bloomberg.com/news/newsletters/2026-03-01/apple-...
This, a thousand times this.
For me, what AI brings is augmented humans. Just as we don't calculate on paper anymore, what is the reason of doing things by hand when a machine in X times better.
Want to code by hand, as artisans of old? Suit yourself.
I, for one, love the smell of burning chrome.
I typically use python ML libraries like lightgbm, sklearn, xgboost etc.
I also use numpy for large correlation matrices, covariance etc.
Are these operations accelerated? Is there a simple way to benchmark?
I see a lot of benchmarks on what look like C functions, but today in my jobs I rely on higher level libraries. I don't know if they perform any better on apple HW, and unless they have a flag like use_ane I'm inclined to think they do better.
Of course chatgpt suggested I benchmark an Intel Mac vs. newer apple silicon. Thanks chatgpt, there's a reason people still hate AI.
It mostly doesn't because NPUs are bespoke and vendor-specific (which incents neglect by software devs working on open source numerics and ML/AI infrastructure), and the Apple ANE is no exception. Part of this effort is most likely about fixing that for the specific case of the Apple ANE.
I just think: great it seems like I'm paying for a hardware accelerator that makes Siri go faster. And I use siri on my laptop exactly 0 times in the last infinite years.