Hacker News Reader

Comment by comboy | Hacker News Reader

Bought by Apple for $360M in 2013. I've been wondering when did they start thinking about this feature (or rather this kind of implementation for it). It seems that even when you have a lot of money and experience, there's still a pretty long way from an idea to its successful execution.

View on HN

That matches pretty closely with the three year design lead period for the Neural Engine that Apple revealed.[1]

[1] http://mashable.com/2017/09/14/inside-apple-a11-bionic-and-s...

Money has little to do with it. You can only add so many engineers to a project. Therefore, some things just take time.

Aka "9 women won't give birth to a child in 1 month". An example often used to describe how adding more developers doesn't necessarily solve an issue quicker.

"What one programmer can do in one month, two programmers can do in two months." - Fred Brooks

Credit for this goes to Fred Brooks, and appears in his famous book "The Mythical Man-Month"

Yes it does. I could not find my copy to find the quote I wanted to use so I used my own (not quite as effective) words.

Didn't mean to imply you'd done something wrong, just added to your comment.

I wonder what they are eyeing up to buy at the moment and what will turn up in iPhones in 5 years time.

Some sort of AI accelerator chip. Or clear phone / heads up display.

> Some sort of AI accelerator chip

like the Neural Engine in the A11? :]

https://www.theverge.com/2017/9/13/16300464/apple-iphone-x-a...

Am I wrong or that article gets many things wrong. These chips are not designed to do any learning, but run pre-trained models (much like FPGA's few years back or Google's TPU).

The actual heavy lifting (model training) still happens and the """cloud""".

You are wrong. For one thing, the learning of your face, and learning over time as your face changes (glasses, hairstyle, beard) happens fully locally.

According to Wikipedia the acquisition happened on Nov 24, 2013, basically 2014

The iPhone X will be released in Nov 2017, basically 2018. So that doesn't change the 4 years

Reply on news.ycombinator.com