Hacker News Reader

Comment by ulfw | Hacker News Reader

I have no idea why you're getting downvoted.

If a company decides not to produce wireless routers (which were class leading) or monitors etc anymore, all because they're non "core" to a computer/mobile hardware company then I don't see how building automotive AI is core to a computer/mobile hardware computer company that has zero core competency in either car manufacturing or AI.

View on HN

I suppose they're getting downvoted because "why are you working on X when Y has problems" is kind of a silly argument. The problems with the butterfly keyboard design are almost certainly not a result of all the keyswitch engineers having been foolishly reassigned to work in the automotive group.

It's very much a focus and resource allocation problem. Double the number of engineers working on the keyboard and it doesn't happen

Doubling the number of engineers can also make progress slower, once communication costs dominate.

(Or did you mean the keyboard doesn't happen if number of engineers is doubled? I don't think it can be that bad.)

…and 9 women are known to bear a child in one month

Adding people to a department actually helps if it's understaffed.

But I guess the keyboard design is a managment problem not a engineer problem.

"The mythical man month" should not be interpreted dogmatically.

The existence of the Autonomous Car Unit was also a management problem.

The unstated truth in the "mythical man month" is that you can't rescue a project by throwing people at it if the project is out of control because of poor management.

Stellar management can do almost anything.

Mediocre management bumbles around creating failures.

> Adding people to a department actually helps if it's understaffed

Eventually. At first it makes matters worth. And that's assuming your predicate is correct, that the department is understaffed.

I can almost guarantee that doubling the number of engineers working on the keyboard would make the problem worse, rather than better.

Replacing the current ones would work.

Well there has to be a reason, people are assuming it’s resources, because incompetence which isn’t noticed and fixed by management is hard to understand

If management isn't competent, they won't be able to discern competence from non-competence.

It would seem to me the issue is because of bad product design decisions. Apple decided having user serviceable reliable keys was less important than making the thinnest laptop they possibly could.

It's left for the market to validate whether it was the correct decision or not.

Having returned my 2018 "improved butterfly keyboard" MBA because the keyboard died after 3 weeks, I suspect that the issue isn't being taken seriously. A minor design revision that didn't pick up or resolve a recurring issue is a death march for a product.

Okay, but on the flip side I used the 2017 MBP for a year with no problems whatsoever. It's hard to figure out realistically how much of a problem this is for them.

Take the dead ones to bits and do a post mortem. There's enough of them.

By this logic we should just triple the cancer researchers and then it doesn't happen

Besides what the commenters are saying, physical keyboards have been a solved problem for what, 2 decades now? Even slim ones. There's no new basic research needed and even the engineering should be quite obvious by now...

Maybe I'm missing something.

They stopped making routers and monitors not because they are non “core” products, they stopped making them because there is not much money in it. You buy routers ones every 10 years probably, and margin is limited, but supporting them throughout this period is quite expensive. As for monitors, there is so many competitors right now, that it’s hard to make it worth building your own monitor, Especially if you want to win the market share you need to distinguish yourself from competition, and there is very little you can do to make your monitor distinguishable from your competitors.

Automotive AI meanwhile is a future and they have to think about future with all the money they have now.

Another option, assuming core Apple processes don't need more funding: they could pay out that cash as dividends, and let the market think about the future (getting the money to companies who's core actually is about automotive AI, if investors actually think so).

The other way of thinking about it is that autonomous cars are so obvious and will eventually be so commodity that Apple not sticking their logo on this market would be about as dumb as them not making the bigger phones the market demanded.

  autonomous cars are so obvious and will eventually be so commodity

Apple's core business (smart phones) will be commoditized way before autonomous cars - we're already seeing smart phone improvements becoming more and more incremental, longer device lifetimes, people less willing to spend for the latest and greatest

Would you buy an autonomous car from a company that can't build a reliable keyboard, and is still struggling with its maps?

I would be scared to buy an Apple car. Let there be a design flaw: It will take a class action suit for them to acknowledge it. Until then, anyone who drives to their Apple Car Store will be told to pay out of their own pocket for fixing it, that they obviously mishandled the car, Apple is not to blame, it is not covered under warranty.

As opposed to Google with their world class customer support or Uber with their great reputation for safety and ethics....

Or a GM car where they do the math to see if its worth it to issue a recall for cars that shut themselves off out of nowhere or Ford where they did a cost benefit analysis on a recall of Pinto fuel lines.

In other words, like every other business....

Apple does pay out dividends, as well as pursuing stock buy backs which pushes share prices up, rewarding investors.

If Apple truly cares about the user experience they would have kept the wireless routers in their portfolio. A great opportunity missed to have a product that is more user friendly than the competition.

I'd argue the wholesome user experience will prove be more important for Apple in the long run than having a big profit margin on every product.

Sure, they could have been the ones launching the Unifi line of routers and switches, instead of Ubiquiti, but that would mean caring about and supporting non-mac equipment and computers in the offices.

The “user experience” for most users trying to use a third party router instead of the one bundled with their internet service where support comes from their ISP would be worse.

> Automotive AI meanwhile is a future and they have to think about future with all the money they have now.

Automotive "AI" is not a future, more than so than the automotive "cloud." It does not even takes one to have a computer science background to figure out that claims of artificial "intelligence" there are a joke.

That's going to be an unpopular opinion on HN, but I really agree. Getting self-driving car to where they are currently is impressive, but the technology isn't close to where it needs to be.

The technology is 90% there, perhaps more, but the last percentages that it will required for this to be a mass market product is going to take 20 or 30 years to reach. The current solutions are built using ideas and research grounded in the field of "Artificial Intelligence", but they are not "true AI".

There's also no AI in your Huawei phone. There's software using machine learning, which is a discipline under the umbrella of "artificial intelligence" as a computer science discipline.

non “core” products, they stopped making them because there is not much money in it

The idea of Mac was always “it just works”. You could go to the Apple Store and get everything you needed and know that while it wasn’t cheap, it was amongst the best you could buy. You pay the premium for that experience and you remember how nice it was and come back next time.

Now that experience is broken, you need to shop around and make it work yourself. And if you need to do that then the experience is no better than anyone else’s.

> I don't see how building automotive AI is core to a computer/mobile hardware computer

I don't see why Apple needs to stay within the bounds what what you think their "core" products are.

They don't, but as a public company they do have obligations to their shareholders which at least suggest they shouldn't be neglecting their core business.

I think the shareholders are the main reason why Apple actually stopped making routers and monitors. There isn’t much money in those products, and shareholders wants to see revenues, they don’t care what products you build as long as they make them money

Do you think before the iPhone came out that Apple was experts in basebands and antenna design? Were you bitching about that when iTunes was a PoS in 2005? Were you ranting that the Wi-Fi bugs in 10.4 weren’t getting fixed because they were working on phones and tablets which they had no core compentency in back then? Point being, everyone is calling doom for Apple saying they’ve not invented the next great thing, yet when they work on the fields that will likely be the next big thing, they get ridiculed by folks like you. Seems they are damned either way, no?

I kinda feel cars are core to Apple. Essentially they are a screen company if you think about it. Their job is to own all your screens and make a quality experience across them, both the bonus of upselling services to said screens. A car dash is the last (others?) of the more used screens they haven't covered yet. So probably a market they dont want to miss out on, especially as it ties in with other areas so well like music, maps and in the future apps.

They do have Apple CarPlay though.

Which is actually pretty good. Had one in a rental a month or so back and I was, wow, a vehicle entertainment system that actually works rather than being primarily occupied with sucking the contacts off my phone.

It's really not that confusing. Routers, discrete monitors, or cars are not core to Apple's existing businesses. But of the three, only cars represent a market large enough to be worth pursuing. So that's where Apple invests and attempts to develop competency.

It's hard to see them making any money on cars anytime soon. Autonomous cars are at least 20 years away. Profit margins on cars isn't even close to that of an iPhone. Also it would be entering an entirely new field of business that not only require the R&D to make the car, they also need the infrastructure to handle sales and repairs.

Apple does have the money to pull it off, if they want to. I would just question if they'd profit from the project.

“We’ve learned and struggled for a few years here figuring out how to make a decent phone, PC guys are not going to just figure this out. They’re not going to just walk in.”

Former CEO of Palm.

>then I don't see how building automotive AI is core to a computer/mobile hardware computer company that has zero core competency in either car manufacturing or AI.

Autonomous Vehicle, is basically a Giant Mobile iPhone that you can sit in it. There will be lots of custom silicon doing software calculation, LiDAR making 10s to million of input every second, a gigantic battery, and lots of softwares.

Although I still think AV is at least another 5 - 10 years away in real world usage.

> Autonomous Vehicle, is basically a Giant Mobile iPhone that you can sit in it.

Ugh. What's next? It seems in the future everything will look like an iPhone and is built by one company. Why can't we have more modularization in the economy, i.e. smaller companies that do 1 thing well, which sell to potentially multiple other companies that do 1 thing well too? Perhaps we should set a hard limit on the number of employees working for a single company, e.g. 1k.

An iPhone is made by many different suppliers. Apple only makes a handful of chips. Important chips, but hardly everything in the phone.

Apple doesn’t even “make” the chips, they design them. TSMC and GlobalFoundry have been the fabs historically that have done the actual “making” of them.

The point is that if Apple's obsession with supply chain integration continues, then at some point even the fabs might be assimilated. This can happen e.g. if their CPU design division needs tighter control over fabrication, or if they develop a key technology at the silicon level that they want to keep out of the hands of their competitors.

That's quite the understatement.

I wonder what is taking Tesla so long to get their factory up and running.

You have core now, core in the future, and never core.

When work started on the iphone, phone weren't core, but there was a vision that it could be. Routers were killed because there was no future where routers would ever be core.

Apparently Apple sees (saw) a future where automotive AI might be core.

I think ulfw’s argument about product line focus makes sense, but the parent was implying Apple can’t make laptops, whereas compared to most manufacturers their products are incredible (but still imperfect).

Reply on news.ycombinator.com