Hacker News Reader

Comment by kevindqc | Hacker News Reader

Huge performance difference - how did they not notice before?

The article made it seem like this only affected a small number of units? Their quality control probably doesn’t look at every single unit. Although it’s interesting that they didn’t find it at all, since reading HN and reddit gave me the impression it was very widespread.

Generally you can’t get a sense for how widespread an issue is from social media, because it’s a perfect medium for confirmation bias. Confirming links and stories are shared/upvoted/forwarded while non-confirming ones are ignored or dismissed (or outright attacked for some reason).

This is basically the story of the modern world as viewed through the distorting lens that is media (and especially social media). I wish people would take the two sentences you wrote and then use them as a corrective lens that undoes this distorted perspective.

Yeah it’s really hard to correct the perspective to the truth though if all you have is social media. Honestly I’ve been putting less and less trust in almost all my sources of knowledge recently.

It's so, so strange that 10-20 years ago my parents told me not to trust what I read on the internet...and now that they're all using it they have seem forgotten their own advice.

I would not trust any coverage of supposed issues with Apple hardware. There is a huge incentive for negative coverage. Remember Antenna-gate with the iPhone 4? They continued selling the iPhone 4, completely unmodified, for years after that whole debacle. That's how widespread the supposed issue was.

Apple settled a fairly large lawsuit related to Antennagate and was forced to send out a bumper case to affected users so they could hold the phone the "Jobs" way and get reception rather than the way you would normally hold...literally any other phone before, during, or since Antennagate, including newer iPhone models.

TBH, the "Jobs" way was the normal way that most people held a phone. It was the people who smothered the whole handset in some kind of crazy death grip that had attenuation problems.

I got my bumper in the mail, but never needed or used it.

Most of the people I know who had iPhones had issues with the iphone 4 antenna, and only one or two of them are death-grippers. The problem was Apple's poor design, which Apple itself acknowledged by settling the case, offering free bumpers, and redesigning the antenna in future models.

They settled for $15/phone or a free case (they had already offered a free case to customers of their own accord). They kept the iPhone 4 around for 2 years after it was no longer their marquee model. Surely if Antennagate was actually a problem for a significant portion of customers they would have stopped selling it as soon as possible (aka, upon release of the 4S). But they didn't.

I think you severely underestimate the difficulty in retooling supply chains for relatively minor issues. It wasn't worth the cost of doing what you suggest when the issue could easily be fixed with a free bumper or by getting people to hold the phone in a non-ergonomic manner.

> It wasn't worth the cost of doing what you suggest when the issue could easily be fixed with a free bumper or by getting people to hold the phone in a non-ergonomic manner.

Exactly. It wasn't that big of a deal.

I mean, we haven't even gotten into how a bunch of other phones which were on sale at the time exhibited the exact same behavior.

And, as was pointed out on these very pages, immediately after they denied it was a big problem an ad for an antenna engineer appeared on their pages.

A more relevant example would be the Dell XPS 15 where owners and a few reviewers have reported extreme thermal throttling with i7 models.

>As mentioned in our detailed review, the XPS 15 9560 with the Core i7-7700HQ processor is prone to two types of throttling:

    Thermal throttling of the CPU or GPU (generally the CPU) when temperatures get too high

    Voltage Regulator Module (VRM) throttling caused by it getting too hot and being unable to deliver enough power

https://www.ultrabookreview.com/14875-fix-throttling-xps-15/

You have to wonder why a well known laptop that can't run a mainstream i7 without severe throttling is a story that couldn't gain any traction.

It isn't has bad as the Apple one though. But of course, Dell should also fix their shit.

I'd say that a similarly priced high end laptop that cannot adequately cool a four core i7 is a bigger deal, especially given that this is their second year to ship devices with the same issue.

>The service manual for the Dell XPS 15 9570 is now available and it reveals something that we wished would improve in this generation — that the XPS 15 9570 carries over the same 2-pipe heatsink assembly from the previous generation. Not just that, there is also no indication of improved thermal dissipation from the Voltage Regulator Modules (VRMs) implying that we could be in for some serious throttling issues.

The Dell XPS 15 while earning accolades for being a very capable multimedia laptop that can also game, has also earned the reputation for throttling heavily under load. Both the earlier XPS 15s suffered from both thermal and power-limit throttling. The only way to fix this was to repaste and undervolt the CPU and also apply thermal pads on the VRMs to drive away all the excess heat.

Unfortunately, it looks like Dell has chosen to retain the existing heatsink design for the XPS 15 9570. Compounding throttling fears is the availability of a Core i9-8950HK variant.

https://www.notebookcheck.net/The-Dell-XPS-15-9570-might-inh...

You would think that after two years with the same issue, the mainstream tech press would pick up on it if the issue is throttling and not just chasing clickbait stories.

Not many manufacturers at their scale charge their prices so their QA procedure should be unorthodox as well.

Apple always claims their issues affect a small number of units. I have seen zero people that tested the i9 MBP and did not see major throttling. The issue is clearly the result of a fundamental defect in how they were managing power and thermals.

The official explanation seems to be that the key management defect which lead to thermal management problems did not occur in the development machines used during the engineering process. This seems possible as presumably every machine used internally goes through a different provisioning process than those released into customer hands ... Thus the claimed cause of the problem seems to indicate that the flaw did not affect the internal benchmarks they ran. It would be _very_ concerning if the benchmarks they ran for engineering/marketing production were insufficient to catch this issue.

I think the official explanation seems pretty likely. If the flaw had been present in the machines used for benchmarking than either their benchmarks didn’t catch the issue (unlikely) or the benchmarking numbers were fabricated which also seems unlikely as it would invite an infeasible amount of legal risk for an organization of apples size and mindshare ...

> It would be _very_ concerning if the benchmarks they ran for engineering/marketing production were insufficient to catch this issue.

Clarification -- I meant to say -- 'It would be _very_ concerning if the benchmarks they ran for engineering/marketing production were insufficient to catch this issue on a system configuration in which the issue manifests.'

Engineering, especially at this level, is not supposed to work like that. You certainly not do dev proto, then switch to prod both in one step and without even rechecking the tests you did on the proto. So their official explanation might be kind of right, but then it shows major deficiencies in their quality process.

Which given the track record of quality issues they have had, is possible.

> You certainly not do dev proto, then switch to prod both in one step and without even rechecking the tests you did on the proto.

Exactly. I work in software engineering for cars and even for things that are not mission critical or life-threatening, the engineering and quality process is insane because so much money is on the line. I wouldn't be surprised if macbook pros are worth more in profit, or at least revenue, than a car line and should therefore see an equal engineering process.

I can't understand how the firmware was released without the proper key, unless there was a bad merge or they dropped a pilot firmware that was testing without thermal management... maybe a clue there is a desktop in the works and the desktop firmware made it onto the macbook pros? Either way, like you said, their quality process has a major deficiency.

> affect a small number of units

Same story with the keyboard that was a fundamental enough design flaw to include 2 design bandaids in as many years.

Reply on news.ycombinator.com