Hacker News Reader

Comment by integricho | Hacker News Reader

Intel transformed from a technology company to a management company. That somewhat explains their highly unreasonable behavior.

View on HN

The part of the lifecycle where the clueless have taken over:

https://www.ribbonfarm.com/2009/10/07/the-gervais-principle-...

That is an excellent bit of writing. Thanks for the link!

Also part basis for Erik Dietrich's Developer Hegemony.

A masterpiece of writing! A tour-de-force!

I highly recommend all of ribbonfarm and especially:

- ancient rivers of money

- The Return of the Barbarian

- Can Hydras eat unknown unknowns for lunch

There is also a more recent one where he writes about young people and their attitudes about money and property ownership and "life scripts" that was here on HN and metafilter and other places but I don't know the name of it ... was quite good.

Sounds like "The Premium Mediocre Life of Maya Millennial" https://www.ribbonfarm.com/2017/08/17/the-premium-mediocre-l...

No, I don't know how I came to be replying about Ribbonfarm a day late to a comment deep in a thread I haven't read most of.

Management can seriously make or break a company in various ways. It sucks... :/

Could not agree more, unfortunately I’ve seen the latter more often than the former. When you find good management it is truly inspiring however.

There's some bias to this, though. Good management is much more "everything and everybody just seemingly by if self seems to be more or less in the right place" and less "Steve Jobs screaming at everybody about rounded corners and then everybody realising five years later that this is actually a much bigger deal than everbody except Jobs saw at the time" -- in other words, pretty invisible.

I'm curious, how were the round corners important? I like them, and I buy his argument that nothing in the real world has sharp corners, but Android has had sharp corners forever. (I'm asking because I'd like to know the answer, not because I'm asserting the opposite.)

I really don't know. It just seems to be once of the anecdotes people use to illustrate the genius of Jobs.

Best i can tell, the latter crops up because of boardroom meddling pushing the focus towards quarterly earnings rather than long term sustainability of the company.

Yea... and what about every other chip maker who also have the same exploits? Can someone explain whether AMD and Qualcomm licensed this tech or if they also independently made the same mistakes? Either way, we really haven't seen much activity from them either.

https://en.m.wikipedia.org/wiki/Whataboutism

There are fewer people trying to attack AMD's chips, they comprise such a small chunk of the installed CPU base out there that most attacker's won't be going after a system containing a CPU from AMD. This may change in the next year or two though, at which point the vulns that are likely in their Platform Security Processor may bite them in the ass.

Qualcomm on the other hand has taken the "fuck security" approach and failed to clean up their code and get it mainlined for nearly all of their common SoCs. Hence why Android phones are often trapped on old versions of Android when launched or shortly thereafter, long term security through the device lifecycle is a bad joke when Qualcomm and most other ARM vendors are involved.

> Hence why Android phones are often trapped on old versions of Android when launched or shortly thereafter

The reason why most android phones don't get upgrades (or only one) to newer android versions has absolutely nothing to do with qualcomm lack of care regarding their cpu's security.

Actually it's one of the major reasons. When Qualcomm stops providing support for a SoC, the OEMs can't update the device even if they wanted to. It's not the only reason, but it is a very major blocker.

It has been quoted by Google developers on why Nexus 5 fell off the support track so early, around the same time Nexus 4 did: it took a herculean effort to get the binary drivers to even remotely work right on Android 6 due to being unable to drag in newer kernels.

This is partially why Treble is being worked on, to just completely disconnect the actual Linux-based HAL from "the OS", thus trying to negate a lot of the oblate stupidity from vendors like Qualcomm and have a shim in-between your actual phone, and Android itself, in hopes to fix stuff.

We seriously need more phone SoC vendors.

How am I running Oreo on my Nexus 5 with lineage os?

LineageOS (ex-CyanogenMod)'s flavour of Android 7 works just fine [1] on Nexus 5.

On the other hand - yes, hammerhead is stuck on Linux kernel 3.4, which is no longer maintained, and current best effort to forward-port to mainline kernel [2] is far from complete.

Among other things this means that there is no workaround for Meltdown on this hardware, which it is likely vulnerable to. [3]

[1] https://download.lineageos.org/hammerhead

[2] https://wiki.postmarketos.org/wiki/Google_Nexus_5_(lg-hammer...

[3] "11 stage integer pipeline with 3-way decode and 4-way out-of-order speculative issue superscalar execution" - from https://en.wikipedia.org/wiki/Krait_(CPU)

That is just an excuse from Google that many blindly accept.

First, of all there are legal obligations known as contracts. If Google was serious about Nexus 5 support from OEMs, it would have had contract in place that would assure the necessary support regardless of what OEMs would decide.

Secondly, Treble is yet another solution that will die, because Google, again, decided to leave to OEMs the task of delivering Treble updates, without any kind of legal obligation.

Additionally Android devices that get upgraded to Oreo aren't obliged to be fully Treble compliant, only the ones released with Oreo.

So, many new devices planned to be released in 2018, are based on Android 7.0.

> First, of all there are legal obligations known as contracts.

That kind if snark is unnecessary and degrades the conversation.

> If Google was serious about Nexus 5 support from OEMs, it would have had contract in place that would assure the necessary support regardless of what OEMs would decide.

I'm not sure who you are referring to with "OEM" here, but Google took on the responsibility of maintaining the OS images for the Nexus devices, not the companies they contracted to build them.

> That kind if snark is unnecessary and degrades the conversation.

May be, but apparently discussions about Android's and OEMs usually leads to hand waving regarding the legal options available to Google, if they really cared about enforcing updates.

> I'm not sure who you are referring to with "OEM" here, but Google took on the responsibility of maintaining the OS images for the Nexus devices, not the companies they contracted to build them.

Like the customers they have abandoned after TI cut down support on their chipsets?

Had Google taken the legal requirements to enforce TI to maintain support or face monetary sanctions for the three years it was supposed to take place, this would not have happened as it did.

It always circles down to how much Google cares.

Nice whattaboutism there.

You’re conflating Spectre with Meltdown. Linus’ response is specific to Meltdown, which applies only to Intel chips. Spectre applies to numerous other architectures.

Uh, Linus is very specifically talking primarily about Spectre. IBRS, IPBP, and retpoline are all Spectre specific mitigations. The Meltdown mention is a sidenote.

The issue is how Intel is handling these problems. If other vendors handle their issues similarly, they may deserve a similar response, though Intel, as the market leader in x86-64, has a special responsibility (and corresponding culpability), as other vendors may have to conform to remain competitive, as the marketplace is capricious in rewarding virtue.

Reply on news.ycombinator.com