And every 2 years you can count on them cutting support for whatever devices you bought from them because they got bored and moved on to something else.

In their defense, OnHub routers launched late 2015 and was discontinued/rebranded Google Wifi in late 2016, but they did end up supporting the older models for approx. 4 years of updates, including phone support. That was a product that made me trust Google just as Apple was getting out of the router business. Google Home, Nest and Dropcam have all had older products remain supported, so far.

Meanwhile I'm rocking a 10 year old DSL router that is still being updated with security patches.

4 years is nothing in the consumer market with the exception of maybe smartphones.

Nobody has the money to buy all electronics new every 4 years and looking at the current Android version distribution, a lot of consumers don't even buy new smartphones every 4 years.

> 4 years is nothing in the consumer market with the exception of maybe smartphones.

Smartphones that expire rapidly is a Google innovation.

6 years in Apple's case is also way too short.

Home appliance lifetime is typically measured in decades.

According to Wikipedia, Apple sold the iPhone 6 until September 12, 2018¹, and it did not get iOS 13, which was released September 19, 2019² — barely one year of OS support.

¹ https://en.wikipedia.org/wiki/IPhone_6

² https://en.wikipedia.org/wiki/IOS_13

It may not support iOS 13, but Apple has been providing updates to iOS 12 (12.4.8 was released less than 3 weeks ago).

>barely one year of OS support.

Measuring support spans this way is... odd. It's like saying "ubuntu only comes with 1 year of support" because you installed 16.04 today. If we're going by this metric a lot of electronics have 0 years of support. Regardless, apple is still has the longest support, at least when it comes to phones/tablets.

You're blurring the line between software and hardware support, which is a very important distinction.

Ubuntu 16.04 may only have one year of support left, but my current desktop hardware supports both 16.04 and 20.04. It's much more reasonable to sunset old software to reduce maintenance burden while providing an upgrade path than to lock people out of any upgrade path for their hardware.

There is also the distinction that Apple sells closed hardware and prevents you from installing a custom OS with longer software support. And that Apple products cost $$ while Ubuntu is free, which leads to different reasonable expectations. But I don't think either of those are needed to make my point.

> You're blurring the line between software and hardware support, which is a very important distinction.

Isn't hardware support just called "warranty"? That remains the same regardless of when you bought it.

>And that Apple products cost $$ while Ubuntu is free, which leads to different reasonable expectations.

My point wasn't whether canonical/apple's support lifecycle policies were reasonable, just that that defining support periods as the last day it was sold, rather the first day it was sold was unusual/misleading. If you want something to be supported a long time after you get it, you don't grab the oldest version, simple as that. Consumer products don't have a support subscription like enterprise products do, so limiting support to old products is really the only economically viable choice that companies have.

FWIW, the 6 was released September 2014. The 6 alongside the 5s (released a year prior in September 2013) are still receiving patches for iOS 12, the most recent being July 15th.

For that reason I think it's more than fine to continue selling it discounted until 2018, it's not like they were marketing it as flagship.

Additionally the 6s from Sept 2015 is receiving iOS 14.

Yeah, it feels like "damned if you do, damned if you don't". Folks would complain "why can't I buy a simple phone like the 6!" if Apple only sold the latest and greatest, but then when you sell something older, folks want to complain if it's not able to run the latest OS.

Not that Apple's being altruistic here, obviously they want to segment the market and get money from wherever they can. But, can't expect them to sell low price point devices that are as future-proofed (future-resisted?) as top end hardware.

I don't know about anyone else, but I'm using my phone a lot less now that I just stay home all the time.

Not really. In the case of eg. fridges, only parts of it (the compressor, usually) is warrantied for that long.

Isn't it the opposite? I thought the compressor was the thing that was not warranted because they break so frequently.

So I actually went on bestbuy and looked at what the waranty policy for a few of the refridegrator brands were, and they were nowhere close to "decades" like the grandparent comment claims.

samsung: 1 year https://www.bestbuy.com/site/samsung-22-6-cu-ft-french-door-...

whirlpool: 1 year https://www.bestbuy.com/site/whirlpool-25-2-cu-ft-french-doo...

lg: 1 Year General, 5 Years Sealed System, 5 Years Compressor, 6-10 Years Linear Compressor. https://www.bestbuy.com/site/lg-23-5-cu-ft-french-door-count...

bosch: 1 year https://www.bestbuy.com/site/bosch-800-series-20-5-cu-ft-4-d...

kitchenaid: 1 year labor, 10 year parts https://www.bestbuy.com/site/kitchenaid-25-8-cu-ft-5-door-fr...

>I thought the compressor was the thing that was not warranted because they break so frequently.

given the sample above, it seems like at least with LG, that they have a lot of confidence in their compressors and "sealed system" (probably the refrigerant system).

Warranty is not the same thing as how long it lasts in practice. I would never buy a fridge from a manufacturer again if it stopped working after a year and a half.

I think the mental hurdle here is that while the warranty might be short, the fridge doesn't just stop working after the warranty period is over. And if it does, well there's a 100+ year old industry of fridge repair. But, given the track record in software of not just ceasing to support the products but actively killing them, the move into hardware might give one pause as the devices may well be bricked. At least that's my interpretation of my own POV re: almost anything Google wrt service. Then there's the data privacy element...

2 at the most... and that is pushing it

Home Furnace is probably the longest lasting Home Appliance and it has an expected life of 15-20 years. Common Kitchen Appliances (and most other appliances) are about 10 years...

Maybe I'm the exception but all my home appliances (washing machine, dishwasher, fridge, freezer) are around 25-30 years old with the exception of my oven and stove. Everything is working as expected and I suspect they will last even longer.

That may be true for most smart home gear, but I place routers in a slightly different class of home electronics. Like gaming consoles, for performance and upgrades to the latest standards (Wi-Fi 6E, for example) you’ll likely need a new router every 5-7 years. There was a short period where 802.11ac routers were initially not very good and then the second-gen 802.11ac routers arrived and were substantially faster/better chipsets. If I recall correctly, the Apple AirPort Extreme was an example of the former and the Google OnHub an example of the latter. Now we have “mesh router sets” that trade features and performance for smaller sizes and cheaper components. Hopefully in the Wi-Fi 6E timeframe we’ll go back to performance routers that don’t look like spiders. It’s not that I don’t expect devices to get updates, but so much in networking gear depends on working around chipset bugs, I can’t imagine every router manufacturer supports their routers equally well... the number of routers I’ve bought that didn’t last even two years could probably fill a medium-sized box for me. Which highlights how hit-or-miss buying such gear tends to be, I suppose.

I know plenty of people that get along just fine with their 802.11b routers.

And even a 10+ year old b/g router can handle 4k streaming.

You're both making observations about different segmentatis of the market. For some gamers, latency (both %-ile and long-tail) is incredibly important -- in particular FPS / MOBA / fighting games. Upgrading every few years to stay on the cutting edge can make a notable improvement to these consumers, though I suspect routers are a bit less rapidly changing that things like CPUs and graphics cards. Think of these like the luxury smartphone users that will get value from upgrading their iPhone annually.

For consumers that just watch Netflix (even at 4k), latency is irrelevant and you just need the bandwidth to be high enough; for these usecases, sure, a 10-year old router is fine. This is the consumer that keeps their Android phone for 2-3 years until it breaks, and only then upgrades.

I don't think the GP was claiming that everyone needs to upgrade frequently, just that for some consumers it makes sense.

Both segments can coexist, and so it's valid for different products/companies to tailor their offerings to one or both. It's pretty clear that you shouldn't buy Google products if you want to set-and-forget for 10 years. But if you're going to upgrade to the newest revision of smart-home gadget in less than 5 years anyway, then Google's product-ADD isn't a particularly relevant negative to you.

Definitely a factor to make sure that consumers are aware of when purchasing though!

> For some gamers, latency (both %-ile and long-tail) is incredibly important

I've been on a reliable average 7ms and peak 9ms latency for twelve years. It's not necessary to replace your router hardware to get low latencies.

That’s not on Wi-Fi is it? Or your router is in the same room as your device, and it’s the only device using the signal?

I agree you shouldn’t rip out 5e cables to install 6e if what you have works for you and there aren’t any substantial improvements.

To that end, I’ve personally skipped Wi-Fi 6 and will wait for 6E before considering an upgrade. Even then, I’d like to have at minimum 3 devices that support Wi-Fi 6 before I upgrade. I’ve two such devices right now, so it’s not out of the question for me.

I’m in favour of keeping devices automatically updated even years later, but I suspect the only economical way of doing so is emulating devices in VMs, a giant test suite and automatic security update testing (no humans required beyond reviewing code diffs). But setting up that kind of automated testing infrastructure is a high bar for most companies, so it’s cheaper to put the cost on the consumer to consume again within 3-5 years for the next product update lifecycle...

> That’s not on Wi-Fi is it

I get this question every time when someone sees my latency, yes it is on WiFi. I don't use cabled and haven't for at least 6-7 years. My latency was always low except when I broke something in my local networking stack.

There are at this point 40+ other visible access points in my vicinity, and my router is two rooms away. My walls are made of concrete or brick.

Everyone should stop thinking WiFi is slow, because it just isn't.

As the sibling comment said, I was referring to WiFi. Should have clarified that.

Sure, if you're wired, there's no benefit to upgrading your router.

I was also referring to WiFi.

You need to replace all of your devices, too, in order to take advantage of those new WiFi specs.

Do you buy all new devices when you buy a new WiFi router? I don’t.

Right now I’m waiting to buy a new router even though I do have 2 of 30 or so devices with support for Wi-Fi 6.

Previously I did buy a few new devices to get 5GHz support from an ac router, after I moved to a location that had very poor 2.4Ghz connectivity. It can happen but it’s rare, I’d say?

I agree. 4 years is nothing to brag about. Not for the sorts of devices that people keep for longer. I've had my Asus router for 5 years, and I'm still getting updates. The hardware is still good, so the support should keep up.

I was on an iPhone 4s up until late last year. It was a backup phone and of course I broke my newer phone and needed something quickly until I could replace it. They stopped updating it however so I really don’t like to use it.

Your DSL router is getting security patches because the primary business model for DSL routers is for them to be owned by the ISP. It's thus an enterprise product which tend to get decades of support.

Depends on what you mean by "support". The OnHub never ended up using its smart home radios or built-in speaker for anything. I also ended up needing to buy multiple other boxes to work around Google WiFi's crashy PPPoE implementation and lack of VLAN tagging.

And then there was that day Google accidently deleted everyone's router configuration.

It worked great as a dumb access point. Everything else was shambles.

True. I stopped using mine because of the PPPoE support & VLAN tagging required for gigabit fiber optic internet was missing/didn’t perform well. And I remember having to reset not just my OnHub but my relatives’ too. But back when I had a standard coax cable connection, it worked great! :) I was able to chat with a live human when I needed support with mesh at one point.

I've been very happy with our home networking since ditching our consumer networking gear for Ubiquiti stuff. It's not very expensive, and it's worked extremely well.

Sad that Google never made amends for that; not even a token gesture like some Play credit.

Is OnHub no longer supported? Google never told me, and mine still works with Google WiFi app.

I made my statement based on when it was last updated. It appears it’s in a stable, security-fix-only mode. I consider this similar to discontinued products of other kinds — Google recognizes here that there’s more goodwill in making sure OnHub still works, but it really wants you to buy new hardware. This will be more visibly true when a Wi-Fi 6E device appears, I expect.

My Netgear router is about 10 years old and works fine. Once a product is released, it should remain supported for it's useful lifetime, not some arbitrary cutoff.

> my data gets scooped up and resold

Just to let you know, data is not resold by Google. It's used for interest-based advertising.

Potato, tomato.

Not really. If it's sold it's in the hands of god knows how many parties. If it's only used for targeting, it's still with one party. Maybe you find any sort of monetization of data icky, but the latter is definitely more preferable to the former.

This is something I don't get. Once I've clicked on an ad, the company has all my information to correlate with other databases they bought from someone other than Google, right? It takes me to their website, where they fingerprint my browser, drop cookies, etc. How is that materially different than Google selling them my info? If I interact with it in any way, they get all my info. (And in fact, with ads that include Javascript, do they even need me to interact with the ad?) This seems like a distinction without a difference to me.

That's definitely technically doable but that doesn't require Google in the mix. You clicked and went to a website that used those databases and your browser to fingerprint and identify you. Not sure how Google also having this data is somehow relevant to any of that.

The parent comment's point is that you can get info for the user from google by doing this. ie.

1. google sends you a bid request, which contains the user's demographic/interest information

2. you respond to it with a bid, which has a unique tracking id

3. if your bid wins, your ad gets served. from there, you can serve a javascript payload that fingerprints the user. at this point you can create a huge database mapping of device fingerprint and ip (probably unique, but also pseudonymous) to interest/demographic data.

4. (optional) you can correlate the data you've gotten with third party databases to augment the data you already have. I'm not sure how viable this is. You can get more demographic information, but deanonymizing the pseudonymous identifiers from step 3 might be difficult to impossible.

All of this only really works if you can fingerprint the user, which might be against the ToS of most ad exchanges. They're probably also on the lookout for this type of thing, because you can imagine this is pretty important business data. However, if you're big enough, you might be able to convince the ad networks to let you fingerprint for "fraud prevention" purposes.

You know who actually sells your data? The senators and congressmen who are constantly going on about how Google and Facebook are selling your data. I've donated a small amount to many campaigns, and I always use a different email address for each one since I own my own domain with a catch-all.

A few months or a year later, I get emails from groups like "stop republicans" or from local campaigns in states I don't care about to elizabethwarren@[mydomain.com]. And I only have used that email once to donate to Elizabeth. And it's not just one politician, it's all of them. Even ones we feel are so idealistic, like my Andrew Yang donation that I did based on his Reddit plea now receives emails targeted to my demographic about how it's my fault the democrats will fail to stop Trump, or lose California, or whatever.

Financial companies also sell my data. I get phishing emails to turbotax2014@[mydomain.com] which is what I used to file that year's taxes. [Edit: in retrospect, this one is probably just a hack or incompetence rather than intentional selling]

And which emails have never been sold or given out? The ones I use for Google or Facebook. Never in the past 15 years.

This is a really good point; I have had the same experience. The fact that every political campaign is happy to sell my personal information to other groups has completely turned me off from donating.

The same applies to volunteering. I made the mistake of giving my real phone number to a congressional campaign that I signed up to volunteer with. Since then I've gotten dozens of spam texts from many different organizations.

Its not an infosec problem, but a privacy problem

In what way? What privacy breach are you worried about?

For example are you worried that a real human at your home security company will be invading your privacy? When they do are you worried that real human will take personal action against you? Or are you worried about automated systems building knowledge about your life based on home security data? Or are you worried about your home security data used in aggregate? Maybe you are worried about attackers getting access?

Internal infosec and privacy controls/policies are required to protect you from some of these. Some of these threats may be more likely from Google, but I'd say some of them are more likely from other home security providers (keep in mind that a lot of home security products are actually intended to have another human monitoring your home).

I would say these are the obvious ones for Google:

> Or are you worried about automated systems building knowledge about your life based on home security data? Or are you worried about your home security data used in aggregate?

Good breakdown. For me the main issues would be around how technically capable Google is to ingest all of the content coming from ADT monitoring infrastructure, extract meaningful information and then merge that in with their existing knowledge of my life.

There’s a number of direct answers below, so I’ll give you a more philosophical one: privacy is in direct conflict with Google’s business model and thus it is impossible for them to deliver privacy in good form without jeopardizing their business.

I'd think it's both in the age of the smart thermostat that has a camera/mic built in.

... and motion sensor

> I'm on a "low Google diet" myself as I'm not at all a fan of how my data gets scooped up and resold.

Does google ever actually resell data? Or just monetize it? There’s a big difference selling the personal information on someone versus letting people buy targeted ads.

Depending on the level of targeting allowed the difference is meaningless. If you can target people based on something like sexual preference, you’ve effectively sold out everyone that clicks on that link.