Hacker News new | past | comments | ask | show | jobs | submit login

They can easily fix this. If they don't, it speaks volumes about how Nest/Google/Alphabets values customer relationships. Take all the other google services and imagine what would happen if google were just as reckless, especially with long life products like home installed devices.

Home automation installation services or even contractors would also be cautious about Nest.




These shenanigans by Google and the numerous reports of the problems with Nest and their smoke detectors have basically guaranteed I don't include Google devices in my home automation setup. I would rather manually muck around with a raspberry pi and some random sensors bought online than rely on a Google product at this point.

There's a lot of room for a vendor to come in with open products, a pledge to maintain that open-ness, and decent integration to really take over the home automation market. I really feel like all the existing products are too proprietary to be of any interest to us home automation geeks.


Yeah, unfortunately, there's a huge middle ground not being tackled. I use INSTEON devices, which are proprietary, but can only interface outside through my PC's serial adapter (or a cloud hub, which I obviously didn't buy). I want consumer-quality switches and such in my house, so this was the obvious choice.

But then the open source hardware stuff is almost entirely limited to 'build it yourself' boards. And sure, I could buy some Arduino and Raspberry Pi hardware (I actually have a lot of it lying around) and spend a bunch of time working on it, but then it wouldn't be a nice wall unit like my switches I have.

There's a huge open space here. It's a market just asking for someone to step in and make products.


I have ALC (On-Q/Legrand) switches for the same reason. Hard-wired control, but reasonably nice (at least not clown) looking and operating switches in manual mode.


That's a key thing too that things like Wi-Fi lightbulbs and crud can't offer: Your light switch should still ALWAYS work like an ordinary light switch!


Companies sunset SAAS products (with subscription fees, one time setup fees etc) all the time when the cost to support them becomes too high, growth stalls etc. Clearly Alphabet/Nest could have done a better job of customer service and offered users a partial refund or discount on a replacement product. That said, is it realistic to expect a company to support a software service indefinitely (when the business case no longer makes sense), just because it's tied to a piece of physical hardware that you paid $199 for a few years ago?


Depends how you sunset it. Depends on the reasonable life of the product. In Europe at least Consumer Legislation has an expectation of a reasonable life. It's not reasonable for instance for lots of your fridges to die after only a couple of years. Car makers have to provide spare parts for, I think, six years after the end of production of a model.

I think a lot more could be done to sunset products gracefully, and with consideration for the paying customer. We'll be ceasing updates for this in a year, we'll keep the server active for x years. Be aware there will be no security fixes. Like the phased EoL of Windows releases.

I think quite a few examples of smart things are getting dangerously close to, or already passing, the line that would fit the legal definition of unreasonably short. SaaS only? We need to wait for case law and legislation to catch up I think.

Eighteen months life, then intentionally break your $300 product, with only 6 weeks notice? I think it would make an interesting test case with reasonable chance of success.

I trust Google very little as far as sunsetting their many releases.


Damnit. If you're going to remove my ability to use it after 18 months, then just sell it to me on an 18 month contract for $20/month and let me decide whether to sign. It's already clear I don't own the hardware, so why not make it official?


It would be tough for a hardware startup to do that, since they must buy the hardware upfront before they ship it to the user (and returns always cost $). What if the manufacturer stated clearly that your upfront purchase includes a free license to the SW for x months?


... then I would fully expect that I can replace their software with that made by others. Then feel free to make a business decision that part of the original product is no longer profitable and you're leaving the marketplace for sw/support - I still won't like it but at least the consumer has options to live with their not-fully dead hardware.

As a consumer - "hey too bad you didn't want to make a business of this but you suckered me in with your shiny new fangled thing and I gave you money for it - I didn't give you permissions to then decide that you didn't make enough money so now you're going to take your ball and go home leaving me with an expensive door stop"


I agree that it's annoying when a company sunsets any product (HW, SW or other), but why is it so much more annoying when it's a hardware product? Is it because you still have the physical thing? Most SAAS products have upfront/monthly costs, and when they fold consumers lose these sunk costs, but nobody is writing medium articles about that...


Again, depends why.

They went bust or are closing? OK that sucks, but it happens. Can't really be helped. Even if it's a hardware thing it's going to happen if there's a server-side component rather than entirely stand alone. Best case release some code, specs or open source to enable some hackery to keep hardwareThing viable.

We got bought, OldThing is boring so it's closing in 4 weeks, we're truly excited to be working with FaceGoog on the future of NewThing. Seriously not ok. People should be writing Medium articles about it. And yelling.


> but why is it so much more annoying when it's a hardware product?

Because applying SaaS business model to hardware is simply fucked up. Also, for many of those products one has to expend significant effort to make them not hackable by end-users, so if a company decides to make a proprietary, cloud-dependent device that then gets bricked when they move on, it only shows they actually worked hard to fuck their customers. Not to mention that in many of those products the cloud is not needed for any sensible reason and is in fact bad engineering - it's included only to make more money off people.

Sadly, as long as customers play ball this nonsense will continue - the market sells what people buy.


Having worked on "smart" hardware, this can't be done. For the reason of regulation.

Essentially, the law demands that no matter what you do with the software, the product won't malfunction (not just "not explode", it needs to work). Obviously that means no software that isn't made by the same engineering team ... also once the hardware is out that means no changes to the software unless someone does make their fridge explode. Even if they change the temperature 50 times a minute with the door open that can't affect the lifetime of the product. Needless to say, you block this in the software.

Unless courts start putting responsibility for the things people do with their own hardware on those people, this won't be forthcoming.


This makes sense into you throw OTA updates and cloud services into the mix. I'm not sure how those would even be allowed by the "the software must not change" rule. But even if you interpret it generosity, shutting down the service or pushing an update that bricks the device is very clearly an update that breaks functionality. Wouldn't that be forbidden by the above rules as well?

Also, there are "hackable" products on the market, e.g. the OpenWRT routers. Those still seem to be allowed.


Yep I would agree that breaking updates are a risk. In the case of a cloud service though, I'd bet the contract would specify a minimum maintenance duration very clearly spelled out, and that's going to be a few years, no more.

OpenWRT routers have been in the legal clear zone for a while (due to FCC exception for the WiFi bands), but that's changing [1]

[1] http://www.cnx-software.com/2015/07/27/new-fcc-rules-may-pre... (FCC page https://apps.fcc.gov/oetcf/kdb/forms/FTSSearchResultPage.cfm... )

People are somewhat confused about this. For instance you can change the firmware of smartphones, no ? Well, no, you can't. You see, what people call a cell phone contains a 100% locked down actual cell phone, and a PDA that's completely independent (and in some cases, more than one locked down system. E.g. NFC is it's own locked down system as well). These are fun things, they're smaller and smaller. In the last 2 years the actual cellphone, everything except the antenna fits in 0.5cm x 0.5cm x 0.1cm package that actually contains multiple chips. These things are connected to the audio devices of the phone, and therefore have sole control over what comes out of the speaker and where the microphone goes. They also have a serial network interface these days for 3G/4G, and the sim connects directly to these chips.

So technically android (or ios, or windows phone, or blackbery, ...) does not run on cellphones, it runs on the PDAs that we insert into cellphones. Therefore it does not need approval.

It also means that Android/IOS do not actually control (for example) the audio on smartphones, or whether a call is answered. Or whether, when asked for the location of the phone or the contents of SMS'es, or to call out and send out the audio the microphone receives, whether those requests are answered. Android/ios would never see those requests (though you may find the microphone suddenly doesn't work anymore in apps). And yes, this works even when there is no sim in the phone.

Whilst there are cellphones that don't provide law enforcement (I wonder if it's only law enforcement though) with these features, they're older models and are getting phased out. If you read this site, I bet your phone has these features. And like all features you don't control, you wouldn't like them if you knew what they were.


They're not a startup any more. I agree, they were in a tough situation. It's difficult for Nest, it's more difficult for consumers.


> when the business case no longer makes sense

Doing this kills ALL future business of this type. Who would buy such a device knowing this could happen?

Unless they don't plan on making consumer appliances anymore.


I don't like this at all, but I think you've got the calculus wrong. If you truly believe in IoT, then the percentage of future Alphabet customers who will ever hear about this scandal is tiny.s


Oh don't worry. I'll do my part in making sure more people hear about this.

Seriously though. Highlighting the ridiculousness of the current approach to IoT (including laughing off startups in this area that follow the SaaS/cloud model with hardware) is probably the only reasonable way of this mess.


I'm not saying it's an ideal outcome for a company to sunset a product, but why is the consumer expectation so different for pure software vs a connected hardware device (i.e. software + hardware)? Consumers experience sunk cost when any product gets sunset, whether it's hardware or not.


What bugs me is how many companies these days want to include reliance on the cloud for features that don't require it at all, in the interest of gaining access to more customer usage data, but then when they turn the servers off the product becomes useless.

This has become a key thing that I look for when buying new products. If the product relies on the cloud for a feature that I think should be entirely offline, the product will be immediately returned and a negative review will be left on the resellers site.

A perfect example was the early version of the Lockitron device. It required cloud access but had a one time payment. Based on that information alone I could not justify the purchase to myself even just to play with because it's a bad business model.


Fitbit. I fail to see why tracking my health, location, etc. is a cloud function, beyond selling my data. At least give me the watch free if you're going to sell my data too.

They then add insult to injury by making me sync via a dongle /connected cable to my laptop; at which point I could just as easily sync to a desktop client vs their otherwise "disconnected cloud".


I wouldn't have such a grudge against fitbit if they would make it easier to log to your own server. Even if I had to hijack DNS to accomplish it, I wouldn't mind. Its not like many people would take advantage of the option.

But the fact that they force you to use their servers to store your data so that they can turn around and sell it, frankly I hope they get sued for something. Anything.

There should be a law against selling customer data without the expressed consent of the customer whos data is being sold, per purchase. Even if the company is acquired.


The Turtlebeach Audiotron relied on accessing Turtlebeach's servers to function. They abandoned the Audiotron, and eventually the server, but did provide a patch so you could still use the Audiotron.

I still use the Audiotron every day. Still haven't found anything better.


How do you determine if a given feature 'should' be entirely offline?


Unlocking my front door? That has to work without internet access.

Turning my lights on and off? That has to work without internet access.

Logging my heart rate et all, ala fitbit? Nope, I should be able to use a server of my choosing, requiring that I use cloud servers for something I paid a one time fee for is an instant deal breaker.

A service that logged all of my access times and determined whether or not I was home based on the result? I would prefer that be run from my own server even optionally with a DNS hijack, but I would understand the need for the cloud in this situation.


Thanks for the info! Have you found many devices that meet this criteria? I can't think of any smart locks that run locally.


Though I haven't confirmed that this was fixed in new models, when I was first reviewing the Lockitron I contacted their support folks with my concerns about paying a one time fee for something that waa easentually a service and about being able to run my own server and (one of) the founder emailed me back and explained that in the next version my concerns would be addressed. That gave me hope for their next iteration but I haven't been back to them yet. I look forward to checking them out again.


Mostly it is because "Connected to the point of bricking" is very new.

Most "connected" devices simply have lesser or limited functionality when services they are connected to disappear.

It also comes down to Ownership. When I buy a device I expect it work until it gives up its electronic ghost. When I subscribe to a service I expect it to work for the month I paid for.

Ofcourse I would have never bought that hub in the first place because I require my products at a minimum to have a open API that can be used for Compatibility later.

Wall Garden should be resisted by consumers, it is sad most people do not even look at that, hopefully cases like this will push consumers to DEMAND manufacturer have open and inter operable systems


Interesting. The reality is that every connected device has an ongoing cost to the company that sells it. Most consumers probably aren't aware of this. Depending on the architecture, # of users etc this could be some fraction of a cent/month in AWS costs or much more. Should Connected Hardware companies/startups disclose this upfront when making the sale? i.e. say "Your $199 purchase includes a free license to our software for x years."


>>The reality is that every connected device has an ongoing cost to the company that sells it.

Not true, that is only the case for devices where the company wants to have exclusive control over the connection and resulting data, There are all kinds of open devices sold where I can connect to servers run by community, myself, or the company (many provide the cloud service and the device separately, so I buy the device then i have to buy a monthly service as well)

>Should Connected Hardware companies/startups disclose this upfront when making the sale? i.e. say "Your $199 purchase includes a free license to our software for x years."

That should be the minimum disclosure requirement.


Hardware has that upfront cost which consumers take as this should work for x amount of time. In the case of electronics I would hope they last a significant amount of time. If my hardware can work still at least give me a way to write my own code for it or use it as is with the "cloud" portions disabled/removed. If I get a new cell phone my old one doesn't need to become a paperweight because the carrier isn't pushing out updates. I can still use it as is or update it myself.


Not sure the analogy works. You would still need to pay some carrier an ongoing service fee to use your cell phone. Here we are talking about a connected device where the software service is offered for free.


Only if you want to use the carrier.

I have old phones that can't make telephone calls, but they still work fine as connected devices via wifi. They still work as cameras, mp3 players, gaming machines etc.


>Hardware has that upfront cost which consumers take as this should work for x amount of time.

What is X?


I think it depends on the product and the cost I am paying. x for a checkout lane electronic keychain that I paid $1.99 for I don't expect to last very long a month or two maybe. When I am paying $300 I would hope to get many years out of it before there may be hardware (rather than fixable software bugs) failure from a bad capacitor. Even then if you have the electronic knowhow you can replace the parts and keep it working. What lifetime do you expect out of a new TV you purchase? Should it stop working after a year or two or just continue to work as long as the hardware works?


The consumer expectation may be different for web sites (as those were always closer to services than products and so you never "bought" them in the first place).

For traditional desktop software and standalone apps, the expectation is not different at all. I still own a lot of old pc games that I bought at some point and play occasionally and would find it extremely irritating if they suddenly stopped to function without any acceptable reason.

Recently, things may be changing with mandatory auto-updating, subscription models and many apps just being front-ends for server-side processes. But as far as I know, most of those changes are pushed by vendors/developers and accepted by consumers as a necessary evil (if at all). It's not at all that customers would suddenly expect something different.


Forever? No. But what is reasonable to you? 0 days? What if it's a home appliance I paid around $5,000 for, like my most recent furnace install? Also 0 days?


I see it as an extension of their consumer software policy.

Have little to no customer service, even when money is involved, letting products stagnate for years and frequently pulling the plug on services.

Which is scary, because this hardware is dependent on Google-owned software and services.


After Google purchased Dropcam, they censored a thread about it's inability to save video feeds locally.


Empathy is not a core competency of google as an enterprise.


Google has always had poor customer service, this is nothing new. Most of their (profitable) product lines don't need customer service representation, so they don't care about it.


It needs to be understood that home automation appliances are going to be assumed to be long-life products. Nobody thinks about their light switch being an old version and needing to be replaced.

I'd say any home automation installer should be very wary of Silicon Valley-based automation "services", and focus on robust products which have been available for a long time.


I've done such installations. The simple solution is to use Honeywell products. It's good stuff that works as intended. The downside is its not even in the same ballpark as modern google/amazon/startup type devices. The customer gets what the customer wants so if someone wants ultra modern then conservatively designed products don't fly. Best advice is to vote with your wallet and stop trusting google products.


Good point.

My Hue lights work without an internet connection, and can be controlled from a device on my local network. And they are also made by an old-school company, Phillips.


There's ups and downs to Hue. If you recall earlier they tried axing out support for third party bulbs from their app. But they also deserve some kudos for basing their app and devices on a open platform to begin with.


Yes, but once people protested, they admitted their mistake and reverted it. AFAIK their position now is that they won't go out of their way to support third party bulbs unless the manufacturer joins their "friends of Hue" program, but their hub firmware isn't blocking them and should work with other bulbs that meet certain specs.


Home Automation is still niche. They probably hope to make up on the volume. It does not matter how badly they treat their current customer: if they find the iPhone of home automation, they will have hundred of millions of customers to make up for it.

They probably hope the guy will find the next big thing. That's more Google working internally as a VC in markets to be. Not that I think it is alright from such an established business like Google.




Applications are open for YC Winter 2020

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: