Nest, the company, died at Google I/O 2019 • Hacker News Board

Re: Nest, the company, died at Google I/O 2019

#111

Post by adjkant » Fri, May 10, 2019, 2:40 PM UTC

If there's anything to be learned from Nest, I think it would be interesting to se a full open source stack for things like Nest as well as Google Home / Alexa. If the NLP is right, the actual functionality is not hard to replicated means that third party integrations can be easily added/customized. Then you also eliminate the always on problem with everything going through your own servers.

What I'm imagining (all open source):

- Hardware Speaker/Microphone device (can be multiple developed of course on one standard)

- Device OS based on Linux

- NLP engine

- Easy Plug and Play Server Box

- Library / Packages for third party integrations (say Gmail et al, Spotify, Amazon, Weather.com)

- An open app store type system where all of these "skills" can actually be used with any device instead of only existing in one ecosystem

Does something like this already exist?

It would obviously take more technical know how, but I think done right it could really be a big step towards keeping the usefulness of assistants and smart home devices while building in control and security both in lasting time and attacks.

Re: Nest, the company, died at Google I/O 2019

#112

Post by 2Ccltvcm » Fri, May 10, 2019, 2:40 PM UTC

justin66 wrote: post #91

daxterspeed wrote: post #3

This article was a surprisingly difficult read. It felt like every other word was a brand name, and the underlying confusing naming scheme surely didn't help. On top of that the author flip-flops between the issues of Nest going away (loss of privacy, and product ecosystem) while excited about Nest functionality moving to Google Home (Google Nest?). I think the main takeaway for me is that it'll probably be another d…

Of course, a decade is optimistic. You could have felt the same way and written the same thing a decade ago. It's possible that there will never be any stability in the smart home ecosystem, and constantly buying and throwing away electronic junk is simply the norm.

Stability in the ecosystem means loss of revenue stream. Disorder must be preserved to generate revenue reliably.

Re: Nest, the company, died at Google I/O 2019

#113

Post by simonebrunozzi » Fri, May 10, 2019, 2:41 PM UTC

ansible wrote: post #78

zelly wrote: post #46

We are all used to SaaS randomly getting obsoleted, but when you buy a physical device you kind of expect to own it forever. But no, cloud and IoT means even your physical stuff are SaaS products. Just stop buying the trash.

You might enjoy Cory Doctorow's recent book "Radicalized": https://craphound.com/category/radicalized-full/ The protagonist starts out by jail-breaking her toaster, which was designed on only toast authorized bread. However, the company that produced it went out of business (taken over by a hedge fund), and the authentication servers were shut down, so now the toaster couldn't toast anything.

That's hilarious, if not sad.

Re: Nest, the company, died at Google I/O 2019

#114

Post by magicalhippo » Fri, May 10, 2019, 2:42 PM UTC

Vaslo wrote: post #15

Has anyone tried Home Assistant to get around things like this? Or are you locked in proprietary technology anyways?

Home Assistant can interface with many different services. I use it with Z-Wave and the (undocumented) cloud API for my A/C, as well as pulling some data from weather services etc.

There are a lot of different Z-Wave devices for controlling lights, motors, heat, as well as sensors for motion, smoke, luminosity etc. Zigbee is another standard which can also be used with Home Assistant, which is mostly for lighting AFAIK.

The way Home Assistant is written it's relatively easy to write a new service which exposes devices that can control or sense. Sensor readings and control is unified, so you can mix multiple platforms in your automation. You could have a Z-Wave motion sensor trigger a Philips Hue (Zigbee) light if the weather service (HTTP) tells you it's past sunset.

You can check out the supported components here: https://www.home-assistant.io/components/

Re: Nest, the company, died at Google I/O 2019

#115

Post by wongarsu » Fri, May 10, 2019, 2:43 PM UTC

scarface74 wrote: post #82

jerf wrote: post #68

Earlier quoted context omitted.

I've occasionally pondered screwing around with various wifi-embedded things like the Arduino, but then I think about how long the light switches have been in my house already, and I find myself wondering, will even Wifi last long enough? Will today's Wifi still be going in 40 years? Dunno. Maybe. But for me to be interested in something it still has to be something that is going to pay back my fiddling in probably t…

I had my house wired with gigabit Ethernet when it was being built. I don’t know how long will that be enough.

According to Wikipedia Cat5e became the common ethernet cable around 2001, it's still in common use and capable of carrying gigabit, or over short distances even 10Gbit ethernet. Today at least in my bubble I see Cat6 and Cat7 deployed instead for 10Gbit and beyond.

I think everyone is getting at least 20 years out of their Ethernet connections, likely more. That's not as durable as electical cables but good enough for most cases (plenty of things need updates after 20 years in a building).

Re: Nest, the company, died at Google I/O 2019

#116

Post by hunter2_ » Fri, May 10, 2019, 2:43 PM UTC

dustinmoris wrote: post #97

Smart people don't buy smart home devices. My relatively poor grandparents and my slightly less poor parents had a hard working life, but none of them had to sleep in a dark cold room because the switch to turn the lights on or the heating stopped working. Ever. This problem has been solved by other people many many years ago. Last month I watched a documentary where a billionaire was showing of his multi million man…

Lighting hardwired to a stuck system and touching physical controllers (phone or otherwise) as you describe is one possible setup. On the contrary, I just have some table/floor lamps with smart outlets between their plugs and the outlets that can be quickly removed in case of failure, and touching my phone is only in case of voice (Google Home Mini) failure. Neither of which fail often, like much less than once a year. The convenience when half asleep is great.

Re: Nest, the company, died at Google I/O 2019

#117

Post by smashah » Fri, May 10, 2019, 2:44 PM UTC

adjkant wrote: post #111

If there's anything to be learned from Nest, I think it would be interesting to se a full open source stack for things like Nest as well as Google Home / Alexa. If the NLP is right, the actual functionality is not hard to replicated means that third party integrations can be easily added/customized. Then you also eliminate the always on problem with everything going through your own servers. What I'm imagining (all o…

Home assistant

Re: Nest, the company, died at Google I/O 2019

#118

Post by thebiglebrewski » Fri, May 10, 2019, 2:45 PM UTC

dustinmoris wrote: post #97

Smart people don't buy smart home devices. My relatively poor grandparents and my slightly less poor parents had a hard working life, but none of them had to sleep in a dark cold room because the switch to turn the lights on or the heating stopped working. Ever. This problem has been solved by other people many many years ago. Last month I watched a documentary where a billionaire was showing of his multi million man…

Por que no los dos? Why not have perfectly functional on their own devices that can also be software addressed? Got that in my house and I'm a happy camper, when I have renters they don't even notice/have to use the smart features!

Re: Nest, the company, died at Google I/O 2019

#119

Post by magicalhippo » Fri, May 10, 2019, 2:45 PM UTC

linker3000 wrote: post #106

magicalhippo wrote: post #79

Earlier quoted context omitted.

I'm mainly going for Z-Wave plus, using Home Assistant as controller for now (though still evaluating that). My A/C is on Wifi though, and the cloudy controller web page is using a REST API (undocumented but simple enough to sniff and fake), so that's the only outlier so far.

I really want to love HA, but I find doing things in Node-Red so much quicker and easy to comprehend due to its visual nature. After a gap of about a year, I installed HA again a few weeks back and tried to setup geofencing using first bluetooth and then third party location services, but gave up due to the inconsistent determination of whether I was home or not. Using Node-red and a couple of tutorials I had things…

Yes, I find HA a bit clunky to use as well. Mostly because the scripting is quite tedious and relies on templating yaml files.

I'll have to give Node-RED a go once I find some spare time to tinker.

Re: Nest, the company, died at Google I/O 2019

#120

Post by AlexandrB » Fri, May 10, 2019, 2:45 PM UTC

dustinmoris wrote: post #97

Smart people don't buy smart home devices. My relatively poor grandparents and my slightly less poor parents had a hard working life, but none of them had to sleep in a dark cold room because the switch to turn the lights on or the heating stopped working. Ever. This problem has been solved by other people many many years ago. Last month I watched a documentary where a billionaire was showing of his multi million man…

> It's just not a problem which I think people have, but somehow the consumer industry has convinced so many fools to buy cheaply fabricated, badly secured, even worse programmed and often not long supported smart home devices which add absolutely no benefit to anyone's everyday life and cause lots of problems.

This. Every time I hear a friend or someone on a podcast talking about their smart home they talk about how now they just need to find a new smart hub/lightswitch/integration service and everything will come together and nirvana will be achieved. Life's too short to spend that kind of time and energy on turning lights on and off.