Hacker News new | comments | show | ask | jobs | submit login
Netflix Switch – dim lights, turn on the TV, order food, and silence your phone (netflix.com)
399 points by PascLeRasc 849 days ago | hide | past | web | favorite | 84 comments

It's really great to see a company's engineers given the time and resources to do some fun tech project like this. Sometimes it seems like unless a project is about directly increasing revenue, its incredibly difficult to get buy-in from management. These projects are great for culture.

Netflix seems like a great company to work for. Recommend reading Reed Hasting's presentation on their culture: http://www.slideshare.net/reed2001/culture-1798664

For another perspective (perhaps a more nuanced/less flattering one), NPR's Planet Money recently did a great episode on Netflix's work culture [1]. I'm not so sure I'd want to work there, but I'm sure it is great for some.

[1] http://www.npr.org/sections/money/2015/08/28/435583328/episo...

When I listened to this before it made me wonder if Netflix's culture will get same sort of negative treatment and publicity that Amazon has been getting recently.

That was super interesting. Thank you for posting it. The DVD/Streaming split was an especially interesting period for Netflix. I didn't know the back story behind it. For anyone curious, it's at the end.

sounds like a bunch of platitudes with a lot of HR speak (a lot of which I REALLY disagree with)

And for having the "top talent" they have little to show for it on the technical side

Little to show for it? What percentage of total Internet traffic would you estimate that Netflix handles at 9PM eastern on a Friday night?

I just think the service hasn't evolved much since they started streaming nor have they really branched out.

Their success is due to their ability to secure rights to TV shows and movies - so thanks legal team

Netflix has massive talent density, no doubt it. The only downside is that they grind through people like crazy.

From what I've heard Netflix has a great culture surrounding it.

Indeed, this is a cool project. Netflix seems to have a great work culture. The button design is neat, it can be used for different projects. A button can do many things :)

I agree, I didn't realize Netflix did side projects like this, it definitely makes the company feel personable.

Is it really just a fun project? Or is Netflix trying to use this as the nose of the camel to get into the house automation business?

I've always thought house automation was cool, but a lot of trouble and likely to wind up a waste of my time and effort once I stop using it in three years because whatever standard I was using became obsolete. But if Netflix can start automating my house one bit at a time and slowly sneak up on me...

It's so far from a scalable product, I would guess it's not even a good test for market fit. Only a small subset of hackers/makers will tackle this. And, there's nothing to even buy from Netflix here...it's all stuff from other vendors.

I'm not saying it isn't on their radar, but this particular project is clearly for fun.

It could be fun, plus a recruiting technique, since it creates awareness around Netflix's culture, and I am sure they hire a lot of engineers.

I honestly can't attribute any profiteering here. This is pure geek fun. Reminds me of the days of the Netflix Prize. http://www.netflixprize.com/

I am sure it was a blast, but it was NOT done "for fun" to amuse the engineering staff. Nor is it an effort at "home automation".

This is part of a marketing campaign not unlike the office depot "easy" button, except they've designed an actual button that sort-of works.

The point is to get people excited about the idea of a Netflix marathon and create some buzz among geeky subscribers. It doesn't matter how many people actually create or use "the button".

Well, I'm sure that was an excuse, but seriously. Do you honestly think the kind of geeks that'd do this are really that important to a company like Netflix these days? Maybe at the beginning...

If I had to attribute anything, it's more of a recruiting move. They want clever hacker geeks working for them. Maybe they are moving into home automation and this will be an interview question? I dunno how effective that is, though. I think someone just said, yeah, screw it, let's do it!

That's a weird couple of sentences, given that the Netflix prize was entirely trying to improve their core business?

Yeah, maybe. I don't think they ever used it though. I think it was just the hacker / experimenter part of Reed that wanted to do it.

A little viral-ness doesn't hurt either. If it makes a dozen or so blogs today and reinforces the idea that Netflix is a service you cuddle down with in your den, that's not such a bad thing.

Step 3 is 'Select the Microcontroller'. I don't understand how they plan for this to be a pure profit driving activity.

Well, here's something that definitely has no market value: http://www.nbcnews.com/tech/gadgets/8-bit-binge-hack-puts-ne... ;)

> Is it really just a fun project? Or is Netflix trying to use this as the nose of the camel to get into the house automation business?

I doubt they would piggyback on others' home automation systems and standards if that were their goal.

We're in the golden age of DIY home automation. Wifi-enabled microcontrollers like the particle photon are a game changer; it's never been easier to make a connected device yourself.

Using the Photon I've built a device that connects to my doorbell and sends a push notification to my mobile when someone rings. (Might be useful when you're in the back yard and don't want to miss the delivery guy)

Yes, most of the stuff is still gimmicky. But when you're tinkering, it's not necessary to turn everything into a product that will change the world or earn a billion dollars. We can just have a bit of fun, and enjoy the things we can do.

I think that hackerspaces might be todays equivalent to the homebrew computer clubs where enthusiasts dabbled with early home computers...

I agree and disagree with you at the same time. IoT, especially with things like the Photon and ESP8266 are major game changers. However, the idea of a baked-in WiFi stack with little/no way to update protocol bugs is rather scary proposition. And I'm also against giving IoT devices a directly routable IP address, even if they are behind a NAT.

I'm looking more at things like the nRF24L01+ board combined with MySensors, which seems to be rather mature. It is a self-healing mesh network with encryption and signing support. And I can get all the data from the net mesh with a serial gateway (inputs/outputs via USB serial). I can connect that gateway to a Linux machine running Node-red and use the MySensors encoder/decoder.

Hook up a relay and you can toggle electronics on and off, or you can make a manual high voltage switch using a servo and 4 contact switches. I'm also using PIR for occupancy along with temp sensors for zone acquisition. I'm looking at developing my own thermostat as well to see if I have the chops..

FYI for the DIYers, there is a little project that I contribute to on the side for fun. I also use it to do things just like the title suggests:


That looks fantastic, I've been hunting for a home automation controller with a modern architecture for ages, and I think you've just pointed it out.

Looks cool! One question - Why android switches on an iphone app? Is it just for the screenshot?

This isn't a native app, that is a very responsive webapp. It uses nuclearjs and the polymer web frameworks, but the backend is all evented python.

You can check out the demo here: https://home-assistant.io/demo/

This is really cool, thanks for sharing.

Speaking of which, do food APIs exist? I've often been tempted to make a generic pizza API that — by limiting options — will deliver a pizza to almost any address via Just-Eat or whatever is local to you.

In 2005, my friends Dave and Adam built a pizza button in our dorm, it actually used the modem to send a fax to a local pizza place because that's all they had. Pushing the button ordered a cheese, but there were two auxiliary buttons underneath to select a topping via a Huffman code. (0 for pepperoni, 10 for sausage, 110 for anchovy, etc.)

I presume there are better APIs today :)

There was this old Perl script for ordering Pizza at Domino's but I don't think it used an API (web scraping if I recall correctly).



Haven't really seen any food api's. It would be cool if it did exist though.. But I don't think Just-Eat like services will bite since you'll be driving traffic away from their website..

What's the problem with the traffic being driven away from the web site? As long as it would be coming through their API (think of centralized API for local food joints), and they get their cut it's all good.

But I see valid reasons for not providing APIs like that. Imagine someone creates a gadget like that and there's a bug, instead of 1 pizza it orders 10 of them. A delivery guy comes with 10, and no one is willing to pay for that many. It wouldn't take long to infuriate enough food joints.

Well, yeah, you could pay with credit card when placing an order... But I never do that, I just avoid using my card on web if I can easily do it :)

I think Seamless offers an API: http://www.seamless.com/business-development/

Away? I would probably use them via screen-scraping. However whilst they would be getting the order, they wouldn't be gaining the end-customer as I would be proxying that.

Gives new meaning to Netflix & Chill.

This can already be achieved with Flic: http://flic.io Flic is a Bluetooth button which uses your smartphone to communicate with smart electronics. If you have Philips Hue or LIFX smart lights and a Logitech Harmony Hub (for IR), this would work straight out of the box! (Disclaimer: I am associated with Flic)

But you're not shipping Flic yet, are you?

Does that require a phone nearby at all times?

I'd also be interested in this. Would love such a button that could connect via WiFi or generic BT on a Raspberry Pi or something like that, rather than a smartphone with a specific app.

For the time being, you need a smartphone. But we are currently looking into other areas to expand the use of our product.

Nice, seems like a cool startup.

> order food, and silence your phone

Really? Around me, delivery drivers have pretty much given up on doorbells. If your phone is silent you may not get your food.

In my city, it's fairly uncommon for any delivery person (pizza or shipping) to ring the doorbell, but it's quite likely they'll knock.

Lots of delivery drivers do call phones, but I see that happening more often when there are many places you could be (e.g. ordered to work, but are they going to come find you at your chair?)

Delivery drivers don't use doorbells where you live? What do they do, call you instead? That seems odd to me, even Fedex and UPS ring my doorbell when dropping off a package.

To be precise, it's Chicago. Being in the city means most deliveries will be to apartments. The door buzzers are often unlabeled, mislabeled, or just broken. But even if you're in a townhouse with a door facing the street, most drivers are firmly in the habit of calling.

Can only be called a Chill Switch.

Others have mentioned the Flic and Amazon's Dash buttons. Here's the one I'm most excited about, which should (when they have an IFTTT channel) make things like the Netflix Switch doable for non-nerds: http://www.dropletlife.com/

A number of years ago I made a project that let me do this for ESPN using IR (Ken Sherrifs awesome library) and x10 units. Rather than a physical button I just had a shortcut on my phones home screen.


It would be cooler, easier, and would yield better battery life to reprogram an Amazon Dash Button [1] to do this. Most Samsung TVs can be controlled from the network side, as well as with an IR remote.[2]

Of course, the Netflix button just brings up power and puts you at the Netflix menu. At that point, you need another device to tell Netflix what you want. Authorized Netflix-compatible TVs already have a Netflix button.

So the logical thing is to have something listening to the TV's state to dim the lights, close curtains, and such. There are home theater accessories for that.

[1] http://www.amazon.com/b/?node=10667898011 [2] http://www.gemote.org/

This all sounds great -- until, for example, you are trying to eat a snack while also watching TV. Or one of a dozen other things where the automation gets in the way. Or, equally worse -- it's not perfect. Sometimes the lights don't change, sometimes there is a delay (so user presses the button multiple times) and suddenly everything catches up and curtains are closing and opening and closing and opening..

Any of these cause WAF to go to zero, at which point, you rip it all out.

So many times when it comes to home automation, otherwise rational software developers forget what people actually want to do, and skip to what the technology makes possible but doesn't actually make anything easier.

There's HDMI CEC which could let you do that even better: the TV remote could have a "cinema" button that sent the signal over CEC to an RPi which controls the lights and curtains.

Step 15: Spend 2 hours browsing for something to watch

Reminds me of something I built: http://garyroumanis.com/2015/03/15/TVRemote-app/

The silence your phone step[1] is much easier if you use something like Tasker[2].

[1] http://makeit.netflix.com/the-switch/do-more#silence-your-de... [2] https://play.google.com/store/apps/details?id=net.dinglisch....

My guess is that this was the result of an internal hackathon. Makes me think working at Netflix could be fun.

Yeah, they've had other hackathons before with good results: http://thenextweb.com/apps/2015/03/10/a-netflix-hack-day-pro...

Most likely the initial prototype was done during an hacktaton... I remember many years back someone that came up with a small TV on a toaster so we could say you can watch Netflix on your toaster, later on that toaster made it into a T-Shirt for the next hackaton!

Any ideas what they might be using to mill the wooden enclosure? I have a few projects that need enclosures and it's been a few years since I looked but the CNC mills were still thousands of dollars.

Why do you believe that wooden enclosure needs a CNC mill?

It requires a few miter cuts and a good chamfer bit.

That's it.

Heck, you could do it with all handtools pretty darn quickly.

(If for some reason, you really want to CNC mill it, and are in the bay area, contact me and i'll mill it on my CNC mill for you)

You could just use regular woodworking tools to get it done.

If you really want to CNC mill something like that then maybe look at the inventables x carve.[0] I got to use a shapeoko for a while which was the previous inventables CNC and it was pretty OK.

I would highly recommend looking in your area for a hackerspace that may have tools like these.

[0] https://www.inventables.com/technologies/x-carve

CNC would be one of the better routes. Like you mentioned, the cost is a barrier. Luckily you can rent CNC time at some hardware stores, university wood shops and some local hacker-spaces. http://100kgarages.com/ can locate you a CNC and answer any questions you may have.

"Get rid of all distractions, to be distracted."

Like the Amazon Dash Button, this is a great reminder of how hardware is still useful in the age of touchscreen interfaces.

This is fun, I like stuff like this; Not too serious and brings that sense of 'doing it yourself' feeling to it.

Damn, this is how you have fun. Both on the tech side and the marketing side. Much respect to Netflix.

This is great and I love it but If the Raspberry Pi had a netflix client this would be amazing.

You can still do the same thing with OSMC and piratebay.

The only, I repeat, ONLY reason why RasPi doesn't "support" netflix is because of DRM. The media files would easily play. So instead of downloading the file each time you want to watch it (netflix), download it our your HD and watch it from there. It's the comparison between O(n*constant) and O(constant).

When you get OSMC up and running, you can add in hooks for node-red. The flowchart for button->actions should be simple to set up.

This makes me think about the swedish startup Flic and thier wireless button.

The hardest path to ultimate laziness, this looks like a fun project!

If you're into IoT stuff like this, check out www.hackaday.io

Maybe Orwell's vision wasn't accurate. The title suggests Huxley got it right.

I'm still not convinced that Brave New World was actually a dystopia, relative to either our modern world or the one that Huxley was writing from. It had some stuff that was pretty fucked up, but was it more fucked up than e.g. the way we treat the poor and homeless? BNW at least dramatically reduced the total volume of human misery in the world.

epsilons are fun, epsilons like to play and run

The WALL-E future is now, just not evenly distributed.


Get more netflix to your life !!

I couldn't even focus on the product after seeing this ridiculousness:


Why would a phone have both capacitive navigation buttons and onscreen navigation buttons? Are they intentionally trying to be funny?

Applications are open for YC Summer 2018

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