How I Helped Potty-Train My Kid Using Twilio and an AWS IoT Button | Hacker News Reader

How I Helped Potty-Train My Kid Using Twilio and an AWS IoT Button

twilio.com

216 pointsgregorymichael63 comments

This is hilarious! But very practical. Buttons are a simple user interface to say one thing. And the phone call on the other side of the system is great. Google's Paper Signals project allow for a more complex input: voice commands. https://papersignals.withgoogle.com/ I can see demand for tons of customized versions of both of these projects. The business model is broken though. Customized hardware or software is expensive. If you could build a cookie-cutter style app creator using emoji-style software blocks (for those with zero programming ability) and stick with off-the-shelf hardware from other creators (Amazon) or even Arduino kits, you could build a business model to support semi-custom products with a mix of hardware and software for many many uses.

Didn't know about Papersignals. Looks awesome (also, for kids). Have you built anything with it yet? What's the developer experience like?

I can't believe I haven't seen this. I consider myself pretty connected, I read HN and have a robust set of reddits and githubs. They should have done more press for this!

This is such a neat project. I want to get some paper signals going up at my house tomorrow. I just so happen to have a jar full of 30 microservos from a project that didn't work out and a couple of the same microcontrollers they are using sitting here... but still, papercraft + IoT. I really like it.

I like the idea of making paper signals for my development team related to sprints, epics, and other task countdown. I think there's something about it that would really make a difference in work.

I would 100% look forward to imaginative kids putting Google's voice recognition through some enhanced QA testing.

I love stuff like this, but I can't help but think you would save a lot of time by just getting an $11 wireless doorbell.

https://www.amazon.com/gp/aw/d/B00M8AQLZS/ref=mp_s_a_1_14?ie...

And likely more reliable. But here's the thing I learned when I got into hacking/makering/hobbying: if you want your widget to be cheaper or more reliable than what's out there, you're going to get discouraged fast.

There's simply no competing with what China can produce. You have to be in it for the joy of the build.

I just looked up the IoT button. $20, non-replaceable battery, not shipping to Canada.

I don't have the skill or resources to assemble it in a nice form factor, but how are there not dozens of cheap Chinese alternatives? An ESP8266 is like $5 and the rest of the parts are probably less than a dollar.

I guess it's just a bit surprising that I cannot find a cheap alternative to buy a dozen of.

The Amazon button probably has higher build quality than anything you can replicate or buy for $5+1. That's probably overkill for most applications aimed at adult users, but it's worth it here.

Is that cheap Chinese alternative toddler-survivable? Will the enclosure break and will delicious shiny metal sharps fall out? Is it reasonable water resistant? Is it coated in cheap lead paint that will come off when (not if) your toddler chews on it it?

Has the Amazon button been validated for any of those things? Ignoring for a moment that we're just assuming product quality of the Amazon product, that's not really my point. For prototyping and hacking, how am I not able to source $5 or $10 wifi buttons with a replaceable battery?

$6 products from China can be really good quality. I mean, after all, Amazon no doubt gets these buttons made for less than $1 in China themselves.

And what's the developer experience?

My approach was to buy a bunch of 433Mhz radio-based "alarm buttons" from Aliexpress. Then I could wire up a simple radio-receiver to an ESP8266 device to handle the presses.

Or of course you could use an SDR-receiver on a PC to do something more complex:

https://blog.steve.fi/creating_an_iot_button__the_smart_way....

https://blog.steve.fi/decoding_433mhz_transmissions_with_sof...

I found this [0] on AliExpress, four buttons in one casing with a rechargeable battery. Haven't ordered it yet though.

[0] https://www.aliexpress.com/item/RAK413AS-UART-serial-port-AT...

You could probably also do this with a small form factor Arduino, a bluetooth module, and a push-to-make button.

With the added bonus that the parts could be re-purposed once potty training is over.

I was thinking a Raspberry Pi W is practically a similar enough form factor for this as well. Some HN'ers might even have one or two of these laying around doing nothing better, so it may work out for them :)

Wondering the same, you probably don't even need a Chinese supplier given how cheap ESP8366 is, you can just assemble a device on a order basis..

Love it, great use of technology and the poop emoji! I probably just forgot but I don't remember our kids getting up in the middle of the night during potty training. Of course we didn't even try until they were 3, you know what they say: potty training a 3 yr old takes a couple days, potty training a 2 yr old takes about a year.

That's not true at all:

"But over the last few decades, the age at which toddlers become diaper-free has been creeping upward. In 1957, 92 percent of children were toilet-trained by the age of 18 months, studies found. Today the figure for 2-year-olds is just 4 percent, according to a large-scale Philadelphia study. Only 60 percent of children have achieved mastery of the toilet by 36 months, the study found, and 2 percent remain untrained at the age of 4 years."

http://www.nytimes.com/1999/01/12/us/two-experts-do-battle-o...

The biggest difference that affects potty-training today is the time available for parents to dedicate to it, it has nothing to do with the children.

I think modern diapers also make a difference. They are so good at absorbing liquid that they barely feel wet at all.

I think toddlers might have been a lot more motivated to learn using the potty when the alternative is sitting in wet cotton cloth.

My son will be three in May. We only just got around to potty training him. He was genuinely surprised to find out that liquid comes out of his penis. The nappy (sorry, I'm British) is so effective at absorbing liquid that he had no idea. A week in, he makes it through the daytime without wetting himself but number twos are catching him out frequently which is fun to deal with.

> potty training a 2 yr old takes about a year

Just so nobody without kids actually believes this:

It's perfectly possible to potty train some 1 year olds, lots of cultures do it, the hard part is communicating with them. Most 2 year olds can communicate verbally and by then it's mostly not a problem anymore.

Then, of course, they're all individuals and vary greatly even among siblings. But the idea that you shouldn't potty train kids before 3 is a strange and possibly damaging idea for new parents.

Interestingly, we found training our 16-month old took only a few weeks. At that age, they're still fascinated with you and what you're doing, so the potty stuff seems to be less tied up in the independence that they're exploring at 2.

To that effect:

I don’t have any kids, but when I do I’m going to write an article entitled: How I helped potty train my two year old with a feather duster and an onion, or some such.

Going through it right now, looking forward to the end of it.

Not having to buy diapers anymore is fucking glorious, but it's also a double-edged sword. No more spending money on Amazon for diapers, but now you have to make sure they pee often, especially before car rides because you could find yourself in an emergency really quickly.

Porta-potty to the rescue

https://www.amazon.co.uk/gp/aw/d/B01JILAFG0

It fits in the rear footwell of the car for zero notice emergencies, has absorbent pads in the disposable bags and doubles up as a toilet seat adapter as well.

I second that. We had twins, bought two. Works great by the side of the highway.

I remember years ago someone on here talking about how companies needed to pony up for 8 devs a year if they wanted 24/7 support - the phrase was "only two people on the planet get to wake me at 2 am without paying".

this is now literally true.

I still rely on on Toddler Lungs 1.0 :-)

> I still rely on on Toddler Lungs 1.0 :-)

As the proud owner of a three year old, I feel you.

My IoT button is stuck to the washing machine. It uses AWS Step Functions and SNS to text me an hour after it's pushed to tell me the laundry is done. Unnecessary? Completely. Useful? Heck yeah.

I'd like you to find a way to get an alert when it's done without having to press the button first.

Use a zwave (or equivalent) plug that can monitor loads and detect when the load is below a certain treshold. I have this for my dishwasher, dryer and washing machine. Works like a charm!

See for instance: https://philhawthorne.com/making-dumb-dishwashers-and-washin...

On the basis that washing machines always break one month after the warranty expires, next time you need to buy one, look for one that beeps and jingles when the program is finished. Mine does this, and it's loud enough to be heard anywhere in the house.

Motion sensor seems like the easy solution there. My washer dryer shake a lot when they're going...

Yup that's what I did in my toy-project:

https://steve.fi/Hardware/washing-machine-alarm/

I tried using a microphone at one point too, but that was not as reliable as just taping a sensor to the back of the machine and alerting after movement stopped.

Huh, the AWS IOT Button doesn't ship to Canada. And it isn't available on Amazon.ca.

Not to be a party pooper... but, huh? Why not?

If author is here, your code for the pip3 install was corrupted by some form of markdown common on blogs. pip3 will not understand —target. That's an em-dash there instead of --.

Remember not to copy commands from web-pages into your terminal:

http://thejh.net/misc/website-terminal-copy-paste

I doubt Twilio would do something malicious, but you're still technically not wrong, of course do you trust your browser plugins not to maliciously change your clipboard contents? ;)

Fixed, thanks for letting me know!

this is funny and a good use of off the shelf apis but it strikes me as a rube Goldberg solution: he's going through AWS to twilio to his cell provider. they're in the same building. he could've hooked up any number of things involving things like RF, Arduino, or whatever. I mean yes you don't get to play with twilio and AWS iot but those are just apis...

Probably because a call is the most reliable way of getting the alert. There are many places in my apartment in which I can't get enough reception to browse a webpage but still can receive and place calls.

I don't think that's what the poster meant.

For example, a much easier solution would be an $11 wireless doorbell

https://www.amazon.com/gp/aw/d/B00M8AQLZS/ref=mp_s_a_1_14?ie...

He changed his original comment. It was more vague hence my reply.

I think the key here is having an IoT button. It's a pretty fool-proof interface, which is an absolute requirement for a toddler, but it does lock you into Amazon's world as your first step - from there, it totally makes sense to jump through the rest of the hoops.

The internet of shit

Remember the 's' in IoT stands for security.

It's kind of cool they sell AWS IoT buttons, guess they liked what one guy was doing with the Amazon Dash devices to keep track of their baby's sleep cycle so he could try to sync naps with theirs. Of course someone else mentioned it's a bit pricey and no ability to replace the battery?... What? If I had to guess the Dash button is likely more reliable and not tied to AWS (you kinda hook it up to a Pi Router and MITM the requests instead - or similar).

I much rather some company would make a nice kit for the Raspberry Pi Zero W to make your own IoT button (with a nice case), at least then you know what you're getting and you feel like you have total control over it.

It's a nice article, but seems like a deliberate cook-up by the twilio PR team. In light of the recently announced twilio-AWS partnership, the article seems even more concocted

True! He probably had the child just because of the AWS kickbacks.

That and your sycophancy here.

I could care less about amazon or twillio. There literally just aren't easily accessible other buttons / SMS tools if you go look. Nothing about this is pandering to them, I just think it's a bit reductive to say this guy made this post as a promo.

Alternate reality version of the post:

Here's my kubernetes SIP/XMPP/WebRTC orchestration to do SMS and an arduino button I designed with STL files so you can 3D print it yourself. Everything is open source to toilet train your kid! I definitely used my time appropriately!

EDIT: After re-review I realize this man could just not have a child at all. It could definitely be contrived and potty training probably doesn't need Amazon's weird button that orders more soap. I guess all I'm saying is that I could see myself trying out the same thing and the solution doesn't feel that contrived to me.

Author here.

I don't want to interrupt a good flame war, but I am just a normal dude who posted a tweet about a hack project and was asked by Twilio to write a post about it after the tweet got popular.

I prefer not to post pictures of my kids publicly out of respect for their privacy, so I won't prove to you that I'm a dad, but some cursory twitter-stalking should provide sufficient evidence that I'm not a marketing shill for either AWS or Twilio.

I dig it. I would try the same thing. Then again I’m a sycophant. :) Thanks for the write up!

I used a cereal bowl filled with mini-M&Ms. 1 for #1, 2 for #2.

> brew install python3

Homebrew now installs Python 3 when you run "brew install python". I'm not sure if I agree with their choice, but it makes your guide a bit simpler.

> I'm not sure if I agree with their choice

Ha! Keep holding out.

Great tip. Thank you. We'll update the post.

Might also be worth changing the setup to use a venv. That way the weird bit about making sure python2 isn't on your path gets handled for you.

Because Python 2 already has an end of life date. You should no longer use Python 2.

One chocolate button for each successful use of the potty during the first week also works. If someone could just build me a Raspberry Pi based dispenser which detects success (and evaluates what kind of success) and then issues the appropriate reward I'll really appreciate that.

After working in DevOps and getting nightly calls from a paging service about down servers, turning my toddler into another down server seems worse than the current situation ;-) That said, this is a great, fun project for an Amazon button.

Brilliant. Though if IoT botnets were bad, I'm terrified of what IoPeePee botnets might be capable of.

why not just use a wireless door bell? cheaper and easier.

Reply on news.ycombinator.com