344 comments

[ 3.5 ms ] story [ 290 ms ] thread
Great article, I totally agree.

One area that I feel is also a "sometimes dumb stuff is better" that the article left out, is that (at least for me) hardware controls are almost always superior (tho certainly less flexible) than software controls. There's no better example than the physical keyboard vs. the soft keyboard. Recently I had to purchase a replacement stove, and I am amazed at how much worse the UX is on this stove than my old one. I don't want to slide my finger to determine heat intensity, and select which burner that applies to by pressing a soft button corresponding to the hotplate. Just give me a physical knob for each hotplate please. The UX there is excellent.

A more painful example is my Ford truck. Almost everything inside of it is software powered, and the bugs drive me insane. At one point my GPS/stereo touchscreen hit some bug, and the only way to "fix" it was to pull over at a rest stop and power cycle my truck (turn off, then back on again :facepalm: ). That is maddening. As a software developer I understand that bugs happen, but as a consumer I just can't tolerate that kind of stuff in my vehicle.

Without question tech has brought us nice things, but with complexity comes bugs (both security-related and non), and with bugs comes software updates (which themselves sometimes introduce bugs in something that was working fine before). Internet-connected things can also be a nightmare for security and privacy. Truly, sometimes the "dumb" version is much better (and way cheaper too).

Car “infotainment” systems are the worst. I drive a VW and the touch screen is maddening. Transition animations are too long and involved, which is frustrating because 99% of the time I just need to force connect Bluetooth and that button combination is muscle memory by now. But every time I have to wait for the stupid interface to do this 3D flip. Oh, and if I skip ahead in Spotify with the on-screen controls, it doesn’t pull the new track data until the next track plays automatically, so it’s always out-of-sync.
The power cycling the car to get the stereo to work isn't limited to Fords unfortunately. I have a "dumb" smart radio and the BT is so bad between my Google phone and the car that I've had numerous glitches some that require power cycling the car, which is insane. I honestly cannot believe something that works flawlessly in nearly every other situation is so screwing in the car, what I wouldn't give to have my aux connector in again.
I work in embedded software, and Bluetooth is the worst. Hardware and software are crawling with bugs. I like to say it's the most unreliable technology in wide use.

What you see is after all the workarounds that some poor application developers had to implement.

That's what happens when you try to design a unifying interface for everything.

Frankly I'm amazed Bluetooth works as well as it does.

WiFi is an example of an almost everything interface that mostly works, though. But it's a different approach because it leaves application interfaces variable, so less big design up front. Its main functionality gap is, of course, low power and very low cost hardware.

That said, I still don't understand why Bluetooth is still flaky after over 20 years of opportunity to improve it.

> WiFi is an example of an almost everything interface

And the gap between almost everything and everything is huge.

I like the phrase I heard a few years ago from friends that played MMOs: “92 is halfway to 100”. Diminishing returns is a painful thing.

> what I wouldn't give to have my aux connector in again.

Your head unit probably has an aux connector in the back, either 1/8" or RCA input as part of video input. You can run a long 1/8" male-male wire from there out to the bottom of the dashboard (what I did on my most recent swap, can't lose the cable this way), or run a 1/8" male-female to an accessory knock-out cover or cut out part of the dashboard. I paid for a higher-quality Kenwood unit, and the Bluetooth, while actually managing to connect (not always the case in my experience, even for factory units), does have problems with interference that sometimes cause sound to drop out for 1-2 seconds. This is as bad as the worst CD-skip-on-hitting-pothole. Bluetooth IMO is complete garbage for car audio.

> Bluetooth IMO is complete garbage for car audio.

Amen. I get a lot of stuttering and random disconnections that drive me insane. Also that's a terrific tip regarding the aux input. Definitely gonna check that out.

The Bluetooth in my car worked reasonable well for a couple of years until one day it stopped working entirely. Thankfully it has an aux connector (and I'll never buy a phone without a 3.5mm jack).
Similar problem in an aeroplane, the Boeing 787 would fail if it had been powered on continuously for 248 days: https://www.theguardian.com/business/2015/may/01/us-aviation...

You don't really think it would be necessary to turn off your plane then turn it back on again, but here we are.

Ha, 248 days is the point at which a 32bit signed int wraps round if counting in milliseconds.
I believe it needs to be a 10ms tick.
And I thought the car analogy was bad. Imagine hearing that excuse at 30,000 ft. (or being the pilot and having to give it) as all the lights in the cabin go off and the actuators stop working (would the turbines stop firing?) "Have you tried restarting it? Maybe boot it in safe mode." Loved the bit at the end 'Operators will perform periodic power cycling at scheduled intervals until incorporating a software update.'

Sad that this story's "thought experiment" https://news.ycombinator.com/item?id=16393740 from 8 days ago may be the future, not a joke. Sorry you couldn't drive to work, we're pushing an update to your car by the middle of the week.

On ground standing still and at 30,000 ft height is probably the safest times for the generators to fail. Plenty of time to fix. Imagine this happening during take-off or landing?
Don't joke. On my first trip flying anywhere ever, on the first leg of the return trip, an AA flight from Austin to St. Louis, we had to abort takeoff while the front wheel was up because all of the on-board computer systems died simultaneously.

I've never been in more of a white-knuckle situation in my life, including the time a tire blew while landing at IAD. We then sat on tarmac for close to 3 more hours while they replaced the computers and we took off again in the same plane.

I've probably told this story before.

I was on a turboprop flight about to depart the airport when we come to a stop and the pilot says that they have a problem with one of the bits of avionics -- I think he said it was for navigation, and that a mechanic was being driven over to swap it out.

This isn't unexpected per se; my knowledge of avionics is that a lot of design goes into making things easy to change out. If you can just swap the chunk that's not working out the plane gets into the air much faster and the electronics can be examined at leisure.

So we're sitting there about fifteen minutes, and the pilot comes on and says "Well, the mechanic says we should try powering down the plane and powering it back up before we exchange it."

They do so. It seems to take a good bit of time, probably five minutes. The pilot comes on and says "Well, that didn't work, but the mechanic wants us to try one more time."

Once again, everything on the plane shuts down, we sit there in the darkness, and then the plane starts back up.

The pilot comes on and says "We're all good now, so we're going to leave the airport."

I felt a bit nervous the whole flight.

(comment deleted)
> I don't want to slide my finger to determine heat intensity, and select which burner that applies to by pressing a soft button corresponding to the hotplate. Just give me a physical knob for each hotplate please. The UX there is excellent.

This drives me nuts as well. I'm the guy going to appliance stores and testing the knobs of every range just to see if they've managed to fuck up even that simple interface.

And some manufacturers have. :-(

You want to be able to turn the knob directly to max from 0, without having to go all the way around, but some have a stop in place so you can only turn it in one direction. Absolutely maddening.

> You want to be able to turn the knob directly to max from 0, without having to go all the way around, but some have a stop in place so you can only turn it in one direction. Absolutely maddening.

At first I was thinking, "Well, when there's a stop in place, that's usually because that plate has one of those turn-briefly-to-11-and-let-the-knob-spring-back-to-10-in-order-to-activate-the-outer-area-of-the-plate" but then I realised how terrible of an interface that is, too! It took me years to figure out, and I have taught it to so many people since, who wish they knew about it earlier...

But what is more important than going from 0 to max, in my opinion, is being able to go in the opposite direction. When I'm holding a pot full of boiling water in one hand and stirring something with my other hand, I want to be able to turn off the plate really quickly.

Doesn't the stop make it easier to go back to 0? Sure you have to turn it more, but you don't have to worry about accidentally going past 0 and leaving the stove on.
That's another advantage of the mechanical knob: a large detent at "zero". Provides both tactile and acoustic feedback. So, no worries of turning past zero.
Exactly, you solve the safety issue with a "groove", so it snaps to 0, and so you get tactile feedback that you did hit 0. Boom, done. No need for annoying stops.
Unless, of course, you view the increase in force required to turb past zero as a sort of soft stop.
This safety feature is far more important than being able to go all the way to max instantly.

But it could definitely be a one-way catch for quick turn-off:

___ _____

0 \_| Max

With the traditional analog controls, the knob is adjusting the spring tension of a thermostat via a screw, so going "below" 0 doesn't really make sense (it would be "more off than off"), and likewise in the other direction too.
The ikea-branded range hood above our stove has touch controls and they drive me nuts. The vapors of the stove deposit on the control surface and cause the “buttons” to malfunction. Sometimes it will switch on full blast by itself and sometimes it will refuse to turn on. The workaround is to wipe down the control surface regularly.

Oh, and every 6 months it crashes and I have to unplug and replug it to make it work again.

You've bought the wrong hood then. IKEA has models with nice clunky physical switches.
Unfortunately physical controls are significantly more expensive to make - we're talking a whole extra dollar or five on the BOM price here, which turns into several times that at retail. Capacitative sensing is cheap and plastic membrane buttons are cheap. They're also easier to waterproof.

Touchscreens in vehicles just seem like an accident waiting to happen.

Finally someone mentions this, everyone seems amazed about Tesla’s big iPad, where you control everything in your car, I haven’t tested it yet, but for me having controls that require taking my eyes from the road sounds like a really bad idea, regular buttons give you touch feedback so you don’t need to look, why is everyone so obsessed with touchscreens?
Because not enough people have read Bret Victor's "A Brief Rant On The Future Of Interaction Design"

http://worrydream.com/ABriefRantOnTheFutureOfInteractionDesi...

(comment deleted)
The Design of Everyday Things already talked about this in the 80s, it's depressing how little his suggestions have been accepted.
I mean, in interaction design is certainly is. It's just that outside of the design world, not enough people realise that there is more to design than graphic design, and that there is more to design than just adding some pretty chrome at the end of the whole development phase.
Yeah, I have been wondering this for a long time already. Can it be that the new people fresh out of school is taking over the design departments everywhere and they grew up with iPhones?
Because they're cheap.

I had a discussion the other day with someone who designs large, moving machinery. His preference is for the actual controls to be physical switches and knobs that are easy to user and learn. He relegates touchscreens for setup & configuration.

My ranges have knobs but my oven is controlled by a seemingly simple set of 4 buttons and a 4 digit disply on a basic circuit board. Just at that vey low level of computerization, the cost to fix a problem goes from $10 to $300.
This is what I don't get about the obsession people have with Tesla, in particular the Model 3. The interface in those cars is inferior and distracting. When I'm driving I need an interface I can operate from muscle memory.

The controls in my 1985 Saab C900 is almost ideal. Everything important at eye height. Direct manipulation analog controls with an immediate and tactile result.

Computers should be here to assist and get out of the way. Wiz-bang fancy too often takes precedence over pragmatism here.

You might be surprised to learn that Tesla has steering-wheel buttons, like other cars.
I'm aware of this, and aware that they can be configured to control things that the touch screen also controls. But some steering wheel buttons are not enough.

Honestly, a giant attention sucking touch screen in the centre of the car is the opposite of the direction manufacturers should be going.

Sometimes the reason manufacturers are going digital is not just to be "smart", albeit it is a nice selling point for them to state it that way. Its actually sometimes cheaper to use digital parts instead of physical parts for certain components.

When you talk about stovetops, the tradeoff between price points on a unit from physical / digital controls is generally if it uses a built in thermostat or not, and the need for more advanced features (built in timers, etc). Its fairly common for kitchens to use purely manual driven knobs that open on/off gas valves on the economy scale since they are reliable and work. For home use stoves, since it also doesn't go through as much abuse compare to a kitchen stove, manufacturers will veer towards digital controls to offer more features and selling points.

Microwaves are the fucking worst. My workplace recently got and LG with hidden capacitative buttons, which don't work half the time, and require various combinations of holding and tapping in unintuitive orders to set anything. A shame because its magnetron is one of those nice inverter-based ones, but I'll be damned if I can figure out how to set the power.

I would pay thousands for a high-powered inverter microwave with two physical dials: power (measured in WATTS, thank you, not percentage), and time.

You can get microwave ovens with knobs, although the power won't be in watts because they don't bother calibrating the power output that closely (and it varies slightly with line voltage and use.)
The only such ones I've found have been under 1000 W, which is too low for me. I appreciate the speed of 1200 W or more models, but those inevitably cheap out on the UI.

Commercial models have a dial and high power, but I haven't found such a one where you can control the power, which is a must for oatmeal and such.

http://www.samsung.com/uk/microwaves/microwave-commercial-cm...

Large and not cheap, but it's 1500W and has 5 power levels.

Thanks for the link. I seem to have cast my net too narrow, as Google turns up this: https://www.webstaurantstore.com/14351/commercial-microwaves... which has quite a few high-powered with dual dial controls (though it's not clear which are inverter-based). Some have instead two columns of buttons, one for power, one for time, which is a design I hadn't considered which does also seem easy to use.
I've found Sharp to have the most intuitive UI, and would probably never buy any other brand. Unfortunately, I'm stuck with a GE microwave built in to my apartment kitchen, but Sharp was what I was used to for many years. You enter a time, you hit "start" or "timer" and it counts down. Other brands seem to have UI designed by Martians - I could write a book on the ways they are screwed up.
> At one point my GPS/stereo touchscreen hit some bug, and the only way to "fix" it was to pull over at a rest stop and power cycle my truck (turn off, then back on again :facepalm: ).

be glad you even have this option.

in my car, sometimes android auto freezes in a way that it will never recover from. this would be a minor annoyance if I could just reboot the thing while driving (and why not? it shouldn't hook into anything essential), but I can't do that. worse still, I can't even hard reset it by power cycling the whole car. the damn thing goes into sleep mode or something and maintains the broken state. only way to fix it is to turn the car off and then wait n minutes for the system to finally shut down.

Especially with vehicles, I wonder if the designers actually use their creations at all, or they do but ignore all the problems and simply replace it with a smug feeling of having designed something so complex.

I have an old car. All electromechanical controls with obvious functions and very pleasant tactile feedback. I don't have to take my eyes off the road to manipulate any of them. IMHO that's how driving should be.

Many years ago (just after I left for college) my parents decided to build a brand new house. The plan was to include a computerized lighting system where every light was individually addressable and all "light switches" were actually banks of 4 to 8 programmable buttons. The possibilities were endless! What could go wrong?

I asked my parents to use a regular light switch in my room. You know the kind that I am talking about - it goes on and off, closing and opening an electrical circuit. It would have been less expensive and (in my opinion) more reliable since after all, it's just a few pieces of metal. I joked that "when everyone else's lights are flashing on and off at 3AM, I'd like to remain asleep". My father accused me (his resident technologist) of being a Luddite and they went on their merry way, finishing construction and installing a fully-computerized lighting system.

A few months after they moved in, I got a phone call from my father that started something like this: "Please don't say 'I told you so' but I have to tell you what happened at 3AM last night..."

(they flipped that house, and when they built their next one, it had regular light switches)

I totally get the spirit of what you're saying, but to be fair a good reliable automated light-system would have an analog fail-safe that allows you to operate the lights normally in case of a malfunction of the central system, meaning that you just turn off the central system, and all the lights keep operating as normal lights that are not connected to it.

But also, centralized light systems are not new. Many corporate facilities have been using these for years and they rarely fail.

Maybe the problem is that your father just picked the wrong vendor/system.

I think the problem is that you don’t really need centralized lighting at home, and the solution is to use regular light switches. ;)
That more or less describes how Philip's Hue works.

If the hub died, the smart bulbs can still still be operated using the regular light switch. If the bulb is in standby (smart "off") just flip the switch twice to turn it on. Then it will act like a regular bulb until the hub returns.

Hue only has one big defect with regards to this... If there's an electrical outage and the smart bulb is in standby, when power is restored the bulb will turn on. Meaning if power goes out and is restored in the middle of the night the lights could suddenly turn on full brightness.

And as an aside, the nice thing about smartbulbs as a concept is that they require zero re-wiring or specialized wiring. So if the manufacturer of your bulbs went out of business or you wanted to change vendors, just replace the bulb and your work is done. Simple.

I've seen televisions have that issue. They can come back on in the middle of the night. To me, this makes sense for a lightbulb. If your failsafe "backdoor" expects it to be wired to a switch, you would turn that switch off in the event of a power outage. What else would it do when the power comes back on?

One thing I really like about "smart" things, when the power is out at home I can check my phone to see if power is restored. I guess that's similar to calling your home to see if the answering machine is back on.

I have a couple dozen Hue lamps, and they regularly “reset” to full bright white, presumably after a flicker in power that is otherwise imperceptible. It happens on average once every couple of weeks, and is quite annoying! I wish the lamps had a nonvolatile memory so they’d return to their previous color and intensity...

(I still love them and am happy I have them rather than regular LEDs though.)

This has been suggested many times and, as far as I can tell, Philips decided the current behaviour is best safety-wise (consider e.g. if your last preset had 1% intensity).
Then reset it after three on/off cycles within 3 s, or something.

But surely the controller iwll sort this out when it gets back online a minute later?

So that everything comes on (annoying enough) but switched off within the minute?

If not that is just outrageously bad and imo ruins the product completely.

Three on-off cycles in rapid succession is exactly what you get in non-urban electrical systems when a tree falls against a wire. The local breaker detects the short to ground and cuts out, then resets after a few seconds on the chance that the tree has continued it's fall to the ground. If there's still a short, it repeats, and after the third or fourth short->cutout (depending on the local utility settings), stays off on the assumption that the tree is well and truly hung up on the line and will need a crew to fix it.

It's pretty common occurrence in many locales where I've lived.

But it's the least smart thing to do. They could at least put an option in the setting to maintain settings after brief power outages.

We had a power flicker last week in the middle of the night and suddenly our bedroom lights were on 100% and our Echo was loudly complaining that the internet was gone. At 3am.

Which is why smart things are dumb. You can't just program features, you have to program context. And context rapidly becomes non-trivial. So you either have a dumb object that no one expects to act intelligently, or you have a smart object that tries to act intelligently but fails because it's not as smart as a human, and humans are random anyway.

For smart lights you need "Return to previous state, and then see if the state was on a timer. If it was, wait until you get an NTP answer (i.e. the Internet isn't down) and set up the state requested by the timer."

Which is fine unless the previous state was a party, it's now daylight, and all the bulbs are still different colours.

Or someone was adjusting the lights and accidentally set them to 1% instead of 100% when the power went off.

Everything becomes an edge case. The most useful solution requires an intelligent default with timer support, with some awareness of what's happening in and around the house.

And it still has to allow for user overrides, because some people - like my partner - sometimes want the lights on during the day, even when the room is bright.

I think you are over-thinking things.

I also have analog switches and dimmers in my house. If I have the lights at the lowest setting when the power goes out, that's where they are when the power returns. If everything was on full when the power goes out, then that's where things will be when the power returns.

Why should smart things work differently than what a lifetime of experience has lead me to expect?

To me, this is the right thing to do. Safety aside, it "does the expected thing" if I flick the switch off and back on again.

And lights coming back on at full intensity after a power outage? Pretty common in analog houses as well. Who thinks to turn off a light that isn't on?

In analog houses, restoring the power means "lights come back as they were prior to outage". No state is lost, no matter how you spin that.
Agreed. I have been woken up from a power outage this way. We should be able to specify the default behavior when the power to the bulb is turned off. Maybe if you flip it on and off quickly within 1 second it will turn off. Or if the power is turned off to the bulb and the bridge is still on, the. Assume it was turned off. It’s not hard to come up with an OPTION for people to choose. I hate that there’s no choice, but I love my Hues.
I suppose they aren't designed for places with unreliable mains power. The auto-on behavior is just a good first approximation of the behavior most people want.
The system works- but it fails due to the surrounding system. Perfection is archieved!

And the lawyers deemed it suitable! Box and ship it.

If you are going to deploy "things" that don't work the way that a traditional device does then you might want to completely mitigate the behaviour or put up with it. A UPS or multiple UPSs are indicated here.

You could go the whole hog and install enough UPS for the whole house and a backup generator.

Yes let's install a whole house UPS so the light bulbs work right...
> Hue only has one big defect with regards to this... If there's an electrical outage and the smart bulb is in standby, when power is restored the bulb will turn on. Meaning if power goes out and is restored in the middle of the night the lights could suddenly turn on full brightness

I guess the reason you have to flip the switch twice, is to prevent this problem. Then, if you get an outage during the night, it should continue to be off...

A power outage looks exactly like flipping the switch twice.
A friend of mine has Hue lights in his bathroom. The fact that they operate like normal lights when the power is interrupted is the only reason that guests have any chance of succeeding in using the bathroom without instruction.
He doesn't have a motion sensor? He could also potentially install a humidity sensor for cases like showers.
See, now it starts to become weird again. I just want the lights on in the bathroom, a motion sensor is an unneeded expense.
Yeah, and robots with laser eyes scanning for people sitting on the crapper...
See, now then you have to provide instructions to guests to wave their arms around vigorously when the lights go off mid shave!

(Just like we have to do every 30 mins in the office when there's 3 or fewer in.)

yeah, waving your arms around wildly when the lights suddenly go out is practically instinctual for many people now :-)
This is so complicated. Especially for a bathroom where there is likely only one entrance, you're no more than 5 steps from the such at any one time and you only need to operate the lights when you enter/leave.
I don't think they want to mount anything on the walls, since it's a rental, but that does sound like a good idea. Smart bulbs seem kinda useless to me, but a three-way light switch with on/auto/off might actually be really nice.
There are a few comments under here about lights that default to "ON" after a power loss. I'm not clear why manufacturers didn't spring for a few bytes of Flash memory on the device to store last state. If my BIOS can maintain the ON/OFF state of my computer after a power outage, I'm not clear why my light bulb can't.
The default setting after a power cut can be configured for Osram Lightify bulbs.
You could potentially pair it with Home Assistant to restore the state of the lightbulbs to 'off', and ensure Home Assistant has backup power in case of an outage (it takes a while to boot on an RPi3 at least).
>Meaning if power goes out and is restored in the middle of the night the lights could suddenly turn on full brightness.

That's the same with regular bulbs, so...

Dimmable, color adjustable bulbs (assuming they exist)? Oh. Didn't think so either: an analog dimmer will resume operation right where you left it, no spontaneous resets to 100%.
No, regular bulbs without a dimmer. Like 99% of households use.
Indeed. Then it's restored to previous state (which could be "on" or "off"), not reset to "always on." Even my computer's BIOS could do that ("on power restored, turn on if that was the previous state, otherwise keep off") in 2000, is the smart lightbulb dumber than that? That's a very unfortunate level: smart just enough to be annoying, but not enough to be useful.
I've never had regular bulbs that were off turn on after a power outage.
While I love my hue bulbs, I have two main gripes (beyond cost) which have kept me from replacing all the bulbs in my house:

1. As good as they have gotten, they still aren’t nearly as instantaneous as a good old fashioned switch. Yes, there are ways to mitigate this, from having it decide when to turn on to alternative controls. But that latency really gets to me sometimes.

2. The system not recalling last state. If you are worried about not being able to default to lights being on, you can make it so that a quick toggle resets it. But I can be convinced of other options/opinions here.

In general, I think the import point is how smart/dumb an object it is. It is more important how “rock solid” it is, and simple to use. I love the mantra “don’t make me think”. Devices should become “invisible” for the task you care about.

> Meaning if power goes out and is restored in the middle of the night the lights could suddenly turn on full brightness.

I'm sorry but this is unacceptable. It's what keeps me from adopting Hue in my own home.

My lighting system literally has one job, if an advancement of this technology can't keep analog reliability, I don't want it (That's a bug, not a feature).

Don't get me wrong, I love the idea of Hue, color changing bulbs, ability to remote control, ability to dim, etc. is awesome, but maybe should stay in a switch format if loss of power can't be handled correctly.

It would be fairly natural to have a small three-way switch onto each light to select between "central control", "override on" and "manual override off", in that order.

That way, you wouldn't have to shut down the entire system.

It would, but it would also be expensive.

Adding a switch changes the assembly process, requires extra firmware and testing, makes manufacture more complicated, demands localised labelling (unless you can work out a symbol system which maybe 50% of users won't get), and requires changes to the manual.

No doubt. This is why I think we need some sort of mandatory class-based reliability certification for a society of increasing IoT reliance. Not mandatory as in "you need to certify as at least class 3" but rather "you need to certify as some class of your choosing and clearly mark this on your packaging." The certification process should probably be cheap and flexible too. Maybe even let manufacturers self-certify the lower classes with heavy fines for abuse.

That way, it would be more clear to the customer why the product with the physical stops and manual overrides is more expensive.

Our office has something like this: dimmable meeting room lights that are controlled by a single rocker switch.

Since the only UI is "hold the switch down for longer or shorter periods, or press it repeatedly", and there are no written instructions and no discoverability, nobody can reliably work it and we routinely hold meetings in the dim half-light.

Same here, except sometimes lights will brighten and dim randomly during meetings.

It's like living in Clamp Tower from Gremlins II.

Well that's your own damn fault for not RTFM! (Had that system at an office: it came with a manual of ~200 pages, but nobody wanted to learn the special variant of Morse code that the lights understood. Somehow we found the commands for "full on" and "full off", and the manual was abandoned. Seriously, 200 pages on operating a light? Technology making lives easier, riiiight there.)
I once stayed at a nice hotel in Bangalore with a computerized lighting system controlled by a LCD with no switches. I just kept the lights off the whole time, I couldn’t figure it out and it was so cumbersome to use anyways. Light switches are an essential luxury!
The way things work in Bangalore, the best thing to do is to call the front desk every time you want to turn the light on or off.
Heh, once upon a time (and this was back in 1996 or so) my parents built a house too. They opted to put an expensive "Hunter" branded ceiling fan in every room. There were no controls on the fan itself, but on the wall where a normal light switch would be, there was a plate with two small buttons. One was for the light on the fan, and one was for the fan.

Each button could be in 4 states: hi, med, low, and off. To get from one state to the next, you had to hold the button for about a second or so. Simply holding the button an less of a time would automatically put the state in 'off' (which would drive you mad in it self). My parents were cheap and rarely used the A/C (in Texas!). I would normally leave the fan setting on medium over night. But that normally meant I would be awoken at 3am to the lights on the fan being fully on....

So they sold their problem to another family without telling them about the problem? Tell your father from my part that he is a moron. Not you. You seem to have common sense. But your father should be less entrepreneurial and start doing things properly.
You have no clue how that deal went down, and you had to jump to your own convenient conclusions to arrive at your judgement.
If you don't have any clue, how are you able to judge my judgement?

He said his father flipped the house. He didn't said her father sold the house. Flipping a house involves the aim of making profit. So I assumed his father said nothing to the buyer. But he can answer me about what his father did if he wants. On the other hand, I can't see the point of your comment.

I agree with your sentiment that not telling the full truth and not giving one owns full assessment is easily equivalent to lying, but the way you brought up that point is not okay.

We don't know how the deal went down, so we shouldn't accuse them of wrondoing. The wording of "flipping the house" is not enough evidence. In the face of that, any accusations can be rejected by everybody, and many rejected your accusation by downvoting.

I definitely wouldn't use any kind of technology that doesn't gracefully fall back to working as an uncontrolled "dumb" device. Z-Wave switches are at least pretty good at this, disconnecting my hub turns them back into normal light switches.
My pet peeve is an even lower level of technology - having multiple light switches for a single light, for "convenience" means that you sometimes click "up" to turn off a light, and sometimes "down" depending on which switch was used last. The logical state and the physical state didn't sync. I found that incredibly annoying and impossible to memorize.
We don't realize how complex a lot of the simple things are.
This is true, I’m not sure the article is saying that the “dumb” versions aren’t complex to make/manufacture/use. Just that sometimes they are more reliable to use in practice.
"One of the most common uses of Amazon’s Echo is to set a kitchen timer."

Truly worth the money. I suspect the greatest value of these things is that they seem like nice gifts. The same thing seems true of the cheaper drones.

I think the greatest value is probably in the speaker itself, most of the other features do just feel like fluff. Though personally I prefer a more “dumb” speaker I can control with my phone.
Same for spellcheck, learn to spell and just turn it off. I also highly doubt I will ever pay money for a driverless car.
Even if you know "how to spell" it is easy to not notice when you've misspelled something. Also, there isn't really such a thing as just "knowing how to spell", it's an open-ended skill. And some people just don't have the mind to remember the precise spelling of every single random word.

I also don't know what driverless cars have to do with this. For one thing, it is possible that you might not have the option in another 25 years in some locations. For another, this actually is an extremely Luddite attitude, given the objectively horrify safety of cars which we've managed to psychologically lock away in modern culture. Not to mention the massive secondary benefits: you get to recover massive amounts of time dedicated to staring at a road, and being sober.

>given the objectively horrify safety of cars which we've managed to psychologically lock away in modern culture.

Ab out the car safety, I see your point mentioned every time selfdriving cars are mentioned but we could remove the bad drivers from the road today without having to wait 20 years or more until all or most cars are self driven. My solution would be to have mandatory systems to check the driver for alcohol or drugs and a system to check for the driver attention plus the collision detections and other driver assisting technology. I know there is a privacy problem but I think is not larger then the selfdriving cars one or we can adopt this technologies in countries with less bad governments.

My point we can improve car safety today if we want and not wait for selfdriving cars.

>My solution would be to have mandatory systems to check the driver for alcohol or drugs

I mean, I don't even know where to begin.

First, good luck getting the 200 million drivers that don't ever have this problem to agree to dealing with this system every single time they start their car. Not gonna happen.

Second, have you actually stopped to think about how this would actually work? BAC is already a pretty vague and inconsistent check of impairment, but it is literally the best available from any category. There are no sobriety tests for other substances. And there definitely are not ones that could be compacted and used on-demand in vehicles.

> a system to check for the driver attention plus the collision detections and other driver assisting technology.

So instead of just letting software do everything, instead you think we should stop at a stage where the human is redundant and honestly a weakness in the system, just so that we can feel like we are doing something when driving? Why not just let the software do the last 10% of the work and free up 100% of your driving time? This just sounds like pure stubbornness, not an actually effective system.

And finally, it completely negates the enormous advantages of an automated driving network. Barring rare and extreme events like multi-lane closures, traffic exists entirely and exclusively because of fundamental issues in human habits and judgement. There would be no traffic if computers were automating freeways. Ever.

You did not understand my point,

1 I am ok with good self driving cars that are better then the average human

2 I am suggesting that forcing this systems on people is as problematic as forcing them to buy an AI car, so say from tomorrow in your city you have 2 choices : buy and install the system I suggested or buy a new car with self driving AI, why is second option easier to force on people?

3 My point is that if you want that in 30 years to force all cars in some regions to be 100% AI driven so forbid people the right to drive there, then why not also force that device I suggested today, you get the benefits today and in 30 years you can force no humans driver at all.

4 My problem is that the problem with car crashes has more then 1 solution, so if we want to solve it today we need to look at something else not only at self driving cars,

I mean I see what you are saying but in the case of spell check and driverless cars you aren’t replacing the dumb version with the smart version you are augmenting it. If your spell checker fails you can still manually do it without any friction.

I think some people get wayyy to reliant on using spell check though (especially with the right click to replace etc, so they never have to actually learn to spell the word).

(comment deleted)
One of such example is when they offered us a "Smart Cup" on our Cruise (Royal Caribbeans). It has built-in chip so it know who it belongs to.

So you go to drink fountain, puts your cup in, LCD display nicely says your name (wow) and then you click which drink you wants and the cup fills in, while the display tells you from now on it remembers your favorite drink. Can't be better huh?

Five minutes later I go for a refil: "Sorry but you drank a cup less than 45 minutes ago. Please come back in 39 minutes".

I came back, but to the front desk to get a refund. Instead, I replaced it with more expensive regular cup that you could go to cafeteria and get refils of pretty much anything you want to as often as you wanted to.

What was the machine dispensing, Woodford Reserve? I thought cruises were “all you can eat/drink”? And why would they sell a limited smart version along side a dumb, analogue version that you could presumably refill as much as you wanted?
This highlights one problem with smart devices. This device truly was smart, but it wasn't smart for you. It was smart for the cruise line, at your expense.
Royal Caribbean has become masterful at the nickel-and-diming of passengers. My wife and I used to cruise with them, but no more, precisely because of this.
Drink fountains used to literally be a tap that let you fill as much or as little as you want. Now they have these "smart" ones that give out premetered amounts, regardless of whether you just wanted a mouthful or have a bigger cup to fill.
I use the tablet to show recipes.. but it causes grief that its blackscreen at just the point I dont want to have to touch a device to wake it up.

The nice thing about paper is, you don't have to wake it up to look at that ingredients list.

Because my iPad locks when the Smart Cover closes, I always just leave the auto lock timer disabled to avoid this situation.
Some recipe apps (such as Paprika) prevent display sleep. I have found that more convenient than some books that always want to close.
Decent cookbooks have put thought into the binding so this doesn't happen. I'm not sure there's a clear winner here.
Since I am a vegetarian, the range of cookbooks is a bit more limited. I have some nice vegetarian cookbooks with binding that closes the book.
...and what no one giving us this "progress" ever seems to realise is how often it is slower.

For instance our Whirlpool microwave - touch sensitive digital and achieving 5 or 10 minutes takes much longer and more clicks than our old one (no 5 min button, but the Panasonic's touch keypad didn't much like damp hands). Setting power or grill is especially unfriendly and rarely succeeds at first attempt. First microwave had a rotary control, now the preserve of absolute bottom of range rubbish, time setting always took one touch and under a second.

I find it's affecting my choices of cars though. Touch sensitive "improvement" of anything basic (heater or fan for instance) is a big negative mark for me as chances are I could formerly achieve it without taking eyes from road, but now I need to look, and you don't want everything on the steering wheel!

> First microwave had a rotary control

Reminds me of the old "ha-ha-only-serious" joke about how a computer scientist would design a toaster:

http://www.fiction.net/tidbits/computer/computer_cs_toaster....

Unfortunately some people interpreted it as a plan for the future.

Most toasters have a thermostat or photocell instead of a timer, so even the simple design is an example of complicators gloves.
A simple timer is under complicated - if you cook two slices in immediate successionm, the second slice will be burned as the coils will still be hot. As I understand it, a good quality toaster will have at minimum both a thermal switch (like a bimetallic switch) and a timer, which only starts when the toaster is hot.

I don't see how a simple thermostat with no timer can do the job by itself.

About the microwave: it seems to be purely American thing. Here in the UK even high-end very expensive microwaves have rotating knobs for setting the time and power. I've seen those monstrosities with a million buttons on the front but they are certainly rare.
It seems to go in cycles... Ten years ago I bought a microwave with a digital knob, and it was fucking fantastic, because if you just turned the knob it increased the time, but it had acceleration built in, so it went in small increments at the start, and larger later. But it was completely deterministic, so you developed muscle-memory for "30 seconds" or "2.5 minutes", and it was so fast to use.

Five years later we moved and I had to buy a replacement, and I found one model that had a digital knob, all the rest had the idiotic keypad instead.

Not just an American thing... At home I got a bare bones one with 2 mechanical knobs (even the "ready" sound is produced mechanically I think). But right now I'm standing in front of a sophisticated one in a Taiwanese home, it has 18 buttons marked with Chinese characters and various numbers or ranges of numbers that make no sense without a reference manual. No symbol, no knob, no useful color coding. I'm unable to operate it without help... Makes you feel like a child needing help from an adult :)
Likewise when shopping for a new hi-fi receiver, the most important feature I wanted was a physical volume control knob.
Grr. The old Denon bookshelf hi-fi had an analogue pot with slow motor tied to remote control, and infinite fine adjustment when using knob directly. After it died we just bought the current equivalent without exercising enough distrust.

The volume is now digital, so we have a choice between "too loud" or "too quiet". Turning the knob gives the same too coarse steps. No motor is required, saving Denon a few pennies and costing them a customer next time around.

The FM tuner (much better quality than DAB hereabouts), uses the digital stage somehow anyway, so analogue transmissions gain about a half second delay. As we often have the same radio station going in more than one room, this can be really annoying.

I had an August smart lock for a while. I used it primarily while running so I didn’t have to carry keys.

The whole process of unlocking the door was way more painful than a key. I switched back to a regular lock.

I find this [type of] lock quite usable, and much more convenient thne keys:

https://www.amazon.com/Schlage-FE575-CAM-619-ACC/dp/B001RFDB...

Digital so I don't need keys, but still keeping it as simple as possible. Most have a power backup so you can power it from a 9V battery externally. I still wish it wasn't necessary to change batteries, but it's not too inconvenient given they last a couple years, timewise well worth not dealing with keys.

We have two of those, as good as they are I wouldn't trust them on an external door[0]. The locking mechanism has to be fairly light weight due to being driven by an actuator and a battery operated one at that, which means instead of a deadbolt-like configuration it is a traditional spring Latch, which could make it susceptible to bumping or even a strong rare earth magnet. Plus the key is pretty simple.

[0] https://www.youtube.com/watch?v=X038GMKwHNw

I've always preferred tactile/mechanical controls (buttons, knobs, sliders) to the touch screens we have today. And I always trust mechanical systems over fly-by-wire systems. For example, I'm very uncomfortable with things like the Tesla's slide-out door handles, electronic locks in hotel room doors, touch buttons in elevators and GPS navigation systems in aircraft.

Smart tends to mean complex, and complex usually means more things to fail. Some things only need shafts, pulleys and gearwheels -- there is no need to bring a microprocessor into the picture...

Bad news: GPS in airplanes is the least complicated piece of kit, and most commercial aviation is literally fly-by-wire.
I can't find the plane/article but I remember reading about a military jet that is practically always in an unstable flight pattern and computers constantly adjust it to make it so the pilots can fly it normally.
The would be the F16 if memory serves me well.
I recall reading that about the F-117, many years ago. Can't remember the source, either, and I've no idea if it's accurate.
I recall the main computer of the F-22 core dumping when it crossed the international dateline. And an earlier version of the F-16 flight control computer flipping the plane upside down when it crossed the equator in a simulation...
I think that description fits anything from the B2 bomber or newer made by the US. Stealth implies on some heavy trade-offs.
Being inherently unstable allows it to be more maneuverable.
I'm pretty sure almost all fighter jets are like that - the more unstable it is the more responsive it is when turning, since responsiveness and stability are kind of opposites when it comes to aerodynamics.
My favourite "smart" dumb device has got to be single sensor electronics. The automatic hallway light to be exact. It's a piece of plastic with a photocell and a LED.

So long as it has power (but a battery backup would be simple enough) it never fails to turn on "if this, then that". Daylight savings, changing seasons, solar eclipse.

It's never a device I have to think about.

And at home depot they sell $80 IoT light bulbs demanding a property app.

I think real ubiquitous computing will require a different technological approach. These Wi-Fi enabled devices that you can access with an app on your phone are clearly not working out too well. I look forward to tech like smart contact lenses, that google jacquard project, and other interfaces... what if instead of telling Alexa to turn on the lights, you could turn your lights on and off by gesturing (think like mouse gestures--you move your wrist down and left in the air, and a bracelet or sleeve knows what you gestured). Or even better, just thinking.

So we're a few hardware inventions away (not to mention, truly enjoyable product inventions away) from knowing what 'Smart Things' really are...

IoT is built on a cruddy stack. Nobody in their sane mind would design it like that from scratch. The problem is technology is much more pervasive compared with earlier eras, and without careful future proofing you end up stuck in a lot of unfortunate designs that cannot be gotten out of, and more and more is thrown into the pit making it worse and worse.
This is why I have been getting superior results working with a $100 tascam digital 4 track compared to what I used to get from logic pro. For music being able to use knobs with eyes closed to record is actually huge to me the GUI is actually a problem for music for my work.
There are few worse design decisions than rotary encoders on a software synth
In almost all music software these days, the knobs are actually linear faders in disguise. Displaying them as knobs simply have the advantage of taking up less screen space.
There's a reason why most music today is still mixed on an analog console. Notice that mixing and recording engineers, even if they use Logic to record, the computer is somewhere on the side, it's not in their face. They don't look at the screen, they just keep their hands on the analog faders and listen.
Sort of related but not exactly.. and I'm sorry if this is controversial... I just prefer my gas car over my electric car. My gas car is older, the cd play doesn't work however the radio does. I could replace it. I just don't care enough and prefer the randomness of the radio. My electric has mp3, satellite radio etc. Don't know if its nostalgia or what I've just grew up using but yea.. I hate my electric car.
I like elemental cars without bloated features as well, but most new cars both ICE and electric have screens and the like. I don't think this is unique to EVs.
Fair. But that's not totally why I dislike my electric car (Nissan). I mention the modern stuff just to highlight the car has significantly better entertainment systems. The nissan compared to the gas (a porsche) is an unfair fight. However I haven't been impressed with tesla when test driving them as well. It's the drive, idk. I know tesla is trying to solve that but so far I haven't been impressed.
I have a feeling simple vechicles (one computer, and maybe a powertrain computer) will be coveted in the near future.

I have a nineties dodge Dakota that I was going to just run into the ground. I never loved the truck, and used to think it was complicated to work on.

After looking at new vechicles, the nineties cars, and trucks are comparatively simple.

I grew up fixing my own cars, and couldn't imagine paying someone to work on my vechicles.

I don't like the complexity of the newer vechicles I'm looking at.

The wealthy will have no problem having a dealership spend 3-hours calibrating the sensors on a bumper, at $240/hr.

I couldn't imagine spending so much. Even if I had the money spending $2400 to replace a bumper is just crazy.

(A buddy of mine called me up, and let the price of his bumper replacement slip out. I told him they must have done more. Well, I looked up his make/model, and that is roughly the going rate. And yes, it went right on his credit card.)

I've never found a shuffle mechanic that works as well as the radio station. You know, where the songs that are new and currently stuck in your head play like... 70% of the time, and then the other 30% it's picking older stuff you love? Most music player shuffle options are pretty basic, and I definitely jump over to the radio at times, despite having my entire music library stored in my car.
> I've never found a shuffle mechanic that works as well as the radio station.

I can't stand ads anymore.

The funny thing is, I now have enough ad free content that I start to miss references because I'm not getting as many ads. Which is to say, I don't really mind some ads here and there, and I can always switch stations during long ad blocks.
On the one hand, the analog watch they show to compare with the Apple Watch is an Omega Speedmaster, the first watch on the moon. Shows that you can do great things with just gears and springs.

On the other hand, it does cost $3,000...

On the third hand, it'll still be running after 50 years and fully compatible with all standard issue eyeballs. Amortized cost of $60/year. Will the Apple Watch still be running in 50 years? At around $300/watch right now, each one needs to last at least 5 years to remain price competitive.
5 years sounds alright, actually. Tech can last that long, despite engineering for obsolescence. But that moonwatch will probably stay in style much longer too.
In 50 years, will it still fit my sense of style? Will I manage to avoid misplacing it or getting lost for 50 years?

I'm not saying I want a Smart Watch (getting even more distracted? No thanks). What I'm saying is that lasting power in a device isn't just about continuing to function.

Fair, but that particular style has apparently already made it 50 years. It's got lasting power at least with a decent sized (or well funded) segment of the market.

It's also an object you can give to children and grandchildren. I have a number of objects like these. I would have zero interest in passing on my 2017 Apple Watch 3 in 20 years.

There's something to be said, IMO, for things that will last generations and not just years.

I got a watch from my grandfather. It was completely wasted on me.
I would bet that Omega Speedmaster, Rolex Bubbleback, an IWC ingenier, or older Rolex submariner will still be in style 50 years from now.

My favorite watches are all made before I was born, and I'm getting up there. Certain watches are timeless.

Sure, but 2,970 of those dollars are just for fanciness.

My $30 digital Timex is fast, can be worn to exercise and then in the shower, doesn't need charging or winding, and is (IMO) easier to read than analog.

It just doesn't look fancy.

I have done some neat hobby automation projects, and I have some thoughts:

1. With some more work it wouldn't be so, but when the power goes out it is like rebooting Jurassic Park to get my sound system back the way I like.

2. The proprietary power grab is worse than any video format war or chat program territory battle. I've tried to DIY as much as I can, but Google Home is a no go for DIY. I can emulate on/off Phillips Hue stuff for Alexa Scenes but for Google Home you have to go through the net.

3. I can't ask Alexa to do anything when the (rare) internet is out, but I can control everything fine with my phone or IR. Most automation projects seem to require outside the firewall stuff.

4. Would be cool if I could come up with a port knocking thing for Alexa Skills to talk back to the LAN with some kind of reliable security and encryption... But Skills seem to be only available to call out insecurely or to a running EC2 instance say but I don't have a free tier account being an early adopter of EC2.

5. In general, the products I will buy are ideally LAN only. Maybe that's unrealistic like a non smart TV?

Yes I would like a Starship Enterprise-style computer in my house. Who wouldn't? But it has to be the the ad-free version that works for me. It won't monitor the activities of my family and then sell the anonymized data to selected third parties. Then supply the de-anonymized data to government agencies and school boards at the drop of a hat for 'safe-guarding' or 'anti-terrorism' purposes.

In the meantime I recognise that there's some health value to getting up periodically to flick switches.

No one's going to build this for you. If you want an assistant built around your needs... start building. AI is not going to magically make a piece of software work the way each and every person on Earth would like, but making one for yourself is far more possible. :)
Then there's the ST:TNG episode "Contagion" where the crew spends the entire hour futzing around with the Enterprise's computer only to come to the conclusion of "let's reboot it!"

Sigh.

Was it running a Windows update ;-) : "do not power off your computer" = "tough tits if you were packing up your laptop to go home matey, one of those soi-disant 'updates' is actually a virus scan, which will take hours if you're lucky, or I might just bork the whole frickin' thing, we'll see..."
I think this is kinda possible - if you hook up a raspberry pi with a bunch of relay switches. I think there's a speech-recognition project from Mozilla that's actually trained on the pi.

I vaguely considered it - but honestly, what's the point? Light switches are excellent tech. I think I'll go for a home computer the day somebody makes a roomba that can pick up trash, and clean the bathroom. Before that, all the really time-consuming stuff has to be done manually anyway.

For speech recognition, in android, you can use google now, then with the app tasker you can send the recognised text to your mqtt broker and from there it's easy.

I'm doing that with some relays and LIFX bulbs using Openhab.

This is an interesting voice assistance project that does not rely on calling out to third-party services:

https://snips.ai/

That's exactly what I was thinking about, actually!

I would absolutely never install a non open-source monitoring device in my house - and I frankly think it's insane that people pay money to do so. It's really nice to know that when this technology matures, and becomes the kind of thing everybody has, I'm not going to have to be a luddite.

> Yes I would like a Starship Enterprise-style computer in my house. Who wouldn't?

I wouldn't. What's the point exactly? Take smart lights, what are they for?? When I enter the room I switch lights on, when I leave the room I switch lights off.

I'm not sure when I would need to switch lights on while already in the room. I never stay in the same room for so long that the sun sets while I'm still there not moving.

At night in the bedroom there are switches near the bed (attached to the wall, which can't get lost, contrary to a remote, and don't need sound, contrary to voice control).

My point is, we tend to accept technology even when it doesn't solve any problem we are actually having.

> When I enter the room I switch lights on, when I leave the room I switch lights off.

There have been more than a few times in the last month where my hands were disgusting enough (pets, kids, etc.) that touching light switches, knobs, etc. meant I had to retrace my steps with cleaning supplies.

You still have elbows, a nose and a chin to switch the light on/off.

And how do you use a remote with greasy hands? You have to clean that afterwards.

I use them when I can, but motion sensors in public bathrooms don't have this issue. Well designed public bathrooms don't require you to touch anything in the room. A similar setup at home would be beneficial to households prone to biohazards, which most are in different seasons of life.
I can understand that for a public restroom the motion sensors are a good addition. For a home setup, what do you do when you sit in the couch, or at your desk. Do you wave to the sensor every five minutes, just to leave the light on? I cannot imagine I would ever want that.
Well people can do what they like, but I'd enjoy the public restroom setup in my bathroom, especially if it is a half bathroom (toilet and sink but no shower or tub). It probably makes sense in kitchens, garages, and utility rooms. Living rooms and bedrooms shouldn't run on motion sensors, at least not always.
Both flat and regular light switches allow for the use of elbows or, in a pinch, noses.
What if it were inexpensive, kept the physical controls and actually worked well?

I'm not in any rush to automate lights, but there are lots of people that would appreciate things like an "arrival mode" for their lighting, where lights switched on when they arrived at the house. Not really a big deal, but say you have poor night vision and reduced mobility.

Agree for reduced mobility, but re:poor night vision, I can do most things in my home in pitch darkness, as well as open the door from the outside: selecting the proper key from the key ring in my pocket, "feeling" the keyhole and inserting the key without looking at anything. Whenever I change the key cylinder I verify I can still do it with the new key.

This used to fascinate my wife. Now it just upsets her, she thinks I'm showing off to the kids. I'm not; I'm training for when the sun dies.

I can do the same. Although the 'open' or 'lock up' door in total darkness skill never fascinated my wife, ever. I always got the "just turn on a light..." whine instead.
> an "arrival mode" for their lighting,

Which can be obtained with a basic 'motion sensor' light switch for one critical light near the door, with no need for internet connectivity or smart-phone apps.

Turning lights off remotely is nice when you intended to go back to a room, but got ready for bed instead. You could turn it off with an app or a voice command, but instead you have to get up and walk to the far side of the house to turn it off.
Having to get out of bed and go turn the switch off with your hand builds character.
> I wouldn't. What's the point exactly?

If the smart house/ambient computing stuff ever works as described, it will be really useful for people with disabilities.

"anonymized data"

That's if you're lucky. I don't want to sound paranoid but if there are sensors in your house its only a matter of time before someone wants to use them.

I think the advantage of "dumb" devices is that you can fully understand how they work, at least at a level of abstraction that covers all possible functionality you care about. For smart devices, you can't fully understand how all of them work, so you have to trust that they understand correctly what you want them to do and are actually attempting to do it. Or if you can't trust it, you have to spend time and effort verifying it. And you also have to trust (or verify) that they're not doing stupid/malicious things like punching holes in your firewall or leaking your personal data.

In short, you can simply use a dumb device, but you have to manage a smart device, and that requires much more mental bandwidth.

That's a good point about using a device vs managing a device. We need more "dumb" devices that have optional layers of smart enhancements. Best of both worlds would be my preference for replacing things like a kitchen timer or light switch.
This is definitely how I view the "dumb" vs "smart" automation. Complex things are hard to reason about, and as much as you might claim they "just work" that's only true until they suddenly don't, and then you can't fix them easily. Simple things are easy to reason about. They may not always work, but at least when they don't it should be pretty obvious why. I bring this up a lot in my "why the Linux Desktop sucks" rants, but it sadly applies to a whole lot of technology. So many complex systems introduced to solve non-problems or problems created by the complexity in the first place.

It's enough that sometimes I want to just give up computing and live in a cave.

Soft touch "buttons" vs regular mechanical buttons.

Soft touch buttons are the bane of my existence. With most buttons, there's no feedback as to whether or not the button has been pressed. It's also difficult to develop muscle memory so you can you use the buttons without looking. For this reason I can't stand typing on an iPad's screen keyboard. Also the reason turning on a modern TV involves me dragging my fingers all over the bottom hoping I hit the power button.

The extreme of this is newer Macbook touchbar. Vim is a nightmare to use. Did I even press ESC? Can't tell.

Another great example: http://www.homecrux.com/wp-content/uploads/2015/01/GE%E2%80%...

I have a Ford Fusion 2013, AC and radio is controlled with those soft touch buttons, man I hate those buttons, I don’t want to take my eyes from the road just to change the radio! I don’t get why the obsession about putting “touch” functionality to every single damn thing, they don’t provide a single UX improvement.
Still a remnant of the hype about certain Apple products, it seems.
Huh, downvotes. Seriously, I was there, working in a related field at the time of the iPhone and iPad hypes, and that was exactly when ill-advised uses of touchscreens popped up everywhere. They became the thing every product needed to have to be perceived as modern.

Just to remind you how hypey it was, "tablets will replace PC-type computers for most consumers in the next few years" and "world's first iPad DJ". I wish I was making up the last one.

Considering the number of touchbased phones and tablets sold world-wide, compared to traditional PCs, I think the hype was understated.
> they don’t provide a single UX improvement.

They don't (well, soft touch radio/heat/AC/etc. buttons). And in my opinion, this is one place where some more safety regulations really are needed to force a return to physical buttons.

But they _do_ provide an additional line item in the list of "comparison specs" to show a check-mark for their car and no check-mark for a competitors product. And that is the reason for the 'race' to add touch screens for everything that used to have a hard button. Look at our car, it has so many more "features" than theirs....

That, and the fact that a single touch screen can replace a whole host of hard buttons, eliminating the design, tooling, manufacturing, shipping, and assembly costs associated with all the old hard button devices.

I moved from a cheap Toyota Yaris to an expensive Subaru Legacy.

Not only do I have a touchscreen radio, but the climate control section is a large array of buttons, with the exception changing the temperature with a knob.

Every single fucking action I need to perform in that area requires that I take my eyes off the road. Want the heat on my feet instead of my face or the windshield? Then I need to press up/down and iterate through the air directions. Need to increase/decrease the air speed? Find the left/right arrows and press them. Oops, I pressed the Defrost button which turns on the windshield vents full blast! Oops, I pressed the "Auto" button which does... something. No fucking clue what that button does.

As shitty as my old Yaris was, I had a knob for air position, another knob for speed, and a button right in the middle of the knob for rear defrost. The knobs were huge and clunked into place. After a week in the car I knew exactly what to do without ever looking down, and I miss it dearly.

And my touchscreen radio takes like 10 seconds to boot and it warns me never to look at it while I'm on the road or some bullshit. If I need to change stations I need to look down and find the touchscreen buttons. I'm all mad thinking about it now, wish I could return the car just for this shit.

And in my opinion, this is one place where some more safety regulations really are needed to force a return to physical buttons.

...I've always wondered, if it weren't for some regulations, would manufacturers replace the steering wheel, pedals, and gearshift with a touchscreen? That would be really scary. The value of the feedback of physical controls can't be underappreciated.

> I've always wondered, if it weren't for some regulations, would manufacturers replace the steering wheel, pedals, and gearshift with a touchscreen?

Sadly, I suspect the answer is yes, some maker somewhere would do just exactly that.

Tesla did that. It’s called “auto-pilot”.
After Apple ruined the podcasts app in iOS 11, I switched back to my old way of listening to podcasts: download on the computer and listen on a Sansa Clip+ player (with Rockbox on it).

Especially the first few days were pure delight. A player with.. hardware buttons! I can skip tracks, jump back and forth, pause, everything without even looking at the thing! Several years of using touchscreen-smartphones for this had me totally forget how it is to use an MP3 player with physical buttons. Really, I didn't expect at all how much I'd love it. I've now gotten used to it again and going back to anything touchscreen-based is not an option at all anymore. And on top of that the joy of knowing that there's no ad-company tracking how/when/where I'm listening.

I have nothing against your solution, but for those of us who prefer not to have to carry and sync another device, you can get (wired) headphones with volume + skip buttons for just a few bucks.
My only experience with the Clip+ was as a runner, but I switched away from my phone because I was terrified of breaking it. But the downsides of the Clip+ drove me back to my phone and a good, hard case for it.

- had to hook it up to my Windows computer and download songs instead of just making a playlist on Apple Music or queueing up podcasts in Overcast

- if the sun was out, I had to hold the Clip right against my eyeballs to make anything out

- no 30 second skips, no chapter selection

And honestly, I'm not sure why you'd give up on the phone altogether without researching other podcast apps. Overcast is the best.

Note that I'm using the Clip+ with the Rockbox [0] replacement firmware, which adds features such as configurable skip lengths, etc. Without Rockbox, the thing isn't nearly as good :-)

> And honestly, I'm not sure why you'd give up on the phone altogether without researching other podcast apps. Overcast is the best.

I looked at them, didn't find a single trustworthy one. Overcast is a good example of that: it's often recommended, but it wants me to make an account, which is completely superfluous and probably just a way for them to collect data about my podcast habits. No thanks. I only want podcasts, I don't want a cloud service to go with it.

[0] https://www.rockbox.org/

Just FYI, the creator of Overcast is a notorious anti-advertising and pro-privacy curmudgeon, and he even wrote Skeptic's FAQ for people put off by the account thing: https://overcast.fm/skeptics_faq

It's very much worth it.

A bit dumber than a “smart device”: the new AppleTV remote is a tiny touchpad that communicates with the box via Bluetooth. The problem is when this thing ends up in a cushion or gets squished under a pillow it pauses, stops, or starts rewinding the video. Compare that to a cheaper IR remote where it simply can’t communicate with the unit in those situations. Turns out line of sight is a feature for a TV remote.
Yeah, that's happened to me, too. It's small, slim size makes it fit just about anywhere. I do like not having to point it at the box however, and it means the AppleTV unit can be placed out of sight. It is a tradeoff, though.
With HDMI-CEC, your TV can receive the IR signal and pass the instruction back to the AppleTV through the wire, without having to place the unit in sight.
So that's what's going on when I've nudged the universal remote when using the AppleTV. Cheers!
The Logitech harmony remote also works over radio so it doesn’t need line of sight.

But it’s not so small that it gets lost.

I never use my Apple TV remote. It’s not practical enough.

And I much prefer mine over any other I’ve used. One of us must be wrong ;)
The AppleTV 4 remote is, to me, the worst remote control I have ever used, with a big margin. I don't understand how this thing could ever get through usability/ergonomics testing at Apple.
> Turns out line of sight is a feature for a TV remote.

It can be, but for bad TVs you have to do "remote tai chi" to get LoS between the remote and the TV sensor from wide angles or when there are people or things in the way. I wouldn't call digital signalling better, but there are some benefits.

Often that means you have to replace the remote batteries.
With my old TV, I'd pick up the remote and just point it at the wall behind me so the beam would scatter. It was much more likely to work than pointing the remote at the TV.
I get his point and to some extent[1] I agree, but it seems a lot of his examples are caused by his technology choices. I can run Waze in my car with no issues (just use an Android phone); I don't use smartwatches for other reasons (I don't wear a watch, period); but when I did, it was always-on (years ago!).

[1] I agree that old technologies are always superior in some aspects to new technologies. But, crucially, they are inferior in others, that sometimes turn out to be more important. When that is the case, the new technology will typically gradually become better in all aspects (i.e. with time the inferior aspects get fix; but yes it may take a while)

Smart meters. Since working in IoT on a "dumb" meter I figure the killer feature isn't automatic temperature and power regulation, it's knowledge. Once you know where and when you are using lots of power you can easily figure out a solution that fits your situation, rather than endlessly coddling an intrusive system.
Dumb stuff have always been better. Most of the smart things are useless.
I love how "techie" people needs nurturing to realize what makes sense and what not. So 2 years later, they realize what other people know the moment they hear about some product.

Something Silicon Valley engineers do. They invent silly things just cause they know how to code bluetooth communications. Like a water bottle that measures how much you are drinking during the day for example.

But hey, it's not their fault. It's hacker news fault. Didn't you spend years telling them "Ideas are worth zero. Execution is king". Now you have highly productive dumb people you helped creating.

I divide my officers into four groups. There are clever, diligent, stupid, and lazy officers. Usually two characteristics are combined. Some are clever and diligent -- their place is the General Staff. The next lot are stupid and lazy -- they make up 90 percent of every army and are suited to routine duties. Anyone who is both clever and lazy is qualified for the highest leadership duties, because he possesses the intellectual clarity and the composure necessary for difficult decisions. One must beware of anyone who is stupid and diligent -- he must not be entrusted with any responsibility because he will always cause only mischief.

https://en.wikipedia.org/wiki/Kurt_von_Hammerstein-Equord

A phone with a touch screen instead of physical buttons was a dumb idea until it was executed well.

Smart devices aren’t a bad idea, they just haven’t been executed well enough.

I never said "all smart devices". I was obviously implying some smart devices for some use-cases. A bluetooth connected water bottle is well executed, but poorly ideated.
They're still bad phones - you can't answer or dial them without looking. It's just that they're far more useful as tiny internet tablets and people have in many cases gradually moved to text chat over voice.
"you can't answer or dial them without looking."

... How often did you dial something without looking before? You probably had a handful of numbers memorized. Otherwise, you needed to look up the number in a phone book or address book. Perhaps you memorized numbers for speed dial?

Now you can dial by talking to you phone if you'd like, so long as you have the person's number in your phone's computer. You can often dial a company directly from their page or from the phonebook... which is online.

And finally: Of course people have gradually moved to text. Voice is often not needed, and I would rather not stop everything to talk to most people.

Now you can dial by talking to you phone if you'd like

The Nokia 3310, from 2000, already had voice dialing. And while the recognition may not have been as good (I frankly can't remember), it certainly didn't take the time that my current cheap Android takes to even start listening.

It worked good enough. I pressed a button and said Pizza and it immediately called my local pizzeria. For some reason I’ve never got that to work smoothly with Siri.
I was honestly thinking about landlines when I wrote this :)

Sidenote: I've never used the voice recognition myself, honestly.