I stopped reading at Try to recall a time back when the Internet didn't exist... Where things like cars and phones functioned on their own in a much more reliable manner.
I grew up in the 80s and have owned several cars from that era. They were significantly less reliable than cars of today. Carburetors were still present (though being phased out) and there were always vacuum hoses that, when they leaked, caused your car to run like crap.
Also, phones: Cellphones were extremely expensive and unreliable. Landlines were extremely unreliable at reaching other humans because as often as not, the recipient wasn't home. Also, I remember all kinds of problems with landline wiring leading to noisy connections. Also, unless you were calling a close neighbor, you paid high per-minute charges.
It's really hard to find something that was better in the pre-internet era. Cars and phones aren't it.
I think you're overlooking the "on their own" part. The OP isn't lamenting that things are less reliable in general, just that they aren't reliably capable of working offline or on a local network.
Phones are maybe an odd example, but it's somewhat true: Traditional phone systems let you make local calls even if long-distance lines fail, or within a building without leaking data externally. Very few people would ever need (or expect) newer voice/video chat systems to be usable in this way, and they largely aren't.
heck, POTS line phones can even function with a power loss since they derive their power from the phone line (unless you have a POTS line phone like a cordless where it has to charge a battery and operate a 900mhz connection to the handset.)
I was able to use the Windows XP Telnet app to chat and send ZIP files to another person when the Internet was down during the 2011 revolution. Hell, it was even better because only party had to pay the call's cost!
POTS is amazing, and its sad to see it going away :(
I am not sure about now but accordingly to what my friend told me at the end of 90s and in the beginning of 2000s in Russia they would wire whole apartment buildings as a very fast LAN with P2P capabilities. So if you urgently needed to download whatever there was very good chance that your request will not cross boundary of said LAN.
I don't think I've ever heard a fast-busy signal from my cell phone, even once. But it happened lots with landlines when too many people were making local telephone calls. Circuit switched networks kind of suck in a lot of ways. I don't think I would say traditional phone systems were as reliable as modern telephone and data systems.
Wait for a disaster to happen in your area - a time when you'll want your cellphone the most. At which point your cellphone will be... well, a good camera.
And personally, as someone who grew up in the 80's, let me offer up a counter-anecdote: I don't ever recall hearing a "fast-busy tone". There was never a time we couldn't use our land line (at least until we had a modem, that is).
I live in a disaster-prone (wildfire) area. I'll take cellphones over landlines every single time.
1) In the event of a major disaster, you've probably evacuated. You can't take your landline with you.
2) Telephone poles burn.
3) Even if the cellular system is completely hosed, you might still be able to get on a data network. All you need is crappy wifi anywhere (cafe, hotel, friend) and you can get an email out.
I have to admit, I'm not entirely sure why there's so much hate for reliable technologies like landlines. All I can guess is that they stole someone's lunch money. Or people feel the need to justify their cellphone purchases by coming up with specious arguments against the "competition".
1) Not all disasters can be evacuated away from. Tornadoes, for example.
2) So do power poles. And those telephone poles are also probably carrying data.
3) You can't dial 911 via email.
Cell network congestion is a well known phenomenon when disasters strike. It's been seen multiple times in the past. Yes, landlines can and do fail as well, but they're more resilient than cellphone networks.
Perhaps this will flip in the future, but it hasn't, not yet.
1) How's landline infrastructure going to hold up to tornadoes?
2) Yes, power goes out too. Cellphones have batteries.
3) Yes, you can dial 911 via "wifi calling". I've done it.
I don't know about "hate" but I do think landlines are overrated. You already have a cellphone. Paying for an extra service and apparatus for the tiny possibility that you need to call 911 when your cellphone can't - why? Who else are you going to call? Do you even know their phone numbers without your cellphone?
I also think you're undervaluing cellphones by assuming they only talk on the cellular network. You just need to get on a data network somewhere and you suddenly have a wide variety of communication options, including phone calls. TCP/IP was designed to withstand nuclear war.
I grew up in the 80's too. What you say is very very dependent on your service area. I have fantastic coverage where I live, but I live in a big city. I've never had a problem with initiating a call on my mobile. It's likely my operator has done a better job of pico-celling my area than yours has. You might consider switching.
I got fast busy tones several times a year back in the 80's, and it usually corresponded to whenever a concert was announced. Maybe my phone company back then was as bad as your mobile services company is now?
Well, that relies on a couple of things. A disaster can take phone lines out as well as electricity. You might not have zero cell phone service if towers are working. Land lines would get clogged at times and simply fail you (I've had this happen).
You generally have some charge on your phone, though, and lots of folks have power banks now - or cars they can charge their phones with.
Don't forget all the times when disaster strikes and the cell phones are a real help: Car accidents, for example. At one time, you were reliant on you - or someone else - stopping by a house and hoping they'd be home. And I'm sure there are lots of other examples of this as well. Now that laws changed, cell phones can pretty much always call 911 as well, even without service at all: This was not the case with phone lines. If the house didn't have service, you couldn't get help.
In large disasters, cell phone networks have historically become so bogged down that even with fully functional towers and phones, you couldn't use the network. Perhaps sometime we'll have sufficient provisioning of cell phone towers to absorb such large spikes in usage, but we're not there right now.
Crucially, however, there is no profit in getting to the point of being properly provisioned for emergencies, and no regulation pushing for it. Land lines, in contrast, are regulated to push for availability during emergencies.
Yeah, I don't think cell phones do fast-busy. They just timeout dialing without a carrier driven noise (mine have usually made the boop-boop noise like when the other party hangs up). Of course, with a cell phone, you can never tell if it's the network is busy or the tower is out of range, or the tower is broken. With a landline, you can tell if the line is dead (totally disconnected), really broken (voltage on the line, but no dialtone), or just pretty overloaded (dial tone, but fast busy for all numbers). And, if you made a habit of calling very busy numbers (like Netzero dial-ups), you could learn the tricks to get better priority; dialing the first N-1 digits, waiting a few seconds, and then the final digit would get me a waiting modem way better than just auto-redial on busy.
Yes, this 100%. I miss being able to talk to people like they're in the same room, with no detectable latency, the way that local landline calls used to work.
I think most people don't realize why phone calls suck these days; it just feels awkward and cumbersome, but it doesn't have to be that way.
Not with VoLTE and FaceTime Audio around now. Inter-carrier calls can still suck, but I now have noise canceling headphones which enables me to get a much more intelligent signal out of the mouth of the other caller than I could with just the call speaker held up to my ear.
There’s a lot that still sucks, but I don’t miss landlines the slightest bit. Scheduled do not disturb is a godsend next to trying to remember to turn the ringer on the landline in the hallway off at night.
You're right that cars are more reliable today but I follow a lot of mechanics on YouTube (Car Wizard, Kilmer, etc) who say that cars were easier to work on back then. These days (at least in the luxury segment) there are a lot of electronic components that can only be repaired by the OEM. We are also in a strange transitional period with the addition of auto start-stop, hybrid engines, etc that also complicate the car.
even that is changing according to youtubers like Kilmer. IIRC he has lambasted his favorite car maker Toyota for having a lower quality today than years past.
New cars are based on black boxes that need to be coded to work together so if they break it is generally the authorised repair shop or the scrapyard.. So sad :-(
Something you're overlooking with your concept of "reliability" is that cars produced before 1980 can be fixed using pretty basic tool. It's conceivable that a car from that era could continue to work well into the 2100s. Because most of what can break on can be remanufactured by a reasonably-skilled machinist.
Yeah, cars need to be fixed much less often, and they are substantially less fickle vis-a-vis temperature and operating conditions. But the moment the software to repair newer cars is gone, so too goes our ability to easily fix them.
This isn't some far-flung dystopian idea either. It's totally possible that the hardware & software used to repair today's cars is difficult or impossible to come by in 30 years. Some of this stuff is five- and six-figure investments. Maybe stuff for GM or Honda vehicles will be accessible, but what about smaller companies like Mazda or Subaru?
Yes! To me, reliability goes hand in hand with repairability. Id rather have something that breaks every 2 years that can be fixed with a wrench than something that breaks every 10 years and cannot be repaired without a $2000 trip to the dealership where they use a $20,000 diagnosis computer to troubleshoot.
There's "fixable in theory" and "fixable in practice". As a former gearhead, I'd argue that cars from the late 70s/early 80s are effectively unfixable - it was the era when emissions standards were ratcheting up, but controls were still mostly analog. Engine bays were a giant morass of rubber vacuum hoses that nobody sane wants to deal with anymore.
Mid-70s and earlier cars tend to be simpler, but they would never pass modern emissions tests. We picked clean air and fuel efficiency over simplicity. I don't think this was wrong.
It's really hard to know what to think about the direction cars are heading. Yeah, there's a lot of closed software and electronics. On the other hand, I have at least one friend with a hacked Tesla. And the huge emissions regulatory and inspection regime that prevents you from making changes to engines doesn't apply to electric cars. So it's a mixed bag.
In the US, older cars generally just need to pass the emissions regulations in effect when they were manufactured.
In practice, that means that most states which have any emissions inspections now just connect to the OBD2 port, check for ready and no faults, and pass the car, with cars that predate OBD2 being either visually inspected for obvious defects or not inspected at all for emissions.
"Easily fix" is relative though. It was never a majority of car owners that fixed their own cars. I would never have maintained a thousand-dollar toolbox to fix and maintain my car - not then and not now. I'm just not mechanically inclined. For us, there's no difference - we need to take the car in to be serviced either way. The complexification of automobile engineering has disenfranchised a small minority of car owners. But for the rest of us, we just don't need to bring the car in for servicing as much. It's a huge win for the majority.
Having a machine that can operate for 30 years with minimal or no maintenance is a type of reliability.
Having a machine that can operate for 200+ years with refurbishment is type of reliability.
There are pros and cons to each. The former is good so long as your country continues to grow, remains wealthy & educated, and retains the ability to constantly produce better replacements. Not every country has that luxury.
The VW Type 2 (what Americans call the "VW Bus") remained in production for the Brazilian market until 2013. That's a poor country and cars are expensive, so the ability to have a vehicle last a lifetime is very important. The Type 2 is something that can be repaired by almost any machine shop with basic fabrication tools.
You'll see this a lot in developing countries. Where "new" automobiles would be recognized as a 1980s or 1990s model by those living in developed countries. Manufactures keep them going because the tooling is cheap and easy to keep going.
Those 80/90s cars are dead reliable too.
Once the spare parts for an American 2020 Honda Civic are exhausted, it's unlikely that new parts can be fabricated. That's fine, so long as there's a 2050 Honda Civic or similar to available replace it, and that such a vehicle can still be afforded.
> I would never have maintained a thousand-dollar toolbox to fix and maintain my car
You might be surprised by how much work you can do on a 90s-era Japanese car with a 14mm wrench and a screwdriver.
I would say that even CIS/K-Jetronic injection (well into the early 90s) will be quite maintainable 50+ years from now. Even L-Jet doesn't require anything particularly sophisticated, but there are likely some already obsolete components in those boxes (though hundreds of thousands or millions of them in junkyards, surely far more than we're consuming).
Your argument is a strawman. Car and phone or bad examples to prove your point because car reliability improved because of factors unrelated to the internet, and phone reliability improved because a better reliability of satellites and computer networks. Better examples are things like watches or desktop applications which seem to work fine before they were connected to the internet.
While it's true that cars are way more reliable nowadays. There are still engineer tradeoffs being made against reliability. Lightweight (fuel efficient) oils are being used, gas tank emissions capture (a system of rubber hoses and solenoids which fail easily), EGR systems, etc all are a detriment to reliability even if they have other positive uses. Today's cars could be far more reliable if that was a bigger design goal.
> Irrespective of why, cars and phones did not function in a more reliable manner back then. The statement is false.
You're missing the point, and misinterpreting that statement to mean something far broader than it actually says. The article is about unnecessary internet dependency, not about the mechanical reliability of cars, and it's a derail to make it about that.
Also, you're interpreting "a time back when the Internet didn't exist" as the 80s. For most people, the internet didn't exist as late as 1995, and I've driven some pretty reliable cars from that era.
Depending on your definition of "internet connected," I'd disagree with your first point. Most new cars come equipped with the hardware for a data-enabled cellular connection and many have been for years. While it's not a feature you necessarily see or know about, it's there. Presumably, the ones that are used by a service like OnStar are constantly polling in the background to get status updates for the account the vehicle is associated with.
But yes, to my knowledge, most (all?) operate fine without a cell signal.
So you missed the other half of that example? "Now consider current times when your thermostat becomes completely unresponsive in the middle of winter because a vital network goes down. Imagine you're not able to operate your car during a crisis because networks are down?"
Something, something, "Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith. ... Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something."
I have a Nest thermostat. It works fine without the network. I have a modern car. It works fine without the internet. Several of my friends have Teslas. They work fine offgrid.
Wasn't there recently a story of a Lamborghini which wouldn't start because it didn't have a cell signal (it was in an underground garage)? Some sensor had been tripped, which made the car require network validation before it would start back up.
And don't Nest thermostats occasionally have an issue where they default back to somewhere in the 60° range due to a server connectivity issue?
As noted in the HN thread, the Ferrari was disabled due to an anti-theft device and couldn't be unlocked remotely due to connectivity. Lack of connectivity is not responsible for the car being disabled; an attempted "theft" (from the device's perspective) was, and a stuck car is exactly what the owner asked for.
Even if there have been a few historical software bugs that caused some Nests to change temperature, it did not "become completely unresponsive in the middle of winter". Worst case is you walk up to the device and turn the temperature back up. Also, there's nothing in that bug report that suggests it has anything to do with internet connectivity.
A quick google search shows plenty of people getting locked in non-internet-connected cars.
The modern world is complex. Even before I replaced my thermostat with a Nest, my furnace was already way more fragile than my fireplace.
Absolutely! But their failure cases weren't due to the inability to access a network or a server. This is whole new set of failure cases which exists over the top of the old ones.
Then big brothers came along, built centralized systems, you bought devices you have zero clue how they work. With your own money you let them collectively manipulate you.
I believe this is what the author tried to explain.
Internet and technological advancement is not evil per se, too much power in a very tiny group is going to be problematic.
Vehicle engines are better made today but now we are introducing new points of failure intentionally. It is as if there is a certain quantity of failure that must be present. In fact a vehicle produced just prior to "everything connected to the internet and the manufacturer can turn it off any time if they don't like how you're using it" probably represents the peak of reliability and now we are going somewhat downhill.
The NEXT Google Outage will involve 32 million cars refusing to run because they can't log in to your Google Account.
Most audio gear was built to last then, my sound system, speakers and amplifier is from the 80's and still sounds so sweet. The only other notable mention was my Nokia 2110 which I had for from 1996 until 2001 and was bulletproof. (but alas not waterproof)
They actually did, but the updated versions were for new sales only.
OTOH, software update is being widely abused nowadays, there's no pressure to get it right off the bat. As long as the update framework is robust, everything else can be weak since you can send another patch whenever you want.
In the 90's game publishers started to make things available via BBS, FTP, AOL and CompuServe and others. And it was already taken advantage of in quality control. Also "mail me a floppy."
Yes they absolutely did, they just couldn't get them. Games still shipped buggy, you just had to get used to it, tough. I work with people who were in the games industry in late 90s and the idea that games shipped without bugs is just funny to them.
Yes, of course modern games are more buggy, because like you said, they are more complex and yes, the assumption is that they can be patched later. But the "old games didn't get patches and they were fine!" Thing is really really not true.
Survivor bias: the games people fondly remember and want to go back and play decades later are generally the ones with few if any problems. The buggy crap got purchased by an unlucky few and then was quickly abandoned and forgotten.
It's also why classic music, books, movies, etc are great - decades of pruning leaving a short list of best-ofs.
Also video games that didn't get regular, free content updates. Top of mind because my partner and I are doing another playthrough of Terraria and there's once again a bunch of new content for us to go through.
Software updates were necessary to bring about internet play without rampant hacking and cheating, and also can provide general fixes to you giving you better value over time, so I’d have to disagree. If you don’t want to play online, you really don’t need software updates anyway.
I do not have 80s experience but starting from 1994-1996 (do not remember exactly) I do not recall any major problems with any of my cars. There was one exception when the transmission broke (I could still drive to repair shop) but from what I see this kind of shit can happen now just as likely.
>It's really hard to find something that was better in the pre-internet era.
Big ticket home items like fridges, washer/dryers, TVs. They're all planned to shit the bed just after their warranty wears out so that companies can squeeze out extra money from them. Nothing is made with pride anymore. There's many facets as to why this is, but as a civilization I don't think we're ever going to be able to go back to the way "the way things were"
So, once upon a time, nothing emitted or depended on receiving EM waves. Then, we learned how to use electricity and started putting it into everything. Then, and only then, did we start developing government regulation to make sure that your new device for sale didn't broadcast EM waves that interfered with any other devices in the vicinity.
We are getting to the point where we will need an Underwriter Labs-style independent body to say, to anybody selling an internet-connected device, "your device is not secure enough to be sold; you are forbidden to sell it until you get your security house in order".
This wouldn't solve every problem, but it would solve a lot of it, simply because the cost required to pay for good security would cause most IoT devices that don't really need it, to get rid of it, in order to cut cost and reduce time to market. Only if it if really added a lot of value would the maker be willing to pay for good security; that does not include most IoT today.
For appliances as they're currently developed (completely locked down, firmware infrastructure owned and operated by the manufacturer), this definitely seems like a good idea.
I do think there's a fine line to walk between these devices and "general purpose computer"-type devices though, and I'd worry about this at least a little bit given the trend that appliance-type devices seem to be taking over the world (especially if you include phones).
Separately, I do still want to see a UEFI-style secure boot architecture for appliances. An appliance can have a default mode where it only accepts firmware signed by the manufacturer. But then if a user wants to install their own key and load their own firmware (because the manufacturer has stopped supporting the device, has gone out of business, writes shitty firmware, or just to use hardware that they own in a different way), most devices should make this possible, and it would be interesting to think about what a legal requirement would look like here. A manufacturer could gate access via a configuration menu, or by cutting a trace/jumper or with a write protect screw like on Chromebooks. And you could have entire classes of devices where open access is explicitly not allowed, like medical devices.
But smart speakers? Manufacturers should not be allowed to cryptographically guarantee that they become trash after two years.
Many or most aren’t. I have a bunch of home automation devices and I’m planning for more. Most of them don’t require any internet connectivity, the only one that does (I still have Alexa connected to Home Assistant, planning to replace her with Rhasspy but didn’t find the time yet) is purely an extra method of controlling things so nothing breaks without internet.
I see people buying internet-required devices as the same people who buy the registry cleaners. They don’t know enough about computers themselves, and the advertorials/computer magazines don’t do them any favors.
> I see people buying internet-required devices as the same people who buy the registry cleaners. They don’t know enough about computers themselves, and the advertorials/computer magazines don’t do them any favors.
In some cases, you don't have a choice. As far as I know, all current Roomba models have their schedule programmed via an smartphone app. I don't have one, so I don't know if it exactly requires the internet, but it's close. Older models (like the ones I have) used buttons on the unit and had no need for any kind of network at all.
Actually I don’t mind things being controlled by a smartphone over the Internet. That’s nice and convenient. What I object to is things being controlled by a company’s internet service, commanded by the user’s smartphone. One just requires my smartphone, the device and an internet connection, the other adds the company as an unnecessary middleman.
> Actually I don’t mind things being controlled by a smartphone over the Internet. That’s nice and convenient. What I object to is things being controlled by a company’s internet service, commanded by the user’s smartphone. One just requires my smartphone, the device and an internet connection, the other adds the company as an unnecessary middleman.
That's not quite right. For all practical purposes, internet control also requires a server sitting on a public IP address somewhere, and that's where things start break down.
1. If you run the server yourself, then you're dealing with patching it and otherwise supporting that infrastructure.
2. If a third party tries to run a server as some kind of service, they're stuck supporting all the protocols and schemas of all kinds of internet connected devices, since AFAIK there are little to no standards. Then there's the problem of funding such a project.
With both of the above situations, the device also needs to support configuring a different server.
3. If a manufacturer runs the server, then you avoid most of the problems of the above options, but then you're at their mercy, and liable to loose functionality at some indeterminate point in the future.
The invisible hand of the market, PBUH, appears to have determined that 3 is the best (most profitable) option.
> For all practical purposes, internet control also requires a server sitting on a public IP address somewhere
We really need to get over this hurdle. In the ipv4 world, yea, users generally have one public IP and NAT complicates things a little. In the IPv6 world, every device should be publicly addressable and it should just be a firewall rule to make it accessible to your smartphone app. That way, no server is required, so the question of who runs it is moot.
Even with IPv6, you're still in scenario 1 from my list. Also, using it would really only help with carrier-grade NAT. Someone with the amount of skill needed to setup a server probably has the skill setup port forwarding and dynamic DNS on their router.
Which reminds me of the DSL modem/router combo my roommates had gotten in my first apartment before I moved in. The router was MSN and had no local configuration page, all configuration had to be done through some external Microsoft server that didn't work by the time I got there (so I could never do any port forwarding).
TP-Link has 2 years of warranty, 5 years for their "business" product line. The terms seem clear.
Probably iRobot has similar terms and conditions when they sell their robots, right?
I mean, eventually someone has to pay for R&D. (Sure, I'm not saying this is the best way, but apparently people really don't like paying for R&D - see how the Ubuntu phone, OpenMoko and similar projects basically flopped.) It turns out to really get money (financial capital) for research, the best way is to be in the market.
Another way to rephrase the question is: why don't
networked devices require signoff from a professionally licensed engineer on its specific software implementation before it goes to market?
You see "forethought" in devices offered to the public when regulation DEMANDS it: think cars, bridges, medical devices, because (or so it's been rationalized), the public needs accountability (a named, professionally licensed, buck-stops-with-them head to roll) when that product can cause harm.
But software harms TOO!. Think privacy, banking, relationships. Those can be harmed. It's always felt to me like a historical accident that networked devices, really most software in general, slipped past this accountability requirement. Without a specific , named accountable person, security seems to fall into the not-my-problem phenomenon, and continues generates articles like from the OP.
In the imagined future where signoff from professional engineers was required, you'd see pushback against a vendor by the engineer until the implementation was secure, because the engineer's licensure was on the line until it was secure. And the vendor couldn't just "shop" for a favorable signoff, because every engineer would be held to the same ethical standard and penalty.
Software engineering licensure is an old debate. To say the least, I can easily imagine it making things worse, where let's say, all software has to be written in Java and get a rubber stamp by someone who hasn't actually built anything for years, and then still nobody gets held accountable when security breaches happen. Meanwhile, a bunch of college dropouts build something that's actually good using bulb, and then just have users sign away their rights in a contract instead of getting their software certified. And most of the world still runs on the Linux kernel which also had no interest in getting certified.
Ah, yeah. Well. Finance and accounting is a very regulated space, yet there are auditors that don't sign off on stuff, and miraculously the same stuff gets signed by a different one a bit later.
Also - and I'm not saying requiring sign off is a bad idea - , but it's not a coincidence that secure walled gardens look like a manicured Black Mirror dystopia.
Software development is one of few areas where a simple folk with little money can start something great and make big bucks. Implementing your proposal on a global would kill it all and whole lot of innovation with it.
As they say "live and let live". We already have this control where it is needed. Leave the rest to manage their own affairs.
This is pretty much why I have yet to buy into any internet of things devices and look really hard at what something can do, will do offline.
Maybe one day when all this is actual infrastructure and all that goes with the idea, standards, service entities who will be there, are regulated, etc.
Maybe then, I will reconsider.
And let's be clear. I have no problem with pioneers pioneering. Have fun! I just have different priorities.
I don't have any problems with devices being connected to the internet, per se. Some issues may arise for devices with sensors like cameras that could be abused.
I have 3 robot vacuums (one is technically a mop). Two are internet connected, one is not. The two internet connected ones are kept up-to-date and get new capabilities from time to time. And one of these is also accessible from the local network too. This is great. At least I know that, should the company decide to support the product, I'll retain some of the capabilities.
Which is something that happened to my Pebble. The community did step in and provide some missing functionality, but it wasn't a seamless transition.
The biggest issue is the amount of e-waste we are creating by having - otherwise perfectly functional devices - being thrown away because their command and control servers were shutdown.
There should be some mandate that companies are required to support the internet connected product for X number of years, just like cars. If they go out of business they should be required to release the control software.
I see all this connectivity and the back end needed to make it work a lot like basic electricity.
Early on we had competing visions, best practices were being learned the hard way and regulation was in its infancy.
Later on, with a lot settled, electricity became utilitarian, but not a given still.
My camp trailer fridge works on 110v or propane. The table radio in it works on 110V, 12V or batteries inside the device.
I have (dangerous) gas lighting in that unit as well as electric.
Power isn't a given and the devices from the era are designed to operate under that state of things.
This all seems very early. We lack foresight because people are pioneering. Users of this tech are also pioneering. Someone is gonna do it. Let's hope they do it increasingly well and sooner.
One day, this will all be real infrastructure, with all the girls along with: regulation standards the usual.
Until then it's kind of a wild west.
Play hard, but play eyes wide open too. Or pass and watch as it all forms to be more robust just like people did with electricity.
Customer onboarding experience is a big one. The internet is a reliable way to communicate with/setup a device. I wish there was a better way, as companies keep having to reinvent the wheel. (Creating ad-hoc wireless networks, weird strings to input, transmitting audio/images to your phone).
I'm pretty sure the onboarding experience contributes to poor reviews, so the goal is to make it as easy as possible.
I remember the days where all devices came with a CD/floppy with software you had to install. (I had a device once that had to port scan my home network to find what it needed to connect to.)
I have to agree with the OP on this one. I recently have been looking at some different home security hardware to add to my home. In particular, trying to find a video/intercom doorbell that I can hookup to my self-hosted Home Assistant server without it reaching out to the internet is a very difficult task.
Even if you know what you want, and you know it's possible, it's difficult or impossible sometimes (without building it yourself) to find hardware that doesn't rely on internet connectivity and some user account.
In the author's specific example, the answer is everybody wants one of those cool security cameras, but no one wants to invest the time to learn some simple networking skills, such as using a router UI to find/assign an IP address... You can create a very nice system using a software package like Blue Iris so that your security cameras function with or without Internet access.
Because the demonstrable[1] security risks for most people of their vacuum cleaner or thermostat getting pwned are non-existent.
[1] I'm not talking about theoretical geek-fantasy concerns about how the Mossad, or some other sufficiently sinister and all-powerful organization[2], can stop your heart by varying the power setting on your internet-connected Roomba, or map out the internal layout of your apartment and fly a poison-dart-tipped AI-controlled drone into it. I'm talking about actual, demonstrable harm to the typical end user. People grousing on the internet about security have done an incredibly poor job of this.
[2] That for some reason is futzing around with thermostat zero-days, instead of just taping a bomb to the underside of your car, or scrubbing your teapot with polonium, or jabbing you with a poison-tipped umbrella.
Consumers in general have no internal model of security. It's enough that there's a sticker on the box saying that "even more secure than last year" or "2021-ready security".
It's cheap, it works for a year or two. It's apparently more than enough.
Laziness, deadlines, and lack of developer experience building things that aren't cloud or web centric. A lot of current developers literally don't know how to build software that runs at the endpoint and can run offline.
Just a thought but I have no internet connected devices apart from PC's and laptops older than 5 years that still work as advertised. OTA Updates only work if someone at the other end is writing them.
On the other hand, electricity replaced many things that were either more dangerous in themselves (like gas lighting) or more physical work (like laundry). And the only significant downside (I'm ignoring the dangers electricity posed) was that it sometimes stopped working, although in those cases you probably had your manual fall-back (candles and a wash-tub).
I grew up in the 80s and have owned several cars from that era. They were significantly less reliable than cars of today. Carburetors were still present (though being phased out) and there were always vacuum hoses that, when they leaked, caused your car to run like crap.
Also, phones: Cellphones were extremely expensive and unreliable. Landlines were extremely unreliable at reaching other humans because as often as not, the recipient wasn't home. Also, I remember all kinds of problems with landline wiring leading to noisy connections. Also, unless you were calling a close neighbor, you paid high per-minute charges.
It's really hard to find something that was better in the pre-internet era. Cars and phones aren't it.