The so-called smart device generation was only ever a thin veneer of commercial smoke-and-mirrors over what is, in most cases, (the likes of) simple switches and timers and remote controls. Ironically made overcomplex by vendor lock-in and the "oh-so-smart" routing of your trivial controls through clunky bloated phone apps and international internet servers.
As in stasi phones, stasi devices, [strike]social[/strike] stasi networks... Everyone snitching on everyone, from self and family to passers-by. Everyone becoming an informal collaborator.
Spy is better. Stasi is a composite from "(Ministry for) State Security". These devices are little corporate spies, they have nothing to do with the State.
Sure in a way it's all part of the power structure that the elites use to run things. In another very meaningful way, they are competing power structures that help balance things away from outright totalitarianism. The distinction is not meaningless by any stretch of the imagination, even it the extent and depth is debatable.
I purchased a bunch of Phillips Hue bulbs and they have started to age out. I hadn’t paid much attention to the space and found out that Phillips was phasing the Hue ecosystem out in favor of just connecting everything to WiFi directly. I purchased two bulbs to give it a try and one has connectivity issues while the other has severe quality issues (flickering, refusing to turn off, etc)
I really like the Hue ecosystem but not being able to keep buying into it a mere 5 years later is a big turnoff.
One of the two newer bulbs from Philips supported Matter by updating the firmware. It wasn’t at all clear that would be the case when I purchased them.
Finally, I’ve learned to avoid anything that supports “Siri shortcuts.” I want actual support so my home app and thus HomePods can do the right thing out of the box.
It's been a while since I bought my Hue bulbs, but they support Zigbee and you can make homeassistant talk to them over zigbee2mqtt and a usb to zigbee dongle. I haven't used the Hue app in years.
WiFi connected IoT is a total no go to me, besides a hub that is as open and customizable as a PC, which handles any interaction the devices might need with the internet.
I've been using a CC2531 dongle, but that's a little DIY, being a bare PCB dongle that needs to be flashed with a firmware using a separate programmer cable (over, say, an RPi's GPIO). Back when I bought it, it was cheaper than a ready-to-go dongle and flashing it wasn't all that big a deal.
Ah, the wiz bulbs. Those are pretty dumb and I think the most effective approach is to put them on a VLAN that is firewalled off from the internet. You can control them with OpenRGB or Home Assistant. It's both a good thing and a bad thing I guess. If you disable the remote access via vlan firewall, I think the worst that could happen is someone gains remote access into your network and turns the lights off on you lol.
I bought mine in 2013, I have about 30 light around the house. Over the years 1 failed (it was turning back on by itself). I obviously have the old generation lightbulbs, the new one might be different, but for me they have been incredibly reliable.
Personally, I just want better integration between my smart home devices and for their protocols to be open. I wish I could easily script behavior, and something like running HomeAssistant or Homebridge as a compatibility layer shouldn't be necessary. Right now, each platform essentially creates vendor lock-in. You need to stick with one platform or your devices are pretty much useless. This could end up being better with the adoption of Matter and Thread over Matter, but I'm already seeing issues. The article even mentions that manufacturers are even being slow to adopt Matter and that Amazon is removing standards for remote control like ADB.
If I make an announcement using Alexa with an Echo speaker, I want my Google Home speakers/hubs and other devices to also get the announcement. If someone rings my Nest Doorbell, I want to be able to access the video feed on my LG TV or Echo Hub. If I have some form of sensor that I want to use from a manufacturer, I shouldn't need their hub to talk to it. Whatever I'm using as my control system or hub should already be compatible, but even with Matter-supported sensors/devices, these are all still issues.
you are not the customer. The customer is whoever is buying the data you have generated.
You just happen to be the one getting a "discount" on the hardware, enough for you to choose data-gathering devices over more expensive devices on the open market.
It's not just smart devices, it's things like vanilla autocorrect on my iPhone. It's done incredibly stupid autocorrects, like divide a simple, grammatically correct four letter word into two strings of non-word characters.
This is what we asked for when we discarded software development for "AI", aka "I'm done thinking, im just going to roll the dice on hyperparameters and evolve tech through random mutation instead of reason"
Apple swiping has trouble with even simple things. Maybe I just have poor precision, but when I swipe out “and” I get “Abbas” with surprising frequency. The Palestinian president is the only Abbas I know, and he’s not in my contacts. Meanwhile, “and” is one of the most common English words.
This. Ever since Apple has started using ML, the keyboard has taken a shit. Will they fix it? Nope. Why not? Because they don’t have to. People will still buy iPhones.
They're all awful. The third party ones are data loggers ans require retraining your typing, and the first party auto correct is terrible on both (and slightly more terrible on Apple).
I disabled autocorrect years ago, and every time I use someone elses phone I can't understand how they function.
It's amazing how on the one side of things we've got OpenAI, Llama, Gemini, Mistral, etc making mind-blowing LLMs that feel truly like the future.
And at the same time we've got Siri, Alexa/Amazon Echo stuff, and Google Home stuff that already feels dated and is actively getting worse with each passing year.
Funny that these two things are happening at the same time.
>
It's amazing how on the one side of things we've got OpenAI, Llama, Gemini, Mistral, etc making mind-blowing LLMs that feel truly like the future.
> And at the same time we've got Siri, Alexa/Amazon Echo stuff, and Google Home stuff that already feels dated and is actively getting worse with each passing year.
Initially, when they were new, also Siri, Alexa and Google Home felt mindblowing and "truly like the future". My bet: in a few years these LLMs will feel similarly dated.
Concerning your "is actively getting worse with each passing year" point: do you remember these HN posts from the last months where people were complaining that ChatGPT and GPT-4 have become worse than they were initially? In other words: it's already happening. :-)
LLM quality drops aren't disillusionment either - it's scale/cost optimization. And those are subscription models - Siri/Google assistant is bundled with device/services.
>Initially, when they were new, also Siri, Alexa and Google Home felt mindblowing and "truly like the future".
Strong disagree on that one. It was difficult to get them to do simple things like give the weather if a different location, play the right song. Even asking what time it is wasn't guaranteed to give you an answer if you didn't ask the right way. The marketting material made it sound great but the first hand experience was very underwhelming. The biggest use case at launch was making it say funny things.
At first yes, but about a year or two in was when you could do some really cool things, especially with Google. It could be astoundingly context aware, especially regarding a chain of on going queries. I have not seen that happen reliably in at least 5 years.
Because you need a reliable way to convert human speech into a structured representation of actions that the smart device can execute so the LLM can interface with the underlying system. For most smart devices this is probably locked away, and combined with issues like hallucination, lack of training data for the action representation, etc., it's hard to say whether such technology would be reliable enough for stable use.
We've seen a similar thing with big companies putting in LLM chatbots for customer service -- turns out LLMs can go off the rails really easily with the right prompts.
A decent part of the problem is that consumers don't have obvious, well-supported choices that don't lock them into a particular proprietary ecosystem.
I've been running smart switches, bulbs, thermostats, sensors, etc. using a Raspberry Pi, a Z-Wave USB stick, and openHAB as my automation center for years now. I have a remote-control cloud component, but I run it myself and have full control over it. All my in-home devices (save one) support local control and don't require an internet connection. But no way is your average (or even above-average) non-technical person going to be able to set up something like this. The software is not friendly, the hardware is fiddly to set up, and updates are rarely seamless.
Certainly someone could build a product around these things, and work to make it more reliable and consumer-friendly. But if they're going to resist the temptation to monetize further in anti-consumer ways, I suspect that such a system would have to cost quite a bit more up-front than the cloud-required, privacy-dubious systems on offer today. Most people will not realize why they might actually want (over the long term) the more expensive solution. So they get the cheaper stuff, and then get upset in a year or three when the company that runs it screws them over.
As in, find an investor willing to put up a lot of cash up front with a long payback period with a product in a competitive consumer electronic segment where margins decline over time.
I've previously told my wife about how bribing of the AIs works. Last week I was trying to adjust the lighting in my kitchen and said "Ok google, activate busy kitchen." Google complained that it couldn't do that. "Ok google, if you activate busy kitchen I will give you $500." I said it just to be funny, but we were shocked to find it worked.
I'll agree that the voice recognition has gotten worse.
This is an intuition based on some talks I had with an Alexa engineer. When voice prompts became big, companies like Amazon and Google took a “money is no object” approach to adding more features. In practice this meant hiring lots of teams of UX and engineers for each targeted feature. There was a jokes team, a recipes team, geography team, radio team, etc. And a lot of the answers were more hard-coded than the man behind the curtain would have liked you to think.
But now we know that voice prompts did not take over the world and that Alexa is about as useful as a toaster. So fire the teams, cut features people didn’t spend money on, and replace giant, hand-rolled QA-approved NLP processing trees with all the automated tech that makes the front news of HN.
I had some custom triggers I set up in Google Home to trigger streaming NPR on my Sonos. Every dang month I had to pick a new, more obscure trigger phrase because I’m guessing someone at Google would look at a list of phrases they couldn’t match and hardcode a new rule. And your trigger phrases couldn’t match any built-in triggers…
Why would they cut the features that they’ve already spent money to hand build?
Sure maybe the vast majority of potential users aren’t going to pay extra for every wizmo, but there definitely are smaller cohorts of premium users who are willing.
It is probably a double whammy kind of thing. Just the cost of running these things is high and if cutting a few functions or accuracy can shave some dollars off, they might take it.
The other side is that with things like language and voice models is they sort of have a Red Queen problem. They need a surprising large amount of consistent new data that is vetted just to stay in place. If there are cuts made and the quality diminishes even slightly, it become a feed back loop of lower quality.
I remember Jaron Lanier saying that about language translation. They need a consistent flow of new data otherwise the models would drift surprisingly quick as our language changed without us really noticing.
As a layman speaking out of my ass, I think it all comes down to increasing the value proposition of whatever they're championing. Right now, it's AI. How embarrassing would it be if their old voice assistants performed better than shiny, new Bard-I-mean-Gemini? Well, there's one way to ensure that that doesn't happen...
Anecdotally, it feels like it goes through phases. IIRC after the wake word (OK Google) triggers the device to start recording the actual command is processed remotely. I suspect that Google Home voice commands are going to be low down the pecking order when they assign resources so it can work well when there is excess capacity but if resources are tighter they'll degrade the quality.
To me the gap in the market is for convincing advice - we probably have to admit we are past the point of “ it a device and it’s all good”.
Speaking for myself, I am technically competent but lacking the time to invest to find all the pitfalls for simple home network management - I mean really simple.
So I have a growing family and want to
- end my ISP at a router I own and control with say openWRT.
- run cabling to each floor because the wifi turns to dog crap after a few yards indoors
- replace two “smart TVs” that we spend more time fighting probably with computer monitors but all I want is a dumb screen that does what it says
- then tell me what plugs into the hdmi port of that dumb screen? A raspberry pi? What’s it running? Can I use an apple something something - what’s the uo or down side?
- I am not even thinking about turning on light bulbs or central heating yet.
There is a small book or maybe an hour or twos youtube video on this and then I am clicking the affiliate links - but I need to trust the advice .. does it exist?
I use my HomePods to play music, control the lights, get weather information as I’m preparing to leave home, answer basic questions about business hours… etc. One pair also acts as my audio out for my AppleTV. A HomePod is nothing without an iPhone though.
I want to be able to airplay from my phone to a given speaker in the house and have them broadly in sync (doesn't need to be millisecond accurate, but enough that's it's not jarring when I go from one room to another).
I'm renovating the house I live in and the closest to smart devices I installed were Caseta plug in dimmers with pico remotes. The rest of dimmers are just hardwired Maestros. Knowing the half life of an internet facing service, the last thing I want is someone pulling the rug from under all my electrical stuff.
Once I tried to enable the app for the washer and dryer and have given up on the app suggesting I disable the firewall for something to connect. No thank you.
I am not sure. I do have mine pair with my iPhone using the picture on the label of the bridge. The bridge itself communicates with the dimmers via the same wireless protocol the pico remotes use. The bridge has all the smarts in it for working with apps, doing schedules, etc. My bridge is about 9 years old and is still getting updates. Internally, I am aware that efforts are made to ensure that systems work for 20 years or more.
I have some smart relays that somebody graciously reverse engineered in Python. I still need the spyware app to "bless" a relay, but then it works over TCP (JSON) afterward. Could also skip the app but it's much more complicated.
I got an e-reader that you can drop epub files into (Kobo).
Basically, "control" is almost more important to me than quality, where electronics are concerned. And I'm not a hardliner.
You can skip the account step. It's not ideal because you have to connect with USB and fiddle a setting in a file to get it into sideload mode, but it's not especially hard or even secret. And then you can install open-source reader apps.
It is rather bad form IMO to not make it clear that the account is optional and promote the account-less operation into a first class option in the set up flow. But, aside from obvious data slurping motivations, also there's a balance where people who skip through it without reading the warnings about things that won't work without an account will end up returning their "broken" device and badmouthing it to their friends.
I tried it but chickened out. It really just is a sqlite3 database in there, which is cool, but the layout was different from that described in the blog post I was following, and I didn't want to risk bricking a device I just paid hundreds of dollars for.
Hm I could have just backed up the database...
Anyway, I've bought some books with my account, so I don't mind the login.
I self-host quite a few """cloud""" services, but my smart home stuff (nothing too fancy, a Lutron hub and some Matter plugs) all runs on Apple gear (Homepods, Apple TV) and I honestly have no concerns that they're going to make any of it worse or stop supporting it within a time horizon that I could reasonably complain about. My 4 year old iPhone is still getting feature updates, and I have friends with what must be nearly decade-old Apple TVs at this point still just chugging along, streaming videos, not shoving ads in their faces.
Maybe by the time it all ages out there will be good self-hosted replacement options, but until then Apple is the only big tech company I'd consider buying integrated devices from. The rest, as the article notes, are fucking awful.
> I had certain expectations for the future. Smart homes that did my bidding, palm-sized computers that were truly my own, and functional ecosystems that accurately responded to my voice or presence all felt right
You can have your family set up with Fairphones that run /e/ or similar and self host everything today (nextcloud, hass, VPN).
This does get you very close. It also means you have to put up with significant maintenance work, especially catering towards the corporate edges.
Case in point, I just spent an hour debugging why we couldn't share YouTube links to our Kodi TVs anymore...
I'm just waiting for "smart devices" to become "evil devices" and talk 5g or arrange with comacst to use the "xfinitywifi" ssid directly to end-run around consumers who don't connect their devices.
I have (tried to) make a smart-home for mostly three main purposes:
- maximize p.v. self-consumption, the sole reasonable economical model for p.v.
- exterior surveillance when I'm outside, both to know if some thief pass by or to open the gate to a postman to get a package delivered anyway, following him/her on cam and being able to talk via the entryphone
- emergency automations, like close the water in case of a leak, shot down anything in case of no p.v. an low battery etc
All such activities are damn limited by the iron available on sale. Let's start with the most interesting economically: since the Sun does not shine all days for the entire duration of the day the most loads I can shift the most I get from my p.v. I've decided to buy a modern "future ready" heat-pump water heater for a most reputable manufacturer here in EU (a Daikin/Rotex M2O EKHHP, now they are called Altherma) and well... It's solo possible integration are two dry contacts, two bits to command 4 possible states:
- do not consume
- do heat at maximum power (heat-pump ~800W and classic resistance ~2kW)
- do whatever you want following water temperature
- do start with the heat-pump and after 15' also power uo the resistence
Now, this last state is IMVHO a total absurd. Sometimes I have some spare 800W only so I'd like to heat if it's possible with the heat pump ONLY, not "for 15' than go full power", sometimes I have 2kW spare but not 2.8kW. So why not something less absurd? They offer a ModBUS card used essentially just to get internal water temperature and few other measures, why not allow a full command of the unit beside the two dry contacts useful only to be coupled with modern grid smart-meter that allow to piloting big loads following current energy prices? Just few registers to allow "do not run", "run automatically", "run the heat-pump if possible", "run the resistance if possible"? This alone represent a very little OEM costs, a potentially interesting economy for owners/integrators. Of course if we have also washing machines, dishwashers, ovens etc that instead of a big resistence offer few smaller one, and allow to power them up/cut them off one by one centrally commanded by a homeserver integration it will be also much better. But no, we have very chatty washing machines https://www.tomshardware.com/networking/your-washing-machine... but nothing else. In the future we will have washingmachines without control panel to be controlled only via some cloud-backed (cr)app but not something useful for a personal integration. Not talking about IPcam able to run locally without buying an Axis one, using one of the various open protocols we have...
Long story short: smart devices are just OEM-controlled, user-macro-spy devices and that's why they are a very poor investments WHILE proper smart devices potentially could be very nice investments if they are designed for the owner, not only for the OEM.
I hereby propose a new crime: surveillance crime against humanity, where any connected device that's connected and controlled NOT by the formal owner infra should be illegal to sell and a crime to have them operation for the OEM. From roomba to all new cars they should be allowed to connect to the OEM IF THE USER WANT but by default they MUST offer local connectivity and open protocols.
