Alexa and Google are always hooked into Google's stuff, whether there's some at least partial local control still available in an outage or not.

I'd say local-first is *fairly* unique. Yes Homekit/Matter devices *can* be controlled locally in a peer-to-peer manner right from handsets, but Thread radios are fairly rare and I don't know if any non-apple handsets support directly talking to those devices without an intermediary.

If you don't have Apple devices, then HomeKit is a mixed bag, as sometimes the onboarding is only possible with an iPhone.

Now when you want to take it to be internet accessible, Home Assistant is a pretty rare software for easily supporting that *without* going through any cloud provider (get a dynamic dns and let's encrypt going, and Home Assistant plugins exist to automate that including renewal for those that don't want to understand how to do that themselves.

Now who is doing something "tech bro" that "isn't suitable for regular users"?

If the vendor's device doesn't support standards based management then I will just ignore them if at all possible. HomeAssistant can update firmware in any of my devices in my house.

The devices don't have a gateway set so they can't 'phone home' anywhere and if that's a deal breaker for them then it's a deal breaker for me.

I do know that there are devices consistent with 'off-cloud' usage, whether Home Assistant is at all responsible I don't know and don't really care.

To address the 'not for regular folks', they made a 'home assistant green' which is fairly decent at being an accessible, self-maintaining package. One of my relatives had a Nest thermostat that Google made stop working, and so I gave them an alternative together with Home Assistant green and they've been pretty happy.

"Only for tech bros" would be nothing but APIs and you have to assemble something yourself. Home Assistant has some tech-bro friendly deployment options, but does offer something akin to the typical cloud connected consumer electronic fare.

The devices generally do not connect to Home Assistant as a server, the Home Assistant connects to them as a client. The devices are generally oblivious about Home Assistant and it's nature.

I have z-wave thermostats. They have no idea what internet even is. They presumed they would be sold into some partner's hub ecosystem, but as a consequence Home Assistant can talk to it direct.

I attached an open firmware based controller to my garage door opener. The garage door opener doesn't know what networking is, and even the open source controller is oblivious to home assistant, just providing a general, locally accessible HTTP api. Home assistant connects to it.

If you are careful, you can generally find networkable components that do not expect to connect to any server, but can be connected to. Matter over Thread is *generally* a safe bet the device in question is friendly to local usage.

However, a lot of devices have firmware hard coded to connect only to their suppliers internet presence. Without an account you can't control them. Sometimes they start charging a subscription. Sometimes they discontinue allowing a device to connect and operate, suggesting you buy the new model after a couple of years. Meanwhile their 'cloud' doesn't add anything that you couldn't have added yourself. Get a free domain and a let's encrypt certificate and you can connect to your house from anywhere, if you want. Or keep it closed off to anything outside your house. Or 'shadow' select stuff into remote access while keeping some things local.

... of a thing to begin with.

It came to fame when some Java guys finally discovered convention over configuration, built yet another web framework around it and bedazzled the world with a 15 minute presentation of Ruby on Rails. The marketing of the ruby on rails FOSS project was the true genius behind all the hype. However, Ruby itself was still struggling with basics such as utf 8 and other details, so people stuck with php, Python or whatever else they were using at the time.

Rails never really caught on in a larger scale. If it had, Ruby would be a thing today. I think it's safe to say that TypeScript has taken its place.

https://volla.online/

I'd say the similarities in name aren't by accident. I wish Jolla good luck they had a nice product but 400 euros for a crow funded phone from a company know for it's volatility probably doesn't inspire confidence in long term viability. Just saying.

Yes, the real reason why netflix used to have a fantastic catalog at low cost was because at the time, the rights holders didn't take Internet streaming seriously and so cheap deal to Netflix was a low risk easy bit of free money.

Then they took it seriously, didn't get the deals from Netflix they thought they should be able to get and started making their own streaming services instead. Probably the first sign of things was when Starz demanded to be a "premium channel" on Netflix. Frankly if Netflix has accepted that arrangement, they might have been the defacto broker of streaming services in one app, though the user experience suffers, but it suffered anyway.

You might call it a tad goofy to turn a somewhat b-movie "hero" into a high quality bronze statue and place it in public display, but this was entirely a private initiative and it's standing in a commercial property and the owners where in on the fun. That's how stuff like this (and weird woke projects) should be done.

I get the joke - RoboCop was also a commentary on the derelict state of Detroit - and would pay the statue a visit.

I suppose I wasn't clear, when I said they are dutifully generating code, I mean they *are* using the AI tools. So the leadership is left with the possibilities that either AI isn't fit for the task of suddenly halving their headcount without any transition plan or the employees are to blame, and so they are deciding the employees are to blame.

The leadership cut entire teams and then just assigned their projects to the other half who had never seen the codebase, never used or talked to users of those projects, no knowledge associated with meetings and emails and instant messages, only tickets and a codebase to go on. AI was the stated answer as to why not only could they double the workload, they could AI-away the traditional need for typical transition efforts.

The trend even in theory doesn't seem to keep pace with the depreciation on the assets.

My subjective experience is opposite, after an initial rapid improvement in LLM behavior, the subjective experience has plateaued. Doing a better job with getting the right stuff into context without having to manually stuff it with certain tools, and that counts for a lot, but given the same context the outputs are about as unreliable as they have been, including gemini 3. If it generates too much code, then it's more trouble to fix that code than it's worth. Digestable snippets are useful... sometimes to save on tedium, but the investment seems to expect to just replace people, and it's not going to be there from what I see.

And the marketing has been obnoxious, pretty much going back to GPT3 I keep seeing some advocates saying "It can write your code for you!" followed with the next iteration of "well, admittedly, it couldn't before, but *now* it can!" GPT4, Claude, Gemini, I keep seeing the claims and then dismissing the previous claim to say "this time it's true!". This is starting to really impact companies, I know of a company that laid off over half their developers without any warning or prep, with execs telling the remaining people "you can just use AI to maintain and improve the code, it's going to be no incremental burden". They seem very disappointed that a bunch of things have stopped happening and have been saying the devs must be some sort of luddites refusing to use AI to just do the work, even as they are dutifully generating code.

Historically speaking, that last part is frequently elusive. Lots of crashing and burning of once "sure thing" tech companies. Some have survived, but it's hardly an assumption that can be made.

Reminds me of when an executive left our company and higher ups were rushing to assure us that we shouldn't be too worried and don't let this hurt morale while mostly we either didn't care or were kind of glad to see the idiot go. Meanwhile the execs speaking would get obviously angry at the guy for betraying them and leaving.

It was clear that day that the executives actually think we give a crap about any one of them.

Not knowing anything at all about Apple and Dye and Lemay, the story seems depressingly familiar and totally believable based on my experience with big companies.

Someone useless occupies a high position because he convinces peers he is somehow insightful, everyone hates him for his crappy 'leadership', his departure pisses off the leadership team so much that all his allies are dead to them.... Yep, all of this absolutely looks like things I've seen at other companies...

Nope, they are going the other way. There's a proposal to dramatically increase residential power rates, in part to fund the 'increased demand due to datacenters". They want residents to pay for stuff instead of making those poor, cash-strapped AI companies have to pay for what they are inflicting...