Apple - that famous server company.

Any other company would be cutting the price so they didn't make as much profit per unit but could shift units.

But Apple just stop producing as many units and retain the same price (which is going to be at least 50% profit, probably more like 80-90%, given Apple's history here).

This product's gonna be dead in 2 years, and they likely won't make another.

Quite.

I started tinkering with it as nothing more than a hobby, and I have some countless dozens of sensors on it now. And most of those are actually just operated in-house and don't require cloud integration at all.

I put three different SDRs on it and it captures multiple weather stations, fridge and freezer sensors, and even passing aircraft (using FlightAware). As well as the other useful integrations like what bin-day is today and what I need to put out.

I obtained for Matter smart plugs for free the other day and just out of interest I joined them - turns out HomeAssistant is a Matter server and requires no further integration or third-party to work. All kinds of switching and energy-monitoring now, and I have door sensors that I've put on my letterbox and parcel box, so I even know when I have a delivery and it triggers the cameras if I want it to.

I even added an ancient projector using "pjlink" into it, so when it's movie-time, I can dim my lights and turn on the projector and switch it to the right input just from my PC/phone.

It's a great little project - but it needs to shake the reputation of all that YAML nonsense being the only way to configure it. They really need to pull that out into the GUI entirely, because even as a programmer I was always put off by the concept of these large code-dumps on discussion forums to "fix" one particular sensor or similar. Turns out, 99% of that isn't needed and the other 1% could easily be folded into the GUI and probably will over time.

I don't even bother with voice, though, because it's just a home-sensor platform to me, I don't want it controlling anything itself and I don't want it listening. I just want alerts and a page of switches/graphs, and it's great for that.

Home Assistant runs my house and there isn't a single Zigbee device on it.

That's just "another module" in their huge libraries of integration modules.

It's "bespoke" because their own hardware that they choose to sell is not just an off-the-shelf device. However it's also just an open-source OS and software that you can put on a commodity Pi or other machine and don't need to buy their hardware or use any given protocol at all if you don't want to.

"Free rides can't last forever".

When the market is saturated, your return on each product diminishes.

But you're still getting free electricity!

Subsidies for solar were always going to be short-lived and rather pointless for the grid beyond a point - you want them to pay you for electricity they literally cannot use?

It's why when I started my little amateur installation at home, I didn't ever care about "feeding back" to the grid. I'd really rather not be tied to the grid at all, even for feeding back. It requires a larger investment - not least the cost of the connection - and it's reliant on subsidies staying around forever and never changing (which just isn't realistic). I'm sure if you got in on day one when panels were ridiculously expensive, and you had the capital to put into it, that the cost of doing so was profitable over many years. But once those subsidies started to dry up and the hardware became commodity, there was never any question of getting that free ride.

Instead I buy panels at my own pace, and increasingly move circuits off onto them, reducing my bill, and one day being "self-sufficient". The excess electricity generated from there? I'll use it for something else (hell, why not just a little Bitcoin farm or similar to acts as a very-easily-controllable on-demand power absorber? Or a heating water tank that you can just dump excess electricity into?), or let it go to waste. Only when I've not switched back to grid power for a few years would I consider severing the connection, but I would never bother to apply for the grants / subsidies / connection as the return is too uncertain because of things like this.

If you are considering it an income of your own, you have to consider it the same as a business, which means you have to take into account that the market changes and may not be profitable in the future.

But you're still going to pay off those assets and get free electricity, no matter what.

I have nothing worth hiding, but I have an even simpler method:

Don't put incriminating things on your phone.
Don't use biometrics AT ALL. Literally just turn them off.

My bank app tries to remind me every 6 months or so and I just dismiss it. Google also seems to think that I can "pay" with a biometric as an option whenever I buy an app or book on their apps... which is interesting because I've literally never given them one. Selecting the option wants me to enroll a fingerprint using my Google password. Nope.

Paypal tries the same - keeps pretending that I can "login" with my fingerprint as an option despite the fact that I have no stored fingerprint on my phone or Paypal account.

The only place I've ever been required to give a biometric was when applying for a passport. That's it. I consider that a reasonable extra metric to have on such a document to help prevent forgery or misuse of my identity if it gets stolen. Nothing else has or could verify my biometrics whatsoever, certainly nothing consumer/commercial.

(I think that too many companies are just too tricky in trying to obtain your details. Whatspp, for example, demanded I furnish them with a ton of information to "confirm" who I was when I asked them to remove data... and 50% of the information requested has LITERALLY never been given to Whatsapp or any partner company... it was just a data-hoarding exercise. "Give us your home address if you want us to remove your data from our servers"... er... no. You have absolutely nothing to confirm that against, and if you do we have a FAR LARGER problem almost immediately).

Disable biometrics, they are not required. Stop faffing around using them as a convenience method, because that's all they are. They are not, and cannot ever be, secure because you're giving them out to every device that tries to read them.

Not storing incriminating evidence on a cloud-controlled mass-market consumer piece of technology is probably a good idea too if you want to avoid such things coming to light.

At the point of arrest, are you suggesting that the police couldn't open a locked door that they suspect has evidence behind it?

That they'd have to arrest the guy, take him away, go to court, get a warrant, in order to open the boiler cupboard?

No, it doesn't work like that. Reasonable suspicion of a crime has rules too, and those allow such actions. It's always "unreasonable" search that's the problem.

Otherwise criminals would put cheap tiny padlocks on everything they own, including their phone, and make you get a court order to open every single one.

Talk about burdening the court systems with nonsense.

"We've arrested the suspected child kidnapper, but we can't possibly open the door to his cellar because it has a small luggage padlock on it, so we have to wait for the court order to come through, and hopefully it does that in 24 hours or we won't be able to charge him and will have to let him go."

If your enterprise depends on a product that can't parse a textfile correctly without appropriate and simplistic sanity-checking, you absolutely and desperately need a new product for your enterprise.

And if that product says "Oh, we can't do that, because it's undocumented and the format could change at any time, so no warranty for that..." take that as a hint.

I would guess that absolutely nobody is paying the kernel team to solve their boo-boos with their third-party, out-of-tree, unnecessary KConfig parser for enterprises that they're charging a fortune for. It's on them to fix it if their parser is so immature that it can't handle a space in an often manually-edited config file.

If it's undocumented, third-party integrators should be asking for such and/or KNOWING that their integration could break at any time and that they should therefore do a lot more sanity-checking.

It's actually incredibly likely that we're NOT.

But the problem is that simple physics gets in the way and the chances of two civilisations existing at the same time, within communication distance of ANY kind, and who notice each other and can do anything about it (beyond having conversations with 4000-year round-trip times), are infinitesimally small, even with a million such civilisations.

Basically, the limiting factor here is the speed of light - and if that's literally the limit of the universe, every civilisation that exists will basically be forever isolated from all others, just by sheer probability. It doesn't matter how advanced they get, how fast they spread through their galaxy, how many millions of years they last.... the chances are they won't meet another, or even catch a glimpse of evidence of their existence.

It's far more likely we're one of countless civilisations, even in our own galaxy, but almost certainly in the countless trillions of galaxies we can see, but that we'll never actually know that. The maths tells us so.

And if someone can "break the speed of light" (without tricks like holes in space, etc. but actually break the speed of light), they could probably also go spend all of eternity locating every civilisation that ever existed anywhere from the point they discover that, and basically visit them at any "time" in that civilisation that they desire. Would they choose 2024 Earth and humans from a universe of possibilities? Almost certainly not.

We're not alone, we're not the most advanced life. But we will likely be entirely unable to provide any evidence of that for the complete window of our entire existence. That's, by far, the most likely scenario for every civilisation in the entire universe.

Assange added 8 years of his own to his sentence.

In reality, he's only done:

April 2019: ejected from the embassy
May 2019: 50 weeks prison for skipping bail
May 2020: everything else starts taking effect

So not even 4 years yet - for a complex, public, international extradition case. BBC News has a story of a rape conviction that took 6 years to get to court only this morning.

That's including a denial of extradition on health grounds, an appeal by the US and escalation to the High Court.

And look... it's about to happen, as the final chance at a successful appeal started in February back at the High Court.

The fact that Asssange himself wasted TWICE AS MUCH TIME AGAIN just to go on the run is his own fault, and self-imprisonment contributed greatly to the mental decline that was later successfully used as an argument in front of a judge to try to avoid extradition and only overruled on appeal.

I learned both early Visual Basic and Java through the sheer fact of reading a) the Visual Basic Language Reference and b) an O'Reilly Java book.

Both had the complete language described, full descriptions of every command, "library", etc. even listing of all the constants.

It seems a really dumb thing, because I'm far from a hard-copy person, but having the entire language in a single book was amazing for reference, and also for flicking through and saying "Ah, that could be useful!". I'd been programming in Z80 assembly since I was a kid years before, and the stand-out moments from years upon years of coding were when I had a complete reference in one place that I could just read and absorb before touching a new language, that I could use as a definitive reference for completeness, and that I could look up syntax and glean ideas from just by browsing it.

Even something like Ralf Brown's Interrupt List was a defining moment - I can remember the first time I got that (downloaded via modem 100 miles away and put onto a floppy disk to bring home) and it was just so useful.

Languages now seem to be in a constant state of flux. Even with code that I'm almost entirely hands-off with, things like PHP change so rapidly that there's almost a guarantee that the code you're managing will break in the next version and need tweaking. C++ has so many iterations it's laughable.

I miss the days when you could describe an entire language in one static reference book (online or not), when a technical writer had spent months of their life describing every available function in detail, it was from an official source you could trust, and it was actually *enjoyable* (with a certain nerdish masochism) to sit and read the documentation.

If only something like that existed for the Windows APIs still, OpenSSL, etc. I would have saved years of my life.

"It was also nice when everyone confidently knew what a man and a woman was."

Know a redhead? Then you know an intersex person, they just haven't told you about it. And with your attitude, I know exactly why.

And were entirely wrong a minimum of 1% of the time, because naturally-occuring intersex variations appear in that proportion in populations the world over. Roughly the same percentage as redheads, diabetics, etc. It's actually a large factor in infertility, because "X" and "Y" are not the cut-and-dry genetics people believe them to be.

And that's before you even get into those who chose to live their lives differently to their assigned biological sex, which has been happening for countless thousands of years (literally women living men's lives and only ever discovered on their deathbed, etc.)

Modern times haven't changed anything that used to happen at all. My workplace (schools) always have stories of headteachers from the past coming out as transexual, cross-dressers, etc. going back literally hundreds of years. It was often kept quiet but it was always there - and kids were not playing out in nice neighbourhoods without incident either. My dad tells a story where his dad wouldn't let him deliver certain papers to certain houses on his own. Grandad was a policeman through the war and knew where certain types of criminal lived, even if nobody else did.

These types of people (and they're still people, always remember, and literally just a type you disapprove of) have always been there. You just didn't know. Same way that you didn't know (and still don't) someone's political affiliation unless they told you. The reason you "never saw them" was because they never told you... and there's a reason for that.

(For reference, not that you deserve to know: I'm biologically and outwardly male, heterosexual).

Scan & Go actually works really well.

You just have to remove the side cases where it doesn't - like loose products as you say. But loose products are, by far, not essential to anyone's shopping. I don't buy cornflakes by the flake or by the gram, there's no need to buy pears individually by weight either.

Since those scan & go technologies come in, however, my shopping is now utterly predictable, far more accurate, less impulse buys, and I'm in and out of the store in half the time. Mainly, it has to be said, because I don't have to unpack my goods onto a conveyor, have a spotty teenager scan them by hand, then repack them back into the exact same place I just unpacked them from. That, alone, makes it worth it for me.

Being able to see the exact price while I'm deciding on the product is another bonus. Whether it's labelled or not, I can see the price and know immediately whether I've picked up something that's part of a deal or not... which I imagine quite a bit of profit was made by companies before by "mistakes" on the human's part. I first started using it when I had no money, and it was great to ensure I was inside a strict budget. I could scroll back and remove that product from 4 aisles ago that would give me just enough money back to stay under my limit, as well as having a running total in front of my face at all times. I'm guessing that doesn't help profits either!

And I simply don't bother with loose goods. I can be in and out of the store in under an hour with an entire month's shopping in one hit, for cheaper than I ever could. I don't even bother to become a member of their special loyalty programmes, because increasingly they were becoming paid-for and because of my single-trip mentality, they actually weren't worth it after a while. I was paying, say, GBP9 a month, in order to get 10% off two shops. So if I spend GBP90 a month, I was profiting. But the programmes got more expensive, applied to less products, and increasingly the terminals wouldn't easily read my codes, discounts, vouchers, etc. and it became a nuisance. So I just cancelled it and actually saved money by doing so because occasionally I go to a different shop instead - their loyalty programme basically forced me to become "disloyal".

Manual checkouts ALWAYS queue. Self-checkouts there's a small queue but it moves fast as there are always more self-checkouts than manual staff. Scan & go never has any queue at all for me. Repeatedly, month after month, I just literally scan-and-go. Along with the packing/unpacking it cuts my shopping time in half, and even the time lost to typing in barcodes, etc. for vouchers and discounts is such a factor still!

Self-checkout is a nonsense outside of tiny few-item purchases - it's for people buying lunch in a store while they are out of the office, that's what it's for, and it works fine for that.

Scan & go is great, but it's for buying a certain subset of things (i.e. no loose items, nothing that requires verification, etc.) and keeping an eye on the bulk.

But manual-checkout... I can't find a use pattern for any more. It's just slower, more expensive and less useful every time I'm forced to use it. I'd rather avoid it and do those bits some other way and somewhere else. If my local supermarket was just scan-and-go, or like these Amazon stores, I'd actually be happier. Even delivery isn't as efficient because they insist on making constant substitutions (what happened to accurate stock control?), and then I have spend ages making the shopping list, be in at a given time (or "click and collect", which is actually less efficient than just scan-and-go), pay a fee for delivery, deal with trying to not delay the driver by getting everything into the house quickly, etc.

Of all the options I want in my future when I get old and infirm and want to just take my time and buy stuff I want without extra nonsense (or unpacking repeatedly), scan&go and home delivery are the options I want, and home delivery will still cost money.