I suspect that your conclusion is correct, because we see that in several industries and branches.

It looks like tech made in the 60s and 70s outlasts everything and anything unless it's abused. It's inefficient, loud and wasteful, but it doesn't break.

It could be survivorship bias, though, and I am not sure how to exclude that for personal observations.

But washing machines, fridges, cars, servers, laptops and a lot of other mass produced tech items seem to be lasting shorter and shorter, and they seem to fail closer and closer to the end of the warranty period, in pretty consistent and predictable ways. Often, the right-after-warranty-failures seem to cluster around only one or two components in every make and model, and these components often share several properties across devices and even classes of devices:

The "failure-prone" components of modern machines are always
a) easily identifiable,
b) essential for the functioning of that device,
c) must be replaced entirely instead of repaired
d) impossible to replace with DIY
e) easy to replace for experts, but with a very predictable, very time-consuming operation
f) has a predictable and linear wear and consumption behavior
g) could EASILY be made ten times bigger, more resilient, more robust etc. by the manufacturer without costing more than a few cents in production or compromising the weight, cost and performance of that machine
h) but is not ever made beefier, even in subsequent generations

In short: every consumer- and SOHO-used machine designed after around 2000-2020 will have a component that will break shortly after warranty, the warranty will exclude "high-intensity use" that could make this component fail earlier, the component is easily identifiable, repair shops will know this pretty quickly and they know that once it breaks, repairing of the device is uneconomical for the layman, but can be acquired by the experts to refurbish in their spare time, so the experts can trot them out again. And the manufacturer will never improve that component to make the device last longer. If the failure-prone component can be easily repaired by experts or replaced with a more robust part or bought from AliExpress for pennies, the failure-prone component will be fortified with something that DRM, digital signatures, DMCA, patents can protect OR it will be entirely re-engineered in subsequent generations to become part of a module or assembly that is.

Since China produces most of the world's steel - and mass produced steel is cheaper than any other material - it will be steel.

We don't know the quality of that alloy, though. "Steel" can be a large number of slightly different materials.

One could have made a custom tool from a very specific well-known material that has been analyzed and documented exceptionally well beforehand. With that, it'd be possible to exclude this specific material from the analysis and by weighing it very precisely before and after use, we would even know exactly how many micro or nanograms of the tool material should be there.

And I guess they did just that, but probably not to a spec that's strong enough or large enough to produce the forces needed to open the damn thing.

And despite not being a scientist, I do know that stainless steel, or any steel in general, is highly UNsuitable to operations in unknown substances that could potentially be corrosive or flammable or magnetic etc.

Something that contains any amount of iron or nickel is very difficult to not be at least slightly magnetic, and if the sample contains magnetic dust, it could be attracted or disturbed by the magnetism of the tool. Steel is always a mix of several elements, and probably has iron and carbon in it, which is always a target for oxidation and thus has a higher chance of one of the elements starting a chemical reaction or becoming degraded by whatever is in the sample. Iron and steel have properties highly dependent on their internal molecular structure, which makes the entire thing (slightly) unpredictable or at least it's expensive to have it be very uniform in all its properties. And above all, steel is hard and throws sparks when impacted, throwing minuscule amounts of burning metal into the sample or its surroundings that can cause whatever chemical reaction with it, even explosions or fire, if the material is gaseous or finely powdered.

Anyone who works with gaseous substances or in flammable / explosive environments will never touch a steel or iron tool. All their tools are made of special beryllium-copper or aluminum-bronze alloys that are far less brittle and will not produce sparks or tiny fragments on impact, but deform or dent instead. These tools don't last nearly as long for these reasons.

Maybe they had a very good reason for using steel that we don't know or the alternative materials for tools had bigger drawbacks.

We bought several DL380 and DL360 generations over the years. They were all designed to last at least 5 years, and a new generation came out every 3 years. In reality, the older models lasted for many many years beyond that, the newer ones don't, and so, unless you threw them all out after their 5 year warranty period (like a good IT operation should, but corporate reality often is much different, as we all know), you'd see all the DL360 generations failing very close to one another.

Every new generation lasts a year less than the previous, so their failure dates are much closer together than they should be.

If you bought the newest server gen every 3 years, you should see them fail after N years, but still 3 years apart, right? But you don't. The older gens die after 5+5 years, the one after that for 5+3 years, the one after that 5+0 years and the latest is pretty much hit or miss if they survive the 5 years warranty without much replacement going on.

Maybe my interpretation is subject to a hefty statistical or psychological bias that I overlooked, but it seems like reliability is going down for all appliances and equipment across the board in all industries. We have almost-brand-new (2 years) commercial airliners from leading manufacturers dropping doors mid-flight. New washing machines failing 2 months after warranty ends. 40.000 USD cars requiring 35.000 USD battery replacements.

If this continues, in a few years, new tech will never outlast their warranty periods.

It is not "not wanting to develop a new single aisle jet", it is "not being able to even if they tried really hard".

Nobody is say why, and I don't either, but please look at how many companies still exist that make the most reliable aircraft to date. We had several companies in the 70s and 80s that did, but many went under after a downturn, a crash or because their aircraft weren't competitive anymore. This is continuing. Our current restrictions for reliability, regulations, economy, ecology and competency have become so dire that the tiny overlap of all these goals is rapidly shrinking.

People only accept the most reliable aircraft of their times. Of course. So if anyone botches reliability, as Boeing does now, their market share suffers tremendously, and rightly so. Some aircraft models remain highly reliable and that's what Airbus can currently sell.

However, regulations are ever increasing, because an insane number of bureaucrats in the US and EU depend on more and more regulations. They cannot allow that regulations ever stop tightening, so regulations are never good enough. In a time where advancements in many established fields of technology have slowed down, regulations have sped up to be faster and faster.

A car sold new in 1970 would probable be legal to sell new in 1980, and maybe even in 1990 with slight corrections for safety and emissions. A car sold new in 2020 cannot be sold new in 2024, because the regulations are turning faster than any engine ever could now. We will have yearly regulation changes that make a car from let's say 2026 absolutely 1000% HARAM CHOPYOURHEADOFF illegal to sell in 2027. And new cars after 2028 will be illegal at all, ever, for all eternity, or until this system finally collapses on itself.

For aircraft, this is no different, because the public has been convinced that these are huge environmental blunders to exist and operate at all, and since no one has an aircraft in their driveway, there's little backlash in pork barrel and regime politics that "make aircraft safer and more environmentally friendly", which, according to published documents by WEF itself is "zero air travel for Plebs beyond 2030, at all, ever". (and it's still a conspiracy theory to point out to the official documents of this organization that operates as some de-facto global government for at least the Western nations.)

And then there's competency. You remember how Boeing pledged to get their workforce to become highly diverse? https://www.reuters.com/articl...

That was in 2020. It's 2024 now and aircraft younger than 2-3 years have insane quality control problems that have nothing to do with any new technology or development, but simple things as "consistently tightening all bolts on the aircraft to their correct torque and correctly checking it twice and documenting the check somewhere and reviewing the bolt torque setting process itself, if too many quality checks light up"

8K video is a niche and extreme-enthusiast use case. Very few things require 8K that cannot be done with 4K or lower, like 360 degree or stereoscopic videos. And a month of video stored?

There's possible use cases for that, but it's not many.

The accident in Tokyo this January was an Airbus A350 that hit a Dash-8 on the runway during landing. From what we already know, the A350 had all the rights to land there and performed a very regular landing until the point that it hit the Dash-8 that suddenly entered the runway at the C5 mark. The Dash-8 was not given clearance to enter the runway and was not cleared for takeoff.

We still don't know why the Dash-8 crossed the runway threshold without being cleared or instructed to do so, but the cause of the accident lies in either a miscommunication between ground control and the aircraft, malfuctioning runway signs and threshold illumination or error on the side of the Dash-8 pilot.

We are pretty sure there were no technical defects for the A350 and probably no errors on the side of the A350 pilot.

So far, it seems that all A350 passengers survived the crash and the intense fire. That alone has prompted several agencies to investigate how this was possible, because given the circumstances of this accident, it is a miracle that all survived. And transport security agencies are trying to know whether the all-carbon structure of the A350 is the reason why, as it seemed to burn slower than older aluminum composite airframes.

The occupants of the Dash-8 all died, except for the pilot and since they were all on a relief mission to Ishikawa prefecture where the extremely strong earthquake hit the day before, it is all the more tragic. As the Dash-8 pilot is very likely to bear some or all responsibility for the accident, and him coming from an official Japanese agency, you can probably guess what he is going to do, as soon as the investigation comes to a close.

Re-using passwords is gross negligence. Using any dictionary word or shorter than 10 characters is at least simple negligence.

Using weak security to protect the access to one's own DNA data is absolutely idiotic. It doesn't matter what the service enforces the passwords to be - the person using the service has to make reasonable efforts to maintain their own security. Only then can the providers be sued.

"Oh McDonald's didn't prevent or warn me from using the five-second-rule for a chicken nugget dropped on their disgusting floor. Let's sue them after getting an e coli infection!"

My car doesn't enforce me driving the speed limit in a school zone. I have to do that.

Long, full-random passwords stored in a trusted password manager are the way to go, and that manager has to have two-factor security to protect itself. Nothing else is feasible anymore. Really. Stop memorizing passwords. Memorize the for the password manager AND the FIDO passkey used to unlock it and the process to access, backup and sync all other passwords stored there. And use a passkey for all important accounts, especially the mailbox used for all the recovery processes.

A regular closet can hold several thousand DVD or BluRay cases. If you use an inexpensive disk wallet case that store just the discs themselves, you can store 1.000 discs in a 30x30x30cm cube (for people without international passports: this is 12"x12"x12").

Your apartment could have that space, if you really had to store 1.000 movies.

If you took the time to rip and convert (or download...) them all, it would be around 10 GB per movie (2 hours, h.265 4K video + one 7 channel audio track). 1000 movies would be 10 TB. 1 HDD. But since it would be risky to store the results of all that hard work and time to convert (or download) on just 1 drive that can fail, you'd need at least the most basic RAID-1 setup for that, with 2x12TB drives. Including a the power brick and little space left for cooling, this will also take up about 30x30x30cm in your room.

You will find that amount of space in your dwelling if you value movies enough.

Streaming is nice to have, but it has become horribly unreliable over the last year or so, as titles are constantly removed, delisted or shuffled to different streaming providers. Some titles are not listed anywhere. ("House MD" anyone for example? Without pirating or paying 8 dollars per episode for nothing but a digital license that can be nulled any time the provider chooses to? No?!)

If you don't have the physical thing in your own possession under your full control, you don't have it.

Digital licenses and streaming services are noting but a pinky promise by them to not remove it next month. And it happened often enough that one finally has the free time to watch a movie to relax, children are at grandma's house, popcorn is ready, blinds lowered and then for some reason Netflix removed this and that and that movie now because a new month started and the money grubbing lawyers behind all this have decided that this title is better moved to another company or it is moved up to a pay-per-episode tier, because regular streaming wasn't profitable enough. And that's when you remember why you built that 18TB NAS for purposes absolutely not related to storing a lot of movies.

Changing a previously well-defined category "MOTHER" from meaning "human with xx chromosomes and a set of more-or-less functional female organs" into a category that also includes "humans with xy chromosomes and no uterus, but wearing skirts and makeup" and "humans with xx chromosomes that have a uterus, but take hormones that healthy xx-chromosomed bodies never produce" and all other sorts of genetic and hormonal status, then the science concerning that becomes LESS precise. It is a generalization to the concept, introduced for political reasons, that either dilute the precision and applicability of the research if they're actually applied and not just used as labels, or dilute scientific precision and objectivity if they're not applied for the entire research, but only labeled.

Science is the formation of categories and assignment of real-world observations - where everything is on a continous spectrum - into these categories. Reality presenting every observable variable as a "spectrum" does not preclude the definition of categories to make sense of it. Defining a set of wavelengths into "Blue" and "Yellow" does not make light not be a spectrum anymore, but it enables human communication to happen about it. And science is the checking, evaluating, re-evaluation and, if needed, re-definition of categories and concepts. And more often than not, sub-division of existing concepts and categories into finer-grained ones with less categorization error.

This categorization error is what statistically-oriented science tries to minimize, without increasing the number of categories too much. Minimize the number of categories while also minimizing the average, median and maximum categorization errors is what makes statistical research "good", because the goal is to make testable predictions from these categories later and that requires a huge reduction from the number of observations to the number of categories. Creating 1 million categories to describe 10 million patients is idiotic, because then there can be no predictions made for any of them, as then there'd only be 10 member of each category and not nearly enough data to make any generalizations about the course of whatever process it is you're trying to describe. Putting all 10 million patients into the same category even if they have huge differences among them is also idiotic, because then generalizations will be too broad to be useful, or more importantly, may be too "wrong" for too many individuals, so a medical treatment developed from it is too risky.

Including men, women, transmen, transwomen, intersex and even more diversely sexualized people in a study relating to pregnancy could be useful, if the actual differences between those categories are the focus of that research or the research is taking care that these categories are not mixed and erased during the analysis. Writing the paper like these categories are all the same or ignoring the differences, despite their hormonal status differing wildly, astronomically, between them - WHILE TALKING ABOUT HORMONES, is complete lunacy.

It is at least as idiotic as talking about the harmful effects of UV-C rays of a very specific wavelenghts on skin cells and then linguistically including yellow and red and blue into the paper and pretending that all "visible light" is identical as far as skin reactions are concerned.

If that is "science", please never produce medicine or safety-related products from the results.

Anecdote: I have it enabled and stored forever. And I will try to make sure my location history survives the announced change.

I have had many instances where a precise recollection of my whereabouts was incredibly helpful to me and I would not want to miss it. What are some examples I can think of right away:

Knowing when my last checkup at this particular doctor was. Their system purged all old data and browsing through the paper folders would have taken them forever, and so they could pinpoint the time frame they have to search. Reduced time from hours to less than a minute.

Knowing when my last visit to this particular shop was. Instantly knowing when the purchase date was of an item I bought there. Saved a lot of time searching through all the receipts, becaue I instantly knew it was beyond the warranty period and bothering with receipts and returns would be pretty much pointless. Saved an hour or more of searching for the receipt and then be disappointed, or worse, not finding it, driving to the store asking them to look up the copy in their system, only to then find out it's out of warranty.

Organizing photo collections from vacations and road trips. EXIF data shows date and time, but not GPS (with the camera, at least). With Google timeline, I can easily and quickly find out where each photo was taken.

Remembering vacations and roadtrips in more detail. Whan it's time to revisit vacation photos, the timeline can show much more detail of where I went that day and that helps me remembering more of what I experienced there, especially for places or circumstances when photos to remember them couldn't, shouldn't be made or simply weren't, because it didn't seem to be relevant at the time.

Filling out time sheets for work. Timeline always knows when I was in the office, at a client's location, how I traveled there, how far and when, for filing expenses and reports.

Retracing my steps looking for a lost item. Arrived somewhere, noticed something was missing and now had a good chance of retrieving it, because I could then walk back the exact route of the day and either remembering where I might have misplaced it or knowing where to look for.

Revisiting tourist sights from vacations years ago or giving recommendations to friends about them. Photos rarely help with that, especially when touring large cities in East Asia, because even if there's text in the photo, it's difficult to tell others about the name of the place and the area is densely packed with places, so approximate locations don't help.

It is some form of an externalized brain that remembers times and places that I could never, ever hope to remember at all, and certainly not with all the details. People with photographic memory might not need any of it, but my linear time and place memory is so bad, that it helps me a lot. Yeah I could write a daily journal instead. I don't have the time, energy and discipline to do that. I am thankful that technology can do that for me.

That Google wants to get rid of the central unencrypted storage of all this location data is hardly surprising. Firstly, because it is probably compute-intensive to maintain. Millions of users randomly accessing their location data piecemeal will be a resource drain, as it will be in a database form and look-up and processing will definitely be difficult, especially with Terabytes of it accumulating and being otherwise pretty cold data most of the time. Secondly, because being in control of user data with a huge impact on privacy and legal matters, is probably a liability for Google. Every law enforcement agency can contact them all the time and try to extract the data from them for whatever investigation they have. If Google has the data, they will be compelled to hand it over. If they don't, they're punished. If they make a mistake with that, they will be sued. If they don't have it anymore, or it's all end-to-end encrypted, they can more easily refuse subpoenas. And thirdly, because even if they now don't count the storage towards the users' contingents, they will of course do so in the future. Google always does that, eventually counting everything towards the contingent. They don't do it right away merely to lessen the backlash, nothing else. And it is reasonable to do so, because every byte stored on their servers must be paid for by someone, eventually. If my location history is X amount of data, of course it's my responsibility to pay for it.

I hate seeing people pretending there is no "normality", no "default", no "usually" about anything and rephrase their language pretending to include the rarest of special cases.

Even if something is true for 99.9% of people, we are forced to pretend it isn't, or that the other 0.001% have enough importance to change every sentence and phrase in every document and every speech. As if human activity, values, speech, society, processes etc would be wrong or even evil, if they did not accomodate for every single particularity in existence, even if it was less than one in a million.

We should allow things, people, concepts to be considered "normal" again.

Passwords are outdated and insecure, per se, all of them, all of the time.

Do not use passwords for anything accessible from the Internet. I repeat: Do not use passwords for anything ANY THING connected to the Internet.

Buy one dedicated, USB and NFC capable FIDO2 device, a one-time 50 USD investment, put it on your keychain to always carry it with you on person, register it for your primary identity provider(s) (usually the email service, where all the password-forgotten emails would go to) and find a good place to store the recovery codes offline and secure, and you're done. If you're using Apple, you need TWO FIDO2 devices, because Apple requires you to register a backup, they don't want recovery codes. If you're smart, don't use a politically active company for the ID provider role that would kick you out of their ecosystem for a youtube comment critical of one of their increasingly numerous "sacred cow" subjects. Repeat: use an ID provider that does not care about your political stance. Big tech does, and no matter your stance today, there's a chance you'll one day be on the wrong side of their censorship. If you choose Google or Facebook, be mindful that they have the keys to all your services and their TOS allow them to lock you out for ONE comment anywhere that they don't like. If you want, register a new empty account with one of them and use it ONLY for the purpose of logging in with third parties and nothing else, ever. That could be safe(er) from censorship.

And then change all other services to log in with that identity provider. Not using passwords, but OAUTH2 "login with google" "login with facebook" etc., which will pull in their security level - FIDO2 and / or other passwordless - to the other services. That way, you have no other passwords for anything else, and everything has the strongest possible security. All the other services won't, can't enable FIDO or other advanced login method, so that is the easiest and fastest way to get rid of passwords and enable a strong MFA for as many accounts as possible.

Intel is more woke. Remember half their stuff is named after landmarks in Israel.

A company that is led by members of the tribe cannot go bankrupt. It will always get more credit from the other tribesmen. They create money out of thin air via the Federal Reserve and fractional banking of everyone's money. And they will cross-finance sell / buy parts of "their" companies (companies led or owned by members) whenever they struggle to one another, so the newly created money is channeled into the parent companies to keep them afloat.

As long as money is being created out of thin air, with no thought to actual reserves (if the reserves at Fort Knox are even there anymore), companies under control of the tribe will always survive AND outperform companies NOT under control of them, until they are bought and integrated into this group.

Look how 90% of the media market belongs to 6 companies. The same with food and pharma. Tech as well. There's are reason why all the tech company icons are blue-on-white or rainbow.