Lifeguards have very short periods of diligence before they take mandatory breaks in an extremely controlled environment. Train conductors operate on grade separated infrastructure. Security Guards do not have to take split second action or die.
Putting a warm body in a mind-numbing situation and requiring split second response to a life or death situation at a random time is a recipe for failure.
Expecting people to be able to behave like machines is generally the attitude that leads to crash investigations.
Same, been kicking myself since I found out it was all gone a few weeks ago. Don’t know why I didn’t make a ‘just in case’ backup / export.
Still infuriating they can just go “oops all gone”. It came through the roll-out fine, I remember looking stuff up in February. As far as I can tell it was a later unrelated glitch.
Google also accidentally deleted a random amount of user’s timeline data if you didn’t immediately catch it and restore from back up last March before the affected backups were overwritten. If you didn’t keep a close enough watch on your timeline to know that that happened, everything before ~Feb 2025 is gone now.
Ask me how I know. Yes I kept up on permissions. Yes I had backups on. No I didn’t have a new device. I even have dozens of available gigabytes of paid storage on Google One.
I’m sure it will only get more stable due to maps and timeline being revenue generators that encourage investment.
I’m sure that there are examples of actually wasted money, but just putting it out there that planning is fucking important. There have been several high profile projects, like Texas high speed rail, where planning was the hard part and the project got canceled as they were ready to break ground because “there was no progress”. Cue* Republicans “the government does nothing” after they stopped anything from happening. Infrastructure cannot operate on election cycle timelines.
Digging in the ground and integrating with existing infrastructure isn’t just a plug and play operation. Leases and liens need to be sorted out. Estimates of current and future demand needs to be sorted out so you don’t install useless networks. Fiber isn’t that heavy, but “can the existing conduits under bridges/roads/etc support it and/or do they have room to without a complete replacement” isn’t a trivial question for backbone lines.
Winging it just causes more problems as you find things you didn’t anticipate and cause delays while having to continue paying contracts so work can resume once the delay is cleared. If you don’t, the contractor is on to their next job and unavailable for an effectively random amount of time. While everyone is mad at you that “no work is being done”.
It could be done faster, but it would cost more. Because planning is really important to keep multi-million/billion dollar projects accountable and on track.
Sure. But in the first Fast and Furious movie they’re not driving syncro-less transmission semis. They’re driving tricked out sports cars in a straight line and somehow having about 14 gear changes in a 6 speed manual.
And what I mean is that prior to the mid 1900s the etymology didn’t exist to cause that confusion of terms. Neither Babbage’s machines nor prior adding engines were called computers or calculators. They were ‘machines’ or ‘engines’.
Babbage’s machines were novel in that they could do multiple types of operations, but ‘mechanical calculators’ and counting machines were ~200 years old. Other mathematical tools like the abacus are obviously far older. They were not novel enough to cause confusion in anyone with even passing interest.
But there will always be people who just assume ‘magic’, and/or “it works like I want it to”.
“Computer” meaning a mechanical/electro-mechanical/electrical machine wasn’t used until around after WWII.
Babbag’s difference/analytical engines weren’t confusing because people called them a computer, they didn’t.
“On two occasions I have been asked, ‘Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?’ I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.”
- Charles Babbage
If you give any computer, human or machine, random numbers, it will not give you “correct answers”.
It’s possible Babbage lacked the social skills to detect sarcasm. We also have several high profile cases of people just trusting LLMs to file legal briefs and official government ‘studies’ because the LLM “said it was real”.
I think because it’s language.
There’s a famous quote from Charles Babbage when he presented his difference engine (gear based calculator) and someone asking “if you put in the wrong figures, will the correct ones be output” and Babbage not understanding how someone can so thoroughly misunderstand that the machine is, just a machine.
People are people, the main thing that’s changed since the Cuneiform copper customer complaint is our materials science and networking ability. Most things that people interact with every day, most people just assume work like it appears to on the surface.
And nothing other than a person can do math problems or talk back to you. So people assume that means intelligence.
That’s what config files are for. It would be a nightmare to hardcode weight and balance and have to recompile the HUD every time you change the loadout or refuel the plane.
Most code, algorithms, etc are not any more sensitive than the concept of desks and file cabinets. No, guidance programs for missiles probably shouldn’t be put on GitHub, but there’s a reason RSA and other encryption algorithms were open sourced. It’s better to have more eyes looking for inefficiencies, weaknesses, and vulnerabilities than to just assume it’s good because no-one on the team responsible is smart/engaged enough to find them.
A lot of functionality can be decoupled from anything that needs to be classified. A HUD is a HUD and no one should be hard coding in performance characteristics of the F-35 into it. I’ve also worked on government projects and holy crap does the code quality vary wildly, even before you get into “it’s still working so deal with the problems, it doesn’t have the budget for updates”.
Using ‘off the shelf’ parts/code can save significant time and money. There’s a reason subs use xbox controllers. Government websites and data interfaces at the very least should have the audit-ability that open source provides.
In theory yes, but not a lot of people are uploading their family photo albums AFAIK.
I think it would be interesting to have some kind of global archive. Even if descendants don’t care “now” has the potential to be the beginning of the best documented era in history. Historians would kill for photographs by random average people from any other time.
A lot of people thought that that’s what the Internet would be, but that’s obviously not the case. And I know the “right to be forgotten” is a thing, and deservedly so, but at some point you’re throwing out the wine with the amphora.
Probably a me problem but kept having problems with that docker on unraid, it’s just in the community apps ‘store’. The vm seemed to just crash randomly.
I switched over to their B2 storage and just use rclone to an encrypted bucket and it’s ~<$5/mo which I’m good with. Biggest cost is if I let it run too often and it spends a bunch of their compute time listing files to see if it needs to update them.
If my family hired an actor to impersonate me at my killer’s trial and give a prepared speech about how I felt about the situation it would be thrown out of court.
If my family hired a cartoonist or movie studio to create a moving scene with my face recreated by digital artists and a professional voice actor to talk about my forgiveness for my death, it would be thrown out of court.
That they used a generative program to do it and the Judge allowed the video to influence the sentence as if it were a statement by the deceased is deeply troubling.
I was averaging ~1-2% blocked using the firebog and a few other lists, I also have ublock origin on everything I can. Added hagezi’s ‘pro plus’ list last month and it’s up to 39% blocked.
I bought a trmnl and it’s pricey but works pretty good. I’ve mostly been using a few out-of-the-box plugins for it.
There is a selfhosted/offline version of the server you can run for it, so it can be ‘offline’ in theory. I keep meaning to mess with it more but haven’t put the time aside.
Building from source is always going to come with complications. That’s why most people don’t do it. A docker compose file that ‘just’ downloads the stable release from a repo and starts running is dramatically more simple than cross-referencing all your services to make sure there are no dependency conflicts.
There’s an added layer of complexity under the hood to simplify the common use case.
You can install key and screen loggers if you want. Could even setup offsite backup and rclone it all wherever you want.
Lifeguards take breaks every ~20 minutes, not just to look down or zone out, to get up and move around. And again, are in an extremely controlled environment looking for a very small number of specific problems.
Elon is making programmers sleep at their desks.