bane

During COVID, by father was dying in a hospice center (from something else, not COVID). Because of the circumstances and the condition of the medical system, it was impossible to get into where he was to see him, and the staff was too overworked and understaffed to find a way to connect us to him. He was alone, fragile, dying, and terrified.

As my mother and I sat in the reception area fighting with the hospice administrator, a medical transport pulled up and unloaded another patient. After putting the new patient back in their room the driver walked up to us as we were sliding into a heated argument with the hospice administrator.

She asked the administrator what the problem was and was told that policy was visitors can't be going into the patient area and it was very firm. They'd had issues with the local government about being slack about it. The driver turned to me and said something along the lines of "here's what we're going to do:

Since I can apparently run around freely in this place, I'm going to find your father and put a star in his window so you can always find where he is.

Number two, I'm going to give you a set of full hazard gear.

Number three" and she turned to the administrator put her finger up into her face and very sternly said, "they are going to hire you as a part time employee, in maintenance, or IT support or whatever, and your hours of employment here are going to be whenever you need to visit your father."

she turned back to me, "but this doesn't mean it's a free pass, you are going to wear all of this hazard gear whenever you come 'work' here, promise me that okay?"

She then took the administrator off to a side room, had a conversation, and I had a piece of paper to sign about 30 minutes later making me an employee of the hospice.

I made it into and out of the hospice without incident for the next week until we decided to bring my father home to die as he wasn't receiving almost any care there. I don't know what the ambulance driver and the administrator discussed, but I suspect it was the absolutely woeful state of the facility.

The look on my father's face when a head-to-toe masked man entered his room the first time, and when I took my mask off to show him I was there for him, and how the terror just simply fell from his face, is something I will always remember as is the kindness of the driver who put herself out there for us.

The period was incredibly hard, beyond the situation with my father, the medical system was in absolute shambles, and as my father's health was rapidly deteriorating, it was among the only kindness we received during that wretched journey.

I don't know why my sense of ethical outrage over the cancerous spread of this worthless garbage technology doesn't extend so hard to LG in this. But I've finally just hit the wall with Microsoft. Between all the forced account sign-ins, ads, random reboots and updates that aren't for the purpose of securing the OS...I'm done.

I haven't seen this level of anti-consumer nonsense from Microsoft since the late 90s when lots of circles called them Micro$oft.

It's a shame since Satya Nadella came in and made a lot of right moves. Support for Linux, open source, etc. I could stretch myself to forgive the other stuff as some kind of wrong headed thinking about a cloud-first strategy. But in the last few years that productive pivot stopped and the company moved back into high-risk money grab steps.

My current daily driver is a Windows machine, but my last few builds around the house have all been some kind of Linux. Last year I moved my home server infrastructure over. I kept a windows VM around for a few things but it ultimately corrupted itself and was replaced by a Linux VM that's been chugging along just fine.

I think when I rebuild my home desktop sometime next year Windows gets relegated to a "run when needed" VM. I've really only kept it around for games and a few other Windows only software but those days are fast fading thanks to numerous efforts by Valve and others and a fallback VM for other stuff works fine in my experience elsewhere.

I think I don't have the same concerns about LG because the relationship with me as the consumer seems different somehow, and I simply expect less from them?

Or if these chat apps could hold a conversation as basic as "I don't know where you are, would you allow me access to your gps?"

A lot of AI models also suffer this flaw.

It also pioneered billions of other users with "hey google" and "siri" uselessness which also copied and then completely flatlined with things to do beyond calling the wrong person in your call list, setting timers, and playing the wrong song.

You have good points!

For me, eliminating the few hours dealing with some stupid config option I messed up is easily worth $250, which unraid basically makes go away. But yeah, most of the things the system does are really just some basic linux distro with various bits installed.

The terramaster NAS's are surprisingly reasonable for pre-built NASs. And I think they'd be fine if you want an <4 bay NAS solution.

There's a few tiers I think of home NAS users:

0 - add an external hard drive to my base machine, maybe share it (or the content) to other users on the network

1 - put the drive on the network, there's a bunch of OEM "drive on your network" options there

2 - the 3-5 bay home NAS user. Used to be a medium user, but now can easily hit ~100TB.

3 - the greater than 6 bay home NAS user.

At some point, I moved from #2 to #3 and decided that I had enough spare drives lying around that it was worth it to invest in a big case and centralize everything in one box. It's around this time that I think the cost efficiency of the NAS hardware really left me behind.

A 6-bay terramaster is around $500 and provides an N95 and 8GB of memory. So basically in rPi5 territory.

A 12-bay terramaster is like $1800 is with an i7-1255U and 16GB of memory.

I built a 16+ bay unraid system this year for around $1500 that included an i9-12900k and 128GB of memory + an unraid lifetime license. I know I'm a #3 NAS user, but the different in price is a bit, for less capable equipment.

You could build the same system, and save about $400 if you put it together and just put some linux distro on it.

I guess that point I was making originally still stands, I think these retailers could build really nice NAS options on a big price undercut. At volume I bet they could negotiate moving unraid licenses for much cheaper as an OEM option.

Note: Yes! I acknowledge none of these choices really include the actual drives, any of these options might allow for a gradual in-fill and replacement of drives over time.

As an American, we grow up almost entirely without this gem of children's literature. I'm so thankful that PBS aired this story when I was a small child. The imagery was so strong that it has forever stuck in my head. When I see other stories like "The Fountain" or Super Mario Galaxy, I immediately think of the Little Prince.

I've yet to revisit it as an adult, but I think maybe it's time?

Unraid weirdly requires booting off of a USB for the base OS. I think it's to manage licensing.

SSDs are generally expected to be used as write-through caches with the main disc pool. However, if you have a bunch you can add them to a ZFS array and it works pretty much flawlessly.

I personally think big-box computer retailers that build custom turn-key computers (e.g. Microcenter) should get into the NAS game by partnering with unraid and Fractal. It's as turnkey as any commercial NAS I've ever used but comes with way more flexibility and future proofing and the ability for users to get hyper technical if they want and tweak everything in the system.

It's wild how much more cost effective this would be than pretty much any commercial NAS offering. It's ridiculous when you consider total system lifecycle cost (with how easy it is to upgrade unraid storage pools).

Looking right now and my local Microcenter builds essentially three things: desktop PCs, some kind of "studio" PC, and "Racing Simulators". Turnkey NASs would move a lot of inventory I'd wager.

I've also considered a side-effect of this. Each generation of software engineers learns to operate on top of the stack of tech that came before them. This becomes their new operating floor. The generations before, when faced with a problem, would have generally achieved a solution "lower" down in the stack (or at their present baseline). But the generations today and in the future will seek to solve the problems they face on top of that base floor because they simply don't understand it.

This leads to higher and higher towers of abstraction that eat up resources while providing little more functionality than if it was solved lower down. This has been further enabled by a long history of rapidly increasing compute capability and vastly increasing memory and storage sizes. Because they are only interacting with these older parts of their systems at the interface level they often don't know that problems were solved years prior, or are capable of being solved efficiently.

I'm starting to see ideas that will probably form into entire pieces of software "written" on top of AI models as the new floor. Where the model basically handles all of the mainline computation, control flow, and business logic. What would have required a dozen Mhz and 4MB of RAM to run now requires TFlops and Gigabytes -- and being built from a fresh start again will fail to learn from any of the lessons learned when it was done 30 years ago and 30 layers down.

Perl "died" through a 1,2,3 knockout of

1. failing to have a coherent path to Perl 6

2. Ruby (on Rails) taking over the workhorse task of serving up dynamic content that Perl had owned before then

3. Python completely dominating the utilitarian scripting/programming world in nearly every niche

Why did this happen? I was a work-a-day developer working in Perl v5 when this transition occurred and from my perspective and recollection v6's meandering development cycle -- which didn't really address the issues of the broader Perl community was the primary choice. Perl 6 was developed in a way that didn't address the broad concerns of the Perl community, and expected people to make a wholesale switch to what was effectively an entirely new language anyways. It forced people to go out looking and what they found were either stronger solutions to specific domains (Ruby on Rails) or a nicer language than what was being proposed (Python).

Where Python really excelled at the time was that it looked and worked very much like the pseudocode that was going around at the time, and had an opinion about how you should write your code. Perl is wonderful to write in, but in many ways is too expressive and permissive and it resulted in an ungodly mess that could be hard to maintain. Perl 6 simply leaned into that problem rather than encouraging a cleaner approach.

I never liked Python much, but damn if I couldn't argue that I was much more productive with it than Perl in the end. Which was weird because when I was really hacking in Perl I could write code almost as fast as I could think giving a kind of illusion of productivity. But Python was easier to integrate into a coherent team development structure and actual productivity is more important than feels.

I miss working in Perl. But I knew it was really finally dead when I was giving tutorial classes to new bioinformaticists who were being given old Perl codebases to update and they were getting through school without learning the language.

You're welcome. I think hearing her talk about her life and experiences in her own way is incredibly rewarding. The ANTIC podcast has done a great job of preservation of people involved in early computing and it deserves a lot of support.

A little late, but man this sucks, cancer sucks.

Rebecca was not only an amazing programmer, but a true hacker from the get go. From what I understand she managed to achieve what she did without even a high school diploma -- a real natural talent.

I first really learned of her from the ANTIC podcast [1] in 2015 and was just kind of blown away by this cool, intelligent, creative and humble human being.

I'm personally sad she's gone, but also really...proud? to see how she went out, with tons of witty communications to her friends and associates in her recognizable voice.

To have such a positive impact in the world is something worth achieving.

1 - https://ataripodcast.libsyn.com/antic-interview-64-rebecca-h...

This is shockingly almost exactly the same conversation that goes on in the retrogaming community regarding pre-HD consoles. "The artists meant for this art to be viewed on a CRT".

The challenge is that everybody's memory is different, and sometime those memories are "I wish the graphics were rock sharp without the artifacts of the CRT". Other times our memories are of the crappy TV we were given as kids that was on its last legs and went black & white and flickered a lot.

The reality is that no matter what the intentions of the original animation teams were, the pipeline of artwork through film transfer to projection to reflection to the viewer's own eyeballs and brain has enough variety to it that it's simply too variable -- and too personal -- the really say what is correct.

Anecdote: one of the local theaters I grew up with was extremely poorly maintained, had a patched rip on one of the several dirty screens, and had projectors that would barely get through an hour of film without needing a "bump" from the projectionist (allowing the audience to go out and get more refreshments halfway through most films). No amount of intentionality by the production companies of the many films I saw there could have accounted for any of that. But I saw many of my favorite movies there.

I've come down with the opinion that these things are like wine. A good wine is the one you enjoy. I have preferences for these things, but they sometimes change, and other people are allowed to enjoy things in their own way.

This is a great reminder that the Eternal September still exists and perhaps mercifully appears to be affecting those with at least some technical exposure.

https://en.wikipedia.org/wiki/Eternal_September

I'm struck with how long the history of Apple's earliest iPhone has shaped and produced long-term damage to the concept of digital ownership. Apple originally didn't allow anybody but Apple to create software for the 1st gen iPhone, and only later was forced "opening" it my market forces.

People who realized they actually owned the thing they bought wanted to do what they wanted, which required circumventing Apple's control or "jailbreaking". This differentiator stimulated Google to "allow" installing on Android without "jailbreaking" the device aka "sideloading", giving the illusion of the kind of freedom that was never in question on normal computers.

It is interesting though how this same conversation doesn't exist in the same way in other areas of computing like video game consoles or other embedded computing devices where the controls against arbitrary applications is even stronger.

The fact that mobile phones aren't yet just a standard type of portable computer with an open-ish harware/driver ecosystem that anybody can just make an OS for (and hence allow anybody to just install what they want) is kind of wild IMHO. Why hasn't the kind of ferver that created Linux driven engineers to fix their phones? Is Android and iOS just good enough to keep us complacent and trapped forever? I can't help but think there might be some effect here that's locking us all in similar to how the U.S. healthcare system can't seem to shake for profit insurance.

I'm sometimes surprised at the plethora of cheap handheld gaming systems coming out of China that support either Linux, Android, or sometimes both, and seem to be based on a handful of chipsets. If anybody ever slapped an LTE module and drivers onto one of those things we'd have criminally cheap and powerful, open phone ecosystem.

That's basically correct. But the more general problem is that engineers simply lost the ability to succinctly package applications and their dependencies into simple to distribute and run packages. Somehow around the same time Java made .jar files mainstream (just zip all the crap with a manifest), the rest of the world completely forgot how to do the equivalent of statically linking in libraries and that we're all running highly scheduled multithreaded operating systems now.

The "solution" for a long time was to spin up single application Virtual Machines, which was a heavy way to solve it and reduced the overall system resources available to the application making them stupidly inefficient solutions. The modern cloud was invented during this phase, which is why one of the base primitives of all current cloud systems is the VM.

Containers both "solved" the dependency distribution problem as well as the resource allocation problem sort of at once.

Containers happened because nobody can be bothered to build an entire application into a single distributable executable anymore - heck even the tooling barely exists anymore. But instead of solving problems like dependency management and linking, today's engineers simply build piles of abstraction into the problem space until the thing you want to do more than anything (i.e. execute an application) becomes a single call.

Of course you now need to build and maintain those abstract towers, so more jobs for everybody!

Looking at the article, my read (which could be wrong) is that the backup was in the same room as the original.

One of my first jobs was helping build an expert system for a, even today, complex computational linguistics problem. The company had a rich corporate library full of academic books on expert systems, decision trees, first gen (pre-winter) AI, and some early books on early ML approaches. I remember seeing this book in particular and its evocative title caused me to look deeper into the library than I would have normally.

Our core system was built of thousands upon thousands of hand-crafted rules informed by careful statistical analysis of hundreds of millions of entries in a bulk data system.

Part of my job was to build the system that analyzed the bulk data and produced the stats, and the other part was carefully testing and fixing the rulesets for certain languages. It was mind-numbing work, and looking back we were freakishly close to all the bit and pieces needed for then bleeding-edge ML had we chosen to go that way.

However, we chose expert systems because it gave us tremendous insight into what was happening, and the opportunity to debug and test things at an incredibly granular scale. It was fully possible to say "the system has this behavior because of xyz" and it was fully possible to tune the system at individual character levels of finesse.

Had we wanted to dive into ML, we could have used this foundation as a bootstrap into building a massive training set. But the founders biased towards expert systems and I think, at the time, it was the right choice.

The technology was acquired, and I wonder if the current custodians use it for those obvious next-step purposes.

Goodness, I have over 100TB at home and it cost less than a two or three thousand dollars to put in place. That's like $25 per TB.

> The stored data amounts to 858TB (terabytes), equivalent to 449.5 billion A4 sheets.

No, the 858TB amounts to under $25k for the government of the 10th largest economy, of one of the most sophisticated countries on the planet, to put in place.

Two of those would be less than the price of a new Hyundai Grandeur car.

> “It’s daunting as eight years’ worth of work materials have completely disappeared.”

So they're clocking in at around 100TB/year or 280GB a day. It's respectable, but not crazy. It's about 12GB/hr, doable with professional, server level hardware with backup moved over dedicated fiber to an offsite location. Multiply the price 10x and you can SSD the entire thing.

Even with data sovereignty consideration demanding an entirely 100% home grown solution rather than turning to AWS or Azure, there's no excuse. But it's not like the cloud providers don't already have CSAP certification and local, in country, sovereign clouds [1] with multiple geographic locations in country [2]

South Korea is full of granite mountains, maybe its time the government converts one into an offsite, redundant backup vault?

1 - https://erp.today/south-korea-microsoft-azure-first-hypersca...

2 - https://learn.microsoft.com/en-us/azure/reliability/regions-...

You may want to check that your Xeon may already support hardware encoding of AV1 in the iGPU. I saved a bundle building a media server when I realized the iGPU was more than sufficient (and more efficient) than chucking a GPU in the case.

I have a service that runs continuously and reencodes any videos I have into h265 and the iGPU barely even notices it.

GPUs are the land that's staked. They're of limited intrinsic value in the sense that they're in the way of getting to the gold (models), but there's a small supply of it so they go for a lot on the market. But if there was literally any other way to make the models even for a few percent cheaper, they model builders would move to that. On the inference side, most of the cloud providers are looking for pretty much any way to server that up more cheaply, with custom TPUs, or other tensor units of some type.

We saw this with crypto mining where truckloads of expensive GPUs were dumped in the trash after the proof of work became so hard it became not worth the cost of electricity to keep on that generation of card.

Of course they've stood it out. The rate of change and the R&D expenditure is off the charts. It buys them marginal utility to hire AI talent at incredible salaries to keep them at table stakes.

Meanwhile, the models are getting larger and more complex, with more users, putting the support infrastructure well beyond what individuals and even small companies can afford to outright buy. You can easily spend well over a million on even basic infrastructure to try to support some of the newer models and make it available to a few end users.

As a point of strategy for individuals and small entities, it really is cheaper in this case to spin up some AWS instances for a bit to do some LLM work and then spin them down when not in use.

So if you were AWS do you mine for gold? Or do you sell shovels?

In some places they act as a kind of local governing authority similar to a town or village council. In much of the US the lowest level of governing body us the county, and those can be pretty huge and diverse areas so HOAs are used to fill that gap.

One of the remedies HOAs typically possess is the ability to put a lien on your property. The end game to one of these liens can be complicated, but losing your house is among them.

We used to eat out a lot. COVID changed it. What changed? We were forced to really learn how to cook properly, then we saw how much money we were spending eating out, and how much healthier and to-our-taste we could make our own cooking.

When things went back to normal, the prices to eat out had jumped so high, it simply wasn't worth it. $15 of fast food to feed both of us turned into $35-40. A $45 dinner out at a restaurant (taxes and tip included) turned into $60-75 meal. Tip expectations had gone from 15%-18% to 20-25%. Add beer or wine or a cocktail and we're instantly at a $100+ night out.

At home $10 of protein, $5 of vegetables and other ingredients and a good youtube video with a recipe, $15 bottle of wine and we were all set.

Weird, I've never had that problem. You do have to remember to set the sync mode to off and it seems fine after that. It may have been an OS setting that was letting the drives spin down if they were doing so after that.

I've got a handful of 4-bay Mediasonic Proboxes, they're about as barebones of a JBOD enclosure as you can get and run about $110 each unless they're on sale. I have stuffed these things with over 40TB of drives and they've worked solidly for many years.

They typically present as 4 USB connected external drives, which you can softraid into whatever. They also support eSATA. At one point I had 3 of these plugged into a spare intel NUC and used it as a NAS running first Windows, then Ubuntu, then unRAID over about 5 years. It wasn't blazing fast, but got the job done for home use.

I love seeing these cheap builds like this. IMHO NAS boxes are way overpriced, and underpowered. I had multiples the performance and capability piecemeal hacking together my old setup than buying the all-in-one systems that will lose vendor coverage in 2 years.

Okay, I see what you are saying and your perspective on it. It's possible I'm approaching the material from a different point of view that's not really aligned with my experience and would like to see some treatment of authoring and mixing software tools, maybe in the Digital Audio section.

The Synthesis chapter is actually very nicely done.