riskable

I work for a huge bank that's investing a non-trivial amount of money (billions) in single family homes. They don't plan to rent them out. They just want to own them.

Now why would a huge, rich bank invest in something like that? Because they're pretty sure they're going beat inflation when they resell those properties later. It's a very safe (if spread across the entire US and Canada) place to park money.

It's not a big deal if one or two banks do this or even a handful of private equity firms. However, when all of them do it at once (like they are now) it can have a major impact on the prices of single family homes. It also creates something called a, "systemic risk" but that's a very large topic that I'm not going to cover here.

The point is that yes: The big banks and big private equity firms (and 401ks!) all own way too much non-commercial real estate in general right now and their expansion into single family homes is a great big societal problem.

...but why now‽ Why haven't they been investing in huge swaths of single family homes since forever? I mean, they've been appreciating faster than inflation since forever with only a few minor hiccups (e.g. 2008). The answer is: It used to be much more expensive to maintain homes that don't have anyone living in them.

Back in the day most homes were unique. In any given neighborhood some homes might have gas heat while others had electric and some others used oil or coal! There were also more fire and flood hazards with more flammable furnishings/building materials and things like washing machine hoses would often just break after a certain amount of time (the seals were only good for like ten years).

These days you have endless amounts of cookie cutter homes in enormous neighborhoods all over the damned place. They're also built to vastly superior building standards and come with appliances and AC that are orders of magnitude more efficient than in decades past.

This means a big bank or private equity firm can buy hundreds of houses in a region and (cheaply) hire a 3rd party to look after them. They just don't need as much maintenance as they used to. They're so much cheaper to maintain en mass.

So how do we fix this problem? There's all sorts of things you can do but some quick and perhaps unexpected things are:

• Raise minimum wage and crack down on businesses hiring illegal workers doing house maintenance work (let them do construction 👍).
• Raise property taxes in general. You could try to raise them for homes without people living in them but then you just end up creating other unintended consequences/problems (which I won't get into to stay brief)
• Force upgrades on unoccupied homes. Air conditioning system is 10 years old? You have to get a new one with improved efficiency. House has gas stoves? You have to replace those.
• Force inspections of unoccupied homes and come down hard in regards to code enforcement (every unoccupied home poses a nonzero fire risk to every neighborhood).

Basically, you have to turn unoccupied homes into expenses again. When that happens the banks and private equity will get the hell out.

There's lots of private equity that will just convert to being slumlords but the big banks do not want to be renting out anything. It's a huge risk for them and looks real bad on their balance sheets from a banking perspective. Also, if a bank is big enough they're straight up forbidden (by law) from renting out properties (though there's various loopholes which I won't get into).

I've been working in information security for many years now and this is standard practice all over the world. I'm sure the State Department is happy to have Asian Americans working for them. This is not that kind of discrimination.

It's a practical thing: It's basically impossible to tell if a person has a secret loyalty to another country. So how do you prevent spies from infiltrating your organization(s) from any given county? You take basic precautions like preventing people who have very direct ties to a country from working for you in that country. It's basically best practices because there's not much else you can do.

If I were a 2nd generation American Chinese citizen (e.g. my parents immigrated there) living in China and I applied to work for the Chinese embassy in the US do you really think the Chinese government would allow that? Hell no!

Would they let me work in Africa or other areas of Asia? Maybe! But the US would be a no-go zone for sure. There'd be no way the Chinese government would trust me not to have a secret loyalty to the US.

Not only that but if someone still has family living in a foreign country that's a big red flag as well. It can get you reassigned regardless of your ethnicity.

Let's say you've been working in the US embassy in China and your sister married a Chinese man who has family in China: You would would have to report that change in family status to the State Department (it's part of the rules). Then they'd make a determination as to whether or not you could keep working where you are.

You've set off my 1984 alarm: Never use the term, "thought police" in reference to private individuals making decisions about others actions. If you actually read 1984 (it's a short book and easy to find online, go read it) the Thought Police were part of the government. As in, policing people's thoughts was a function of the government (in order to maintain the status quo, avoiding change aka extreme conservativism).

Here, you're referring to a collection of people that have decided--on their own--to boycott a comedian because they don't like where he stands on certain topics. That's not Thought Police! Call them snowflakes or "too sensitive" or "hysterical" or some other bullshit (don't care, really) but please for the love of Orwell stop using that term to refer to non-governmental entities or the actions of people that aren't part of any authority.

When the government starts cracking down on people's speech then by all means refer to this action as, "Thought Police".

This has got to be the most boring dystopia thing I've ever read (so far).

It'll be a good while before an AI generates an Oscar-winning script or a whole movie but most movies and TV shows are very formulaic. Would it really be that surprising if AIs were generating the entertainment equivalent of Hannah Montana in a few years? Or the latest Hallmark Christmas special (LOL)?

My guess is five years: That's how long it'll be before we start getting a flood of half-decent AI-generated shows/movies. Where the script is good but the animation/video are "a little off".

I mean, come on: There's so many successful TV shows and movies that are total shit! You think AI can't do better with just the tiniest bit of evolutionary improvements (and better hardware)?

Edit: I expect AI videos to be a revolution! Where we finally break free from the Hollywood and "big mega" cookie cutter stories. It'll give creative people the power to make the movies they want without heavy-handed censorship and executives that require everything dumbed down for the lowest common viewer.

AI will follow a similar curve as computers in general: At first they required giant rooms full of expensive hardware and a team of experts to perform the most basic of functions. Over time they got smaller and cheaper and more efficient. So much so that we all carry around the equivalent of a 2000-era supercomputer in our pockets (see note below).

2-3 years ago you really did need a whole bunch of very expensive GPUs with a lot of VRAM to train a basic diffusion (image) model (aka a LoRA). Today you can do it with a desktop GPU (Nvidia 3090 or 4090 with 24GB of VRAM... Or a 4060 Ti with 16GB and some patience). You can use pretrained diffusion models at reasonable speeds (~1-5 seconds an image, depending on size/quality settings) with any GPU with at least 6GB of VRAM (seriously, try it! It's fun and only takes like 5-10 minutes to install automatic1111 and will provide endless uncensored entertainment).

Large Language Model (LLM) training is still out of reach for desktop GPUs. ChatGPT 3.0 was trained using 10,000 Nvidia A100 chips and if you wanted to run it locally (assuming it was available for download) you'd need the equivalent of 5 A100s (and each one costs about $6700 plus you'd need an expensive server capable of hosting them all simultaneously).

Having said that you can host a smaller LLM such as Llama2 on a desktop GPU and it'll actually perform really well (as in, just a second or two between when you give it a prompt and when it gives you a response). You can also train LoRAs on a desktop GPU just like with diffusion models (e.g. train it with a data set containing your thousands of Lemmy posts so it can mimic your writing style; yes that actually works!).

Not only that but the speed/efficiency of AI tools like LLMs and diffusion models improves by leaps and bounds every few weeks. Seriously: It's hard to keep up! This is how much of a difference a week can make in the world of AI: I bought myself a 4060 Ti as an early Christmas present to myself and was generating 4 (high quality) 768x768 images in about 20 seconds. Then Latent Consistency Models (LCM) came out and suddenly they only took 8s. Then a week later "TurboXL" models became a thing and now I can generate 4 really great 768x768 images in 4 seconds!

At the same time there's been improvements in training efficiency and less VRAM is required in general thanks to those advancements. We're still in the "early days" of AI algorithms (seriously: AI stuff is extremely inefficient right now) so I wouldn't be surprised to see efficiency gains of 1,000-100,000x in the next five years for all kinds of AI tools (language models, image models, weather models, etc).

If you combine just a 100x efficiency gain with five years of merely evolutionary hardware improvements and I wouldn't be surprised to see something even better than ChatGPT 4.0 running locally on people's smartphones with custom training/learning happening in real time (to better match the user's preferences/style).

Note: The latest Google smartphone as of the date of this post is the Pixel 8 which is capable of ~2.4 TeraFLOPS. Even 2yo smartphones were nearing ~2 TeraFLOPS which is about what you'd get out of a supercomputer in the early 2000s: https://en.wikipedia.org/wiki/FLOPS (see the SVG chart in the middle of the page).

Whoah there: Who says AI influencers aren't the result of individual's honest work? You don't need an entire data center of computers to make your own AI influencer!

Don't assume there's a corporation behind every AI persona. It could just be one guy with a lot of VRAM getting creative with prompts in his parent's basement.

I chuckled at this but I would like to point out that we shouldn't dehumanize influencers. They are just as human as Mark Zuckerberg and Elon Musk.

Wait...

AI is stealing the job of immigrants, being blamed for stealing jobs!

Don't see what the big deal is. We humans can also absorb water and other fluids through our butts. Just ask Brett Kavanaugh.

As someone who has worked in government and private industries of all sizes let me tell you the takeaway from my experience: Only organization size matters. The only exception is in the "small stuff":

• Small government entities (municipal stuff) are often extremely efficient and frankly, surprisingly competent.
• Small businesses are much more likely to be wildly inefficient and incompetent. They're also much harder to police and can often exist solely to extract as much money as possible from a government contract while providing as little benefit/output as possible (the bare minimum). Safety is never a priority and anything that can be made someone else's problem will be (externalities).

Big business and big government are both extremely slow and wasteful but in different ways. Big government wastes time and money on simple things that should be cheap but because of various laws and regulations must adhere to regulations of all sorts they end up being expensive (and these regulations often don't keep up with the times). This also slows everything down because you have to wait for the stuff to pass muster before you can use it most of the time (no matter what that thing is... From simple paper products to chairs to industrial equipment to desks to rocket engines etc you name it). This often results in people having to wait (sitting on their asses while still getting paid).

Big business wastes money on 3rd party tools and services that are often completely unnecessary. Usually because the powers that be "have always done things that way." They also waste money by being really, really bad at project management. This is the big one: At any big company something like 9 out of 10 IT projects are considered failures because they just keep going forward (with the project) no matter what. So they often end up with something that needs to be maintained/replaced and ends up becoming a regular, long term expense.

Big business isn't usually corrupt but they will spend loads and loads of money lobbying to make it easier for them to extract profit from whatever it is that they do. Safety, ethics, and things like the general well-being of society be damned. They have no morals except those codified in law whereas the people in huge government organizations are very visible to the people in general and know they have to act ethically or they could get in big trouble (and there's whole entities who's job it is to watch them for bad behavior and inefficiencies).

Related: There's never "too much" or "too little" regulation. There's just good regulations and bad regulations. Anyone who says regulations are bad or insinuates that they're "job killing" is looking to mislead you.

If the volume knob was connected to the amp you'd hear the static from a shitty potentiometer that's wearing out. Instead what you get is a volume knob that occasionally skips steps because it's an electromechanical rotary encoder and doesn't rely on brushes rubbing against a gradient resistive wheel (that literally wears itself away over time which is why car manufacturers switched to rotary encoders in the first place).

The software sucks too (absolutely!) but it's pretty obvious when the problem is one of the following:

• Skipped control (e.g. volume) steps. Indicates that a contact has worn out (oxidized too much).
• The car suddenly thinks a knob is being turned constantly in one direction (e.g. volume suddenly goes up up up or down down down sometimes forever until you move the knob). This can be "bouncing" or just a contact that's getting stuck (because dust/car gunk got in there).

These two things are clear indicators of electromechanical components failing. Not normally caused by buggy software.

Neither of these things happen when you use hall effect switches or hall effect rotary encoders (for knobs).