3,374 karma · joined July 30, 2024
A quick look on Dell's web page shows an entry-level Core i3-14100T system with 8 GB RAM for $747.74, which would be under £600. A $2,230 system from their Alienware brand currently gets you a Core Ultra 7 265F, GeForce RTX 5070 and 32 GB of RAM.
I can't imagine most office workers would need more than the entry-level system, although corporate purchases are known for overprovisioning, especially so when it's a government body, so it wouldn't be unheard of to spend $2,300, especially so if that includes paying an IT worker to set it up.
Most of the work new technology replaces is work we made up after we invented tractors and power looms and no longer needed to perform physical labor. Most of what humanity does isn't required to survive or even thrive. Even before LLMs, when jobs in our lifetimes have been lost to automation, it wasn't because a welding robot replaced someones factory welding job; it's usually an office job that's getting replaced because a company started using software to do something that an office worker had been doing by hand.
Walk into an accounting office at any small to medium sized business that has been around for a while and does their accounting in-house, and you'll find people looking at a number on the computer, typing it into a calculator to do the math, then entering the result into the computer. We switched from paper to computers a career lifetime ago, but that doesn't mean everyone, or even most people, have adapted their work flow. When that job is outsourced to a third-part accounting firm, it's cheaper because they're using spreadsheets and enterprise resource management software the way it was meant to be used, not because they're paying their workers less. Likely, they're paying more. For most companies, this could have been done in the 90's or even 80's, but they tend to adapt slowly, so change more often comes from old companies or departments being replaced by a completely new company that built themselves around the technology.
Pretty much every technology advancement has been what militaries call a force multiplier. It means that an individual or small group can do a job that used to take a larger group, but it doesn't mean that zero people can do what one person was doing alone. This increase in worker productivity leads to people doing things they wouldn't have been able to do before, which makes more room for us to make up new jobs.
Also, art is the fields most sensitive to income loss from this, because it decreases the barrier to entry, and gatekeeping is often the only thing making art expensive. A creative director can lead a group of artists to create a complex and expensive project, but as worker productivity improves, what was a creative director role becomes an individual role creating complex projects for cheap. It used to take a team of people to produce works like radio programs, game shows, documentaries, educational videos, and the like, but as technology has improved, a single person with a few hundred to thousand dollars of equipment has the capitol to produce a one-man podcast or YouTube channel that rivals what those teams did decades ago. Of course not everyone that tries will be good at it, but a lot more of the people that are good at it will try than would have tried in the past, and listeners and viewers now have far more options for a given quality of content, so the value of that content has decreased. That worker isn't worse off, because the income isn't being split between an entire team, but it does mean the workers that would have formed the rest of the team need to adapt.
Granted, the only reason that content has value at all is because we don't have to spend most of our waking hours plowing fields and weaving clothes.
The universities themselves may react to market demand, if subsidies for their services decline, but it's also possible for subsidies to increase, despite the lack of value, as is often the case for a declining industry with a significant political presence.
† Student loans may be repayed by the recipient, but they are not granted based on the market value of the loan, so they act as much like a subsidy as student loans paid by third parties.
Samsung has already worked out the major kinks for folding displays, over dozens of products used by multiple phone manufacturers including themselves, and Android has worked out how to handle the various use cases between the outer screen and inner screen, when closed, open, or partially open, so Apple can pull from the current state of the art and launch with a more fully developed product, as they usually did before the Apple Vision Pro.
It makes sense to have one screen for an as-large-as-possible uninterrupted display, and another for maximum utility.
I like that Samsung doesn't cut as much corner off as Apple, but ideally I'd like to find a manufacturer that doesn't cut corners at all. Even though I'm not likely to buy another iPhone product, I am excited that Apple doesn't cut holes in their interior display, because that'll probably get other manufacturers to do the same.
Anyway, traveling efficiency is a factor of the speed and size of the vessel and the viscosity of what it's traveling through. It's non-linear, and scaling up the vessel size slowly adds efficiency, and speeding up the vehicle or increasing the viscosity quickly reduces efficiency.
Boats have an advantage by getting very big, but a really, really big disadvantage if they go fast, which is why huge boats going slow is their most common use, mostly for cargo, but sometimes for people, when the ship itself can be a destination. Trains can't get very big, but air is much less viscous than water, so they can go a few times faster without major losses. This is why rail is really popular for transporting cargo across the US and Canada, although it's not fast enough for long distance passenger travel.
The absolute peak in energy efficiency comes from a large vessel moving through a very-low viscosity atmosphere, which is what airliners are able to do, by traveling in the stratosphere, eliminating the majority of the drag, and allowing for effectively uncapped speeds. Really, the faster an airplane can travel, the higher it can go, further reducing drag and increasing efficiency. The limit is technological, not physical, but current airplanes traveling around mach 0.8 get very high cruise efficiency from both being large and being in a low-viscosity atmosphere.
Also, it takes a lot of compute. It still involves training a model, which for a very small model requires a GPU with several times more ram than the final model size, which will need to run for days to weeks. Usually you're also running the model that is being distilled from, but in this case it would require a huge database of chat logs, instead.
Granted, the difference in emissions of either form of traveling are absolutely nothing compared to the emissions of Eyjafjallajökull itself.
My bank sent out an a announcement that they will support voice identification, right after a news story went viral of a scammer using a voice cloning model to steal millions of dollars. They pretty quickly followed it up with an announcement that they are aware of cucumbers' concerns and that they will monitor changing technology, but will still support voice identification. What I read from that is that they signed a contract and will keep using the service until after one or more of their customers are exploited.
Turbulence is probably the leading cause of injuries and deaths from modern air travel, but taxiing incidents at extremely rare and pretty much never result in injuries.