What are the use cases for local AI models that actually require running on macOS? Surely a commodity x86 system is more appropriate?

Is there even the software support for LLMs on macOS?

Much like the speed of light through a medium varies with the density and optical properties of masses, perhaps C could also be somewhat "plastic" when traveling through "stretched" spacetime where massive gravitational forces warp the perception of time. If you spend your life in a glass block, light traveling through the air around it will certainly look like it's breaking the speed of light through the medium you're familiar with. Sufficiently gravitationally bent spacetime could surely exhibit similar properties. Especially when considering the conditions after the big bang, "constants" like C may be a bit more spacey-wasey timey-wimey than we currently understand.

Don't forget, this period of human observation is absolutely minuscule compared to the total age of even the observable universe. Without obserations taken at geological timescales (or even longer), how can we be certain that the "constants" we are familiar with have always ben constant throughout the history of the universe?

I know some of those words.

As a school IT technician, there's times the Windows 11 start menu stops working entirely.

Just a few years ago the schools were only teaching basic use of word processors and spreadsheets.

And then they axed that, and now we have a generation who can't word process a letter or use excel. Given the amount of receptionist/secretarial, data manager and general managerial jobs well outnumbers programming and computer science jobs, we have a generation of kids who are ill-prepared for the real world they will be entering.

It's not elemental sodium. It's sodium salt, so less poppy boomy.

I say less, as that salt is sodium nitrate

I wholeheartedly agree. And to respond somewhat to the other poster below, using terms like "Treacherous Platform Module", or "Winblows" is fine in an informal forum where the content is created by users, it's a different matter when it's a supposedly professional press release from a serious organization.

By all means, ridicule it, and tear it apart on a technical level, but resorting to name calling is childish and makes your organization seem petty and immature. The FSF struggles enough to make it's point without having their own press release team undermine the noble and important work the FSF are trying to do. There's very few people who actually care about a truly free-as-in-freedom operating system, and using childish name-calling makes it appear that the organization representing free software also don't care and can't take themselves seriously either.

And anyway, most people still using 32-bit time on a 32-bit system at the end of 32-bit time will be "time traveling" to a date in the past anyway, to keep their systems running. Most cases will be either legacy software in VMs, or the retro hobbyist, neither of which should be affected greatly by mismatched dates

Honestly, the systems where this will be an issue are essentially prehistoric. x64 has been supported by essentially every mainstream x86 processor since about 2005. Unless you have a dire need to repartition a Pentium II box, or some ancient Compaq Proliant RAID, this move will have no bearing on most people.

Of course, it might hit those in the retro computing community quite hard, but again, it's not like the i386 versions will cease to exist. You just won't have the latest cutting edge features and support. But again, those users are likely to be using older versions of the software already anyway

How long until we have Fallout-esque robobrains. Seems this sort of computer would be well adapted to powering a humanoid robot, especially with it's learning capabilities. And then how keen are those robots going to be to living a life of servitude. Are we working our way towards a Kaylon threat of our own?

Given a computer is just a very expensive space heater, capturing the "wadte" heat and using it to heat homes is a very sensible idea. "Waste" and "Byproducts" are almost always just "products" waiting for a use.

I used to run Folding@home in the winter on my old dual CPU rig. Definitely helped keep the chill off. In that situation, it was the heat i wanted and the science research was a nice bonus

"Cloud, "NFT", and now "AI"

Most people understand a star as being the center of a solar system. And again, most people can understand a binary star as having 2 stars at the center, with planets orbiting around both.

This system however throws that out of the window. Sure, the stars are a binary system, but to the planet, the White dwarf orbits like a planet. So whilst it is a binary star system, from the planet's perspective, it's not. The white dwarf is just a big glowing planet orbiting a star.

On another note, it would be a fantastic spectacle to view the parent stars on that planet. Sometimes you'd have day and night, other times you'd have day and not-so-day.

Hydrogen electrolysis is the one i think we should focus on. It would produce fuel for vehicles for practically nothing, and help balance the grid. It's also quite responsive to power increases and decreases, meaning it would be great for power smoothing. It also doesn't have issues with inductance, which (large) motors suffer from, which can cause problems on the grid itself.

It's literally a case of turn the electrolysis on, or turn it off. And of course, if you generate too much, you can just flare it off. Of course you can always just store loads of it and then burn it to make more electricity when demand is high and supply is low.

Power grids are a case of balancing supply and demand. You can either increase demand, or cut supply. Cutting supply is difficult with solar, without turning some of it off. A solution to this is to increase demand. Hydrogen electrolysis is a simple, which can be very responsive to ramp ups and downs. This would have the added bonus of producing fuel that could then be burned in vehicles, or even as an alternative to natural gas.