This sounds like AI assisted bunk. All their supposed evidence (tenuous at best) would also apply to developing swift.

The example code is identical to the Swift code, one just isn’t letting the compiler discern type. I think the author doesn’t actually understand what they’re talking about.

Apple has spent a lot of effort making Swift more secure , especially with Swift 6. And a lot of effort trying to move C++ codebases to Swift. So this sounds highly unlikely as a premise.

Apple spent a lot of effort adapting or creating languages, starting with Clascal/Object Pascal (or Woz' SWEET16 as an alternative high-level assembler-like intermediate language and 16-bit virtual machine) over Objective C (bought from Brad Cox and Tom Love and significantly extended/changed over the years) to Swift. They also used these languages to implement critical parts of the system software, e.g. for the Lisa or the NeXTstep ObjC-based driver framework.

So this speculation has quite some credibility - but, as the author states, I think this is still in an early stage and will take quite some time to become mature.

There are other people designing their own system-level language and OS, e.g. Drew DeVaults Hare and the Helios microkernel and Bunnix Unix clone based on Hare - this is a single person effort, so given Apple's resources, this is definitely feasible.

The rewriting: quite believable, but it will take a looooooooooooooong time, and their goals may change looooong before they get there.

It not being Swift: unlikely, IMO. They’re spending lots of resources on Swift-C++ interoperability, added lots of low-level stuff to Swift, and advertise Swift for embedded use on Swift.org,

I don’t see what another language could bring. I also do not see them rename embedded Swift to something else; it’s hard enough for them to make Swift get traction outside their ecosystem already.

Also note that the “New Apple Language (hypothetical)” sample already is valid Swift code.