These small proof of concept projects just go to show how fundamentally important are projects such as GCC and LLVM, which considerably lowered the barrier to entry of monumental tasks such as developing a programming language that targets basically all platforms under the sun.
It is still very young and I don't yet see any advantages over Rust, except that they want to have out of the box interoperability with C++. Let's see where they are in 3 years or so.
All I see is another wannabe generic purpose and C++ (and Rust? Already?) replacement. I'm not saying they shouldn't try, but a lot have already and it's pretty tough. C alternatives like Zig and Nim don't seem to have taken off, Vala either.
Only glanced through the paper on mutable value semantics (MVS). Which appears to be the differentiator of this from Rust.
The claim is that MVS enables mutable operations safely without the complexity of Rust - but still high performance.
"Mutable value semantics (MVS) sits at third point in the
design space where both goals are satisfied and mutation is
supported, without the complexity inherent to flow-sensitive
type systems. The key to this balance is simple: MVS does not
surface references as a first-class concept in the programming
model. As such, they can neither be assigned to a variable nor
stored in object fields, and all values form disjoint topological
trees rooted in the program’s variables"
If only mutable value semantics are allowed, then how can you represent an object graph with circular references (let's say HTML DOM, for example) while still allowing modifying the objects?
Seems like this would be very difficult to work with in practice.