The scammy part was definitely making all of the claims as if they were already true. A lot of people sincerely believed it. My recollection was that it took months before some of this was fixed to explain that it wasn't available yet, and then more to explain that it hadn't actually been done yet.

There were some really crazy ones, like that it could convert entire codebases from C++ to V and then compile them some multiple of times faster and make them memory safe. How? I'd love to know.

The author to me seems like they are not attempting to scam people, but it's hard to ignore that they were collecting money, from people who thought they had already accomplished these things, because the person collecting it absolutely appeared to be claiming it.

Now doing that once and going "oops, my bad" would be one thing. There would be no controversy. However, over time it has proven to be a bit of a behavior pattern, which is confusing because the author does appear to be producing things that I think are interesting. Just... often buried under some hyperbole.