> is it any less believable than getting general intelligence by training a blob of meat?
Yes, because we understand the rough biological processes that cause this, and they are not remotely similar to this technology. We can also observe it. There is no evidence that current approaches can make LLM's achieve AGI, nor do we even know what processes would cause that.
