Even the clean-room isolation that Phoenix went through isn't legally required, it just makes nuisance lawsuits more difficult. BSD prevailed over UNIX System Laboratories, in their reimplementation of Unix, despite having directly worked with the source code.
It turns out that's not exactly the case! See, e.g., Goldstein v. California, 412 U.S. 546 (1973). Before 1978, state (often common law) copyright used to cover a lot of pre-publication works, and until 2018 (when the federal law was amended) state copyright law covered pre-1972 sound recordings, and state copyright still covers obscure things like post-mortem moral rights in visual art or rights to "unfixed" works. See 1 Nimmer on Copyright §§ A.02 & 2.02. Other forms of intellectual property (trade secrets, rights of publicity) remain mostly creatures of state law, and some states also have trademark systems.
Coreboot is debatable for this, it's fine in the sense that nobody is going to come after you for it, but legally you're not doing a clean room implementation, you're looking at the original and creating a new functional replacement, which is fundamentally different to the Phoenix BIOS clone, and not in a good way.
But as I said, nobody is going to come after you for it so...
Interesting legal question: if Claude reverse engineers the original and writes a spec, and ChatGPT implements the spec without seeing the original, is that a clean-room implementation? Asking for a friend with a trillion parameters
But honestly this is going to have to be tested in court. Also, IANAL.
...yes we are? After all, that's how the whole IBM PC-compatible industry started.
https://en.wikipedia.org/wiki/Phoenix_Technologies#Cloning_t...
AFAIK the later Thinkpads including this one uses a Phoenix BIOS, so it's amusing to see the circularity of how things turned out; and continuing on that path, Phoenix sold its BIOS business to Lenovo a little earlier this year.