User blog comment:Wythagoras/My Turing machines/@comment-6768393-20131012123348/@comment-1605058-20131013115853

They can't know BB(30) as it is far beyond Graham's number. It is possible, though, that they could give an upper bound by analysing all halting machines and proving no of them leaves more than, say, \(f_{\varepsilon_4}(400)\) ones.