Interesting... I think he got a lot of the details wrong. Lisp didn't do well, but for reasons other than surmised.
Does it matter? What was important then isn't very important now. We need better paradigms for programming networks of agents, neural nets and statistical computing with algorithms that run well on thousands of threads.
For a standard to be useful it needs to make some guarantees about interoperability, that if a conforming implementation meets a conforming instance it will work as expected.
Compare Common Lisp to the web designer's toolkit: HTML + CSS + HTTP + JavaScript + TLS + URL + SVG and throw in XML and JSON.
Medley may have a niche or two of three of users who value the relative simplicity and transparency (looking at initials to discover that the horrible mess was last edited by me).