I can totally picture new languages as described doing well today if we had exponentially more single threaded processing power/watt at our disposal than we actually do.
I think there is still opportunity for languages that offer huge leaps in expressiveness w.r.t handling concurrency even at the cost of raw single threaded performance. Though I would not label a language like that as "slower" as it'd allow us to actually make much better use of our computing resources than we reasonably could today without blowing through most of our complexity budget, resulting in faster programs in practice.
I feel like that one simply missed the mark. If anything, there was greater emphasis towards languages that were more performant, even if less productive for banging out a working product.
Remember, IE6 was the dominate browser for years, so very slow JavaScript was the norm.
For example round about 2008 I made a client-side pivot table for a web app, tried it in JavaScript first, even trivial tables would take 10 seconds despite all the optimization I tried (in IE6, IE7 + Firefox).
I rewrote it using XML + XSLT and it was instant. But it was truly gnarly code.
I dont think this happened?