[hga]

Thanks for that link to his criticism of institutional CS, it gets me firmly into territory where I can apply the Gell-Mann Amnesia principle. Upon which I find the thesis sorely lacking, if you accept the principle as discussed in the comments that it's OK for research to be "wasteful" as long as "1%" of it turns out to be useful, especially in the long term (e.g. I do not accept that all interesting computing is going to limited to the context of the current context of the cloud and supercomputers masquerading as smart phones ("mobile"; I started my computing career in 1977, when the 90 MHz Cray 1 was the pinnacle of number crunching, although I have to confess that I don't know the 64 bit floating point performance of typical smartphone ARM CPUs)).

More specifically, his criticism of Haskell seems to be misplaced by his criteria of developing useful software, if you accept that the seL4 microkernel is useful, which I gather it is, otherwise General Dynamics et. al. are wasting money. I can't tell, it's perhaps a bit early to get a list of hardware using it, but previous L4 versions have been used in billions of Qualcomm chips and apparently all iOS devices.

And the related academic Barrelfish OS researchers seem to me to be doing something useful, and the languages they are using are C, with various bits of that generated by Haskell (e.g. hardware descriptions -> C).

It's a pity that Urbit now has no chance of greater success, the SJWs of computing going so far as to say it "has neoreactionary politics hard-coded into the network layer" (https://news.ycombinator.com/item?id=9675512), which is obviously worse than the "monarchy" of superuser vs. user.