Is MIT Dissing Its CompSci Grads, Maybe the Google, IBM, and Possibly AI in General

August 3, 2021

Hey, what does one expect from an outfit which did some Fancy Dancing with alleged human trafficker Jeffrey Epstein? I don’t expect much. It was amusing to me to read “Hundreds of AI Tools Have Been Built to Catch Covid. None of Them Helped.” Why am I laughing? Well, there are the MIT spawned smart software systems populating architecturally disappointing structures in the Boston area. There is also the really nifty teaming with IBM Watson (yep, the smart software systems which is less exciting that RedHat when it comes to tickling shareholders’ fancies). Watson, as you may recall, is allegedly the first artificial intelligence system to be placed on a patient trolley and bustled out of the emergency room exit by a clutch of cancer docs.

The referenced write up makes clear that Covid is either smarter than the smartest people in the world, or the smartest people in the world are dumber than their résumés suggest. The truth, I admit, might be somewhere in the middle of tenure squabbles, non-reproducible results, and good old PT Barnun malarkey.

The write up states:

The AI community, in particular, rushed to develop software that many believed would allow hospitals to diagnose or triage patients faster, bringing much-needed support to the front lines—in theory. In the end, many hundreds of predictive tools were developed. None of them made a real difference, and some were potentially harmful.

Yo, what’s this harm thing? Like Google’s brilliant progress on solving death, the hubris and rah rah about what a PhD demonstration implies, and what those Rube Goldberg constructions of downloadable code deliver are quite different.

The write up drags a reader through case examples of baloney. The write up documents failure. Yep, F. Failure for whiz kids who never experienced a set back which a helicopter mom or proud PhD mentor couldn’t address. A phone call from donors like Mr. Epstein probably helped too.

So the big question is posed by the estimable MIT cuddled write up: What went wrong?

What’s the answer?

Guess what. Lots. Bad data, wonky algorithms, statistical drift, grant crazed researchers.

This is a surprise?

Nope. In a nutshell, the entire confection of smart software’s capabilities is deconstructed in my opinion:

In a sense, this is an old problem with research. Academic researchers have few career incentives to share work or validate existing results. There’s no reward for pushing through the last mile that takes tech from “lab bench to bedside,”.

Should I bring up Mr. Epstein’s penchant for bedside activities. History will have to judge which is the more problematic social behavior.

Stephen E Arnold, August 3, 2021

Comments

One Response to “Is MIT Dissing Its CompSci Grads, Maybe the Google, IBM, and Possibly AI in General”

  1. Watson AI vs the Pandemic - Long Life and Health on August 4th, 2021 6:25 am

    […] some, this does not come as a surprise at all. As Stephen E. Arnold of Beyond Search […]

Got something to say?





  • Archives

  • Recent Posts

  • Meta