My Faults My Own

One's ponens is another's tollens.

IN WHICH Ross Rheingans-Yoo, a sometimes-poet and else­wise a recently-graduated student of Computer Science and Math, oc­cas­ion­al­ly writes on things of int­erest.

Reading Feed (last update: April 2)

A collection of things that I was happy I read. Views expressed by linked authors are chosen because I think they're interesting, not because I think they're correct, unless indicated otherwise.


Blog: Don't Worry About the Vase | On Automoderation -- Zvi concretizes much the the vague disease I was feeling around Automoderation, despite it being an eminently plausible approach to its design specification.


Blog: JeffTK | Slack tool: predict -- Note that Jeff's implementation is of a market mechanism that's not budget-balanced, and rewards marginal improvements of the "last price", rather than marginal improvements of the "current best price". I suspect that these design decisions have the net effect of denoising the signal of predicter quality.

Blog: Schneier on Security | New Gmail Phishing Scam -- "The article is right; this is frighteningly good."

Blog: Marginal Revolution | The Baffling Politics


Burn the Man's Books!

According MIT's Title IX Office, no-longer Professor Emeritus Walter Lewin acted in violation of the Institute's sexual harassment and misconduct policy while teaching an online MIT course open to the public. The Institute announced on Tuesday that it has stripped Lewin of Professor-Emeritus status, and will be removing videos of his physics lectures -- which have been called "legendary" -- from MIT OpenCourseWare and MITx.

I accept without question the reports that the charges were extremely serious and that "this wasn't a borderline case", and I agree with my current CS(@MIT) professor Scott Aaronson, as he writes in a recent blog post:

  • [S]exual harassment must never be tolerated, neither here nor anywhere else. But I also feel that, if a public figure is going to be publicly brought down like this (yes, even by a private university), then the detailed findings of the investigation should likewise be made

1 / 1