Comment of the Day: Rad Geek https://www.bradford-delong.com/2020/03/rad-geek-_the-infovores-dilemma_-in-circumstances-that-lead-to-a-high-risk-of-groupthink-and-overreach-its-a-r.html?cid=6a00e551f0800388340240a51a750f200b#comment-6a00e551f0800388340240a51a750f200b: 'Rad Geek said in reply to Ebenezer Scrooge: "A really expert journo collects an awful lot of signal with very little added noise, even if the originals are mentally accessible and noise-free .... a general-purpose journo in a specialized beat is trouble, even if the journo is smart and fair. ... A good lawyer or journo can manipulate symbolic knowledge far better than those who have the tacit knowledge..." I don't think we disagree about the features that make a really expert journalist very valuable on a well-defined beat. Sure: and under ordinary conditions in an informational ecosystem, that's sometimes a pretty good reason to read newspaper reports more than you read abstracts, etc. But what I'm concerned about is what happens in extraordinary conditions in the informational ecosystem, when—for example—the sheer volume, rapid-fire turn-around times, and tremendously wide scope of issues involved mean that a lot more "general-purpose" journos are suddenly put onto specialized beats, or onto writing generalist stories that really essentially involve complicated issues from a specialized beat. (Call that the demand-side worry.)...
...And, also, even expert journos on a specialized beat do rely on a lot of the institutional and conventional environment around research to have much chance of getting their reporting mostly right most of the time; but under extraordinary conditions the institutions and conventions are likely to shift subtly or break off rapidly (no doubt for perfectly good reasons)—for example, when the tremendously large volume of simultaneous research shifts the environment towards much more frequent rapid-fire sharing of preliminary results, very tentative findings, possible breakthroughs, etc., in huge volumes, on very short turn-around times and often with minimal opportunity for peer review or revision before they are widely disseminated. (Call this the supply-side worry.)
My point is, the more you go towards those kind of situations, and towards stronger cases of these two kinds of worries, the more trouble you can expect from exactly the trouble that you point out, and that this can have some characteristic effects (for example a lot of unwarranted confidence, false precision, errors of groupthink and closed-mindedness towards possible falsification, and in general an unusually elevated likelihood that you'll hit points of diminishing or negative returns, where consuming more information fails to make you better informed).
(The suggestion that there may be better substitutes, e.g., in reading abstracts, depends on an underlying conjecture that these tend to do a better job than reporters of doing things like explicitly stating the simplifying assumptions behind models, stating ranges of possible values in outcomes, acknowledging rival explanations and limitations on the precision or generalizability of findings, etc. And also that these statements are comprehensible to the person reading them. If I'm wrong about the substitutability claim -- either because the sources don't do those things very well either, or because even if they do the reader is ill-prepared to understand or make use of them -- then in either case, that may be a reason for that reader to consume less information under the circumstances and to be relatively more agnostic and less confident in their beliefs about the topics in question.)...
#commentoftheday #2020-03-27
