Neocon propaganda outlet busted!
Plus: Deep-fake defense system, mind-reading AI, climate-change do-over, Jeffrey Sachs comes on the pod, and more!
Nearly a year ago, this newsletter tried to call attention to the unfortunate influence that the Institute for the Study of War—an arms-industry-funded think tank with neoconservative roots—was having on American journalism. ISW, by providing frequent and authoritative-sounding updates on the Ukraine war, complete with detailed battlefield maps, had become a go-to source for MSM reporters. As a result, media reports about the war tended to overstate Ukrainian battlefield successes, understate Russian ones, and in various other ways cloud Americans’ understanding of the war.
Well, it’s taken a while, but the conventional wisdom is starting to catch up with this assessment. This week a number of respected military analysts complained on Twitter about ISW, and their complaints reached a kind of critical mass.
Shortly after ISW said this week’s drone strike on the Kremlin was “likely” a false-flag attack staged by Russia, intelligence analyst Nathan Ruser dismissed this view as a “hunch” and said the think tank was “becoming a major problem in the media ecosystem.” Ruser lamented that this “(bad) opinion” rendered by ISW would be treated “as fact by many journalists who print what they say verbatim.”
Veteran journalist Neil Hauer, citing the same ISW assessment, wrote, “This ISW update is absolutely laughable. Poorly argued, nonsensical, getting basic facts wrong and asserting unfounded theories as fact.”
Two days earlier, Hauer had flagged a dubious ISW claim of territorial advance by Ukrainian forces and written that “there is a fascinating ecosystem that has developed where ISW (and the social media interns who write the UK MOD [British Ministry of Defense] 'daily update' tweets) claim some nonsense speculation as if it were facts, and then it gets laundered into 'confirmed info' by legacy media reporting it.”
ISW has long been a subject of derision by pro-Russia Twitter voices, but Hauer and Ruser are pro-Ukraine, and they have big Twitter followings. Worse still, from ISW’s point of view, Michael Kofman—who in the course of this war has become America’s most prominent and respected expert on the Russian military—gave Ruser’s long thread critiquing ISW his seal of approval, deeming it “constructive criticism.”
To which Hauer replied: “This would all be less of a problem if 80% of English major media didn’t cite ISW constantly and give credence to their speculative nonsense.” That number isn’t much of an exaggeration. At last check The New York Times, The Wall Street Journal, and The Washington Post, not to mention lots of equally famous if less elite outlets, were routinely quoting ISW output.
And that output, ISW assured us this week, is solid. Stung by all the criticism, the think tank issued a lengthy statement in response, saying, among other things, that “ISW applies industry-leading tradecraft.” If “industry” is meant in the sense of “military industrial complex,” that’s hard to argue with.
Attention NZN Members! This week we bring you two paid subscriber perks (and three if you count the Parrot Room, Bob’s after-hours discussion with arch-frenemy Mickey Kaus):
Early access to Bob’s conversation about US foreign policy with influential economist and public intellectual Jeffrey Sachs. You can watch and/or listen to the episode here (or find it, and the other perks, in your members-only Nonzero podcast feed).
Bob’s conversation with NZN staffer Andrew Day about this week’s Earthling items.
Just when you thought you’d heard about all possible threats to humankind that AI could eventually pose, the New York Times brings us this headline: “AI Is Getting Better at Mind-Reading.”
Scientists from the University of Texas devised a “large language model” that, they say, can decode human subjects’ private thoughts by analyzing their brain scans.
In one test of the model, researchers asked participants to silently tell themselves a story while lying inside an fMRI scanner. The AI then translated the fMRI output into language. Here’s how, in one case, the participant’s account of the unspoken tale compared with the AI’s account of it:
Not bad, all things considered. But don’t expect to see this technology at your local thought police station anytime soon. Training the models takes a long time, and models trained on one person can’t (for now, at least) be used on another. So if somebody asks you if you’d like to spend a few weeks in a brain scanner, either say no or spend those weeks lying about what you’re actually thinking.
Are we about to see significant economic impact from the new generation of AIs? This week, two developments suggested the answer is yes: