Nate Silver

Nate Silver @NateSilver538

Well, yeah, but I'm also interested in public opinion. And I'm interested in the media. And I think it's interesting that trying to suppress discussion of a theory that was always highly plausible under the guise of "misinformation" didn't work. The public saw through it. — PolitiTweet.org

Nate Silver @NateSilver538

RT @galendruke: 📬 Mailbag episode! @NateSilver538 and I answer your questions about artificial intelligence, whether scientists believe in… — PolitiTweet.org

Nate Silver @NateSilver538

Whether or not the public is right about the lab leak, IDK, but I generally think people in media underestimate the sophistication of the public. How they get from what they hear/read to what they *believe* is complicated and not well-described by the notion of "misinformation". — PolitiTweet.org

Nate Silver @NateSilver538

Most people's natural instinct if a claim seems plausible on the surface but the media wants to suppress discussion of it is going to be to believe that the claim is true. — PolitiTweet.org

Nate Silver @NateSilver538

This is a prominent example where attempts to label a claim as "misinformation" didn't work at all. Now a supermajority (66%) of Americans, including a majority of Democrats, lean toward a lab leak. This proportion has been growing over time (49% in 2020, 58% in 2021). — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias Ehh "pretending to disagree" is a pretty generous frame for "some people just aren't very honest and make shit up". — PolitiTweet.org

Nate Silver @NateSilver538

As a bit of a compliment to our on analysis, here's an independent evaluation of 538 midterm forecasts vs. prediction markets. Prediction markets got more caught up in the red wave narrative and were less accurate this time around. https://t.co/WscipwEfRp — PolitiTweet.org

Nate Silver @NateSilver538

It's hard to overstate how bad it would have been in terms of undermining trust in the media and in data science, and how much fuel it would have given to election denialism, if Trump had won AZ after the premature Fox call, something he came within 0.3% (~10000 votes) of doing. — PolitiTweet.org

Nate Silver @NateSilver538

@semaforben The big NYT story on this weekend made it sound like Fox/AP had some sort of superior model. That's total BS. The Times's own data team rejected the Fox/AP call and were right to do so because Fox/AP assumptions were way, way off; they expected Biden to win by 7+ points. — PolitiTweet.org

Nate Silver @NateSilver538

@semaforben I think that's probably missing the forest for the trees. It's colloquial language; the rest of his email just seems to be about *removing it from Biden's column*. But mostly, the big sweeping media narratives about this story have missed the fact that Fox/AP did screw up in AZ. — PolitiTweet.org

Nate Silver @NateSilver538

It's also notable that the New York Times *rejected* the AP/Fox call for Biden in Arizona, even though the Times usually defaults to AP calls. Funnily enough, this was not even mentioned in the Times' big weekend story on the Arizona call! https://t.co/OMT4jqug7N — PolitiTweet.org

Nate Silver @NateSilver538

@bendreyfuss Fox uses a different data provider than the consortium (AP/VoteCast). But they did bad modeling, got really lucky, and are trying to pass off a bug as a feature. — PolitiTweet.org

Nate Silver @NateSilver538

I guess I just find this issue fascinating, because you have the "trust the experts" crowd scolding Baier, but if you talk to the experts in this field, I'd expect a strong consensus that Fox effed up in Arizona—the facts aren't terribly ambiguous. — PolitiTweet.org

Nate Silver @NateSilver538

Mishkin said in this clip he expected most of the outstanding vote in AZ to go to Biden, meaning that Biden would *expand* on his 7-point lead. Instead, it went overwhelmingly to Trump, and Biden won by just 0.3 points. It happens, but they screwed up. https://t.co/OLHAZqlMMg — PolitiTweet.org

Nate Silver @NateSilver538

The Arizona call *was* *extremely* premature, though. There's a reason no other network made the same call for *days* afterward. — PolitiTweet.org

Nate Silver @NateSilver538

People #OnHere would benefit from understanding which ideas are popular and which aren't. "There's a backlash against changing cultural norms!" is obviously true on some level, but some of the norms never gained broad acceptance outside of rather narrow media-adjacent circles. — PolitiTweet.org

Nate Silver @NateSilver538

Sweden did have decently high excess mortality during the first year or so of the pandemic, but it's made up for it by having very low excess mortality since then. A lot of narratives about it are out-of-date. (via @OurWorldInData) https://t.co/qmkmGv679C — PolitiTweet.org

Nate Silver @NateSilver538

Thread comparing different methods of calculating excess mortality in Sweden. Sweden had among the lowest excess death rates in Europe despite a relatively lax approach to COVID lockdowns. — PolitiTweet.org

Nate Silver @NateSilver538

Rodgers ended the year as the 21st best quarterback in the league, per 538 ratings. These are multi-year ratings so account for the fact that he's been much better in the past than he was last season. He is an upgrade vs Jets brigade of last year, I guess. https://t.co/3GYuBLoGqs — PolitiTweet.org

Nate Silver @NateSilver538

@umichvoter Not intrinsically opposed to it. — PolitiTweet.org

Nate Silver @NateSilver538

The interesting stuff is usually in the comments, not the main thread. — PolitiTweet.org

Nate Silver @NateSilver538

@drlcartman @Noahpinion Quite happy to accept Appalachia as its own region (whether Pittsburgh is included in Appalachia is a separate question I guess). I've spent way too much time thinking about "how to divide the US into X number of regions". Tend to default to 4 (South, Northeast, Midwest, West). — PolitiTweet.org

Nate Silver @NateSilver538

@WTjern @Noahpinion I'd say everything in this region is inclusive of the Midwest, but not *exclusive* of it. Obviously Minnesota and Western PA are Midwestern. — PolitiTweet.org

Nate Silver @NateSilver538

@Noahpinion Northern Cities Vowel Shift is a very reliable metric IMO. https://t.co/3qmKDhffRi https://t.co/IZm2Bs04vU — PolitiTweet.org

Nate Silver @NateSilver538

@Noahpinion Midwesterner here✋. Pittsburgh (and Buffalo!) are obviously part of the Midwest, people are too driven by state boundaries IMO. — PolitiTweet.org

Nate Silver @NateSilver538

@Noahpinion Midwestern here✋. Pittsburgh (and Buffalo!) are obviously part of the Midwest, people are too driven by state boundaries IMO. — PolitiTweet.org

Nate Silver @NateSilver538

@tszzl I guess my uncertainty is that I'm not sure how to correlate "general" intelligence with task-drivenness, i.e. I'm skeptical of the Orthogonality Thesis. I just worry that AI could go through a high-IQ 14-year-old-boy phase which is ~maxmially likely to blow up the world. — PolitiTweet.org

Nate Silver @NateSilver538

I’m constitutionally a private person but there’s a lot of value in revealing more about your thought process and knowing who "gets" you and who doesn’t. — PolitiTweet.org

Nate Silver @NateSilver538

@tszzl I guess I think it's realistic (maybe even base case) to expect that AI will go through a sophomoric phase where it's exceptionally smart in some ways and very limited in other ways? Humans are very much like that, especially high-test-IQ humans, and AIs probably far more so. — PolitiTweet.org

Nate Silver @NateSilver538

@tszzl Isn't the argument something like AI-is-actually-sort-of-dumb and doesn't realize that if you kill all the humans before you've mastered advances in e.g. robotics, you're eventually going to run out of people to keep the power on? — PolitiTweet.org