Top Hackernews posts from statmodeling.stat.columbia.edu

1. “AI promised to revolutionize radiology but so far its failing” (statmodeling.stat.columbia.edu)
   405 points by macleginn almost 55 years | 387 comments
2. Open AI gets GPT-3 to work by hiring an army of humans to fix GPT’s bad answers (statmodeling.stat.columbia.edu)
   345 points by agnosticmantis almost 55 years | 139 comments
3. Chief scientist of major corporation can’t handle criticism of the work he hypes (statmodeling.stat.columbia.edu)
   337 points by Tomte almost 55 years | 410 comments
4. If you think psychological science is bad, imagine how bad it was in 1999 (statmodeling.stat.columbia.edu)
   275 points by ruaraidh almost 55 years | 228 comments
5. The claimed effect size is about a zillion times higher than is plausible (statmodeling.stat.columbia.edu)
   275 points by luu almost 55 years | 61 comments
6. ML is not that good at predicting consumers' choices (statmodeling.stat.columbia.edu)
   251 points by macleginn almost 55 years | 217 comments
7. The immediate victims of a con would rather act as if the con never happened (statmodeling.stat.columbia.edu)
   230 points by Tomte almost 55 years | 140 comments
8. Bayesians moving from defense to offense (statmodeling.stat.columbia.edu)
   213 points by Tomte almost 55 years | 135 comments
9. Noted study in psychology fails to replicate, crumbles with evidence of fraud (statmodeling.stat.columbia.edu)
   205 points by luu almost 55 years | 102 comments
10. They say that stocks go down during the day and up at night (statmodeling.stat.columbia.edu)
    197 points by kuhewa almost 55 years | 146 comments
11. I disagree with Geoff Hinton regarding "glorified autocomplete" (statmodeling.stat.columbia.edu)
    194 points by magoghm almost 55 years | 265 comments
12. Do simpler machine learning models exist and how can we find them? (statmodeling.stat.columbia.edu)
    191 points by luu almost 55 years | 116 comments
13. Every modeler is supposed to be a great Python programmer (statmodeling.stat.columbia.edu)
    189 points by jeffreyrogers almost 55 years | 126 comments
14. “Chatbots: Still Dumb After All These Years” (statmodeling.stat.columbia.edu)
    187 points by LittlePeter almost 55 years | 162 comments
15. A baffling scale transform on a chart of university course selection trends (statmodeling.stat.columbia.edu)
    174 points by luu almost 55 years | 35 comments
16. A homework question in someone’s 11th grade statistics class (statmodeling.stat.columbia.edu)
    160 points by Tomte almost 55 years | 258 comments
17. High-intensity exercise, some new news (statmodeling.stat.columbia.edu)
    149 points by luu almost 55 years | 167 comments
18. Using Benford’s Law to Detect Bitcoin Manipulation (statmodeling.stat.columbia.edu)
    139 points by luu almost 55 years | 84 comments
19. Which one of these will be the biggest “unicorn” failure ever? (statmodeling.stat.columbia.edu)
    136 points by jeffreyrogers almost 55 years | 187 comments
20. I disagree with Turing and Kahneman regarding statistical evidence (2014) (statmodeling.stat.columbia.edu)
    129 points by acqbu almost 55 years | 77 comments
21. Bad stuff going down at the American Sociological Association (statmodeling.stat.columbia.edu)
    129 points by Tomte almost 55 years | 116 comments
22. Dean of Engineering at University of Nevada wrote a paper that’s bad (statmodeling.stat.columbia.edu)
    123 points by Tomte almost 55 years | 79 comments
23. EU proposing to regulate the use of Bayesian estimation (statmodeling.stat.columbia.edu)
    119 points by tosh almost 55 years | 105 comments
24. A successful example of "adversarial collaboration" (statmodeling.stat.columbia.edu)
    118 points by Tomte almost 55 years | 40 comments
25. Bayesian statistics and machine learning: How do they differ? (statmodeling.stat.columbia.edu)
    105 points by magoghm almost 55 years | 39 comments
26. The examples in the textbooks are not representative of real world problems (statmodeling.stat.columbia.edu)
    101 points by luu almost 55 years | 70 comments
27. Update on the fake story about the river laborers paying people to whip them (statmodeling.stat.columbia.edu)
    101 points by luu almost 55 years | 59 comments
28. “Is there a heuristic we might use to identify and flag questionable papers?” (statmodeling.stat.columbia.edu)
    99 points by pyinstallwoes almost 55 years | 103 comments
29. An accidental experiment that saved 700 lives (statmodeling.stat.columbia.edu)
    96 points by luu almost 55 years | 102 comments
30. Fields where it matters, fields where you can thrive on BS alone, and in between (statmodeling.stat.columbia.edu)
    93 points by johndcook almost 55 years | 121 comments
31. “A Headline That Will Make Global-Warming Activists Apoplectic” (statmodeling.stat.columbia.edu)
    91 points by luu almost 55 years | 72 comments
32. Not Frequentist Enough (statmodeling.stat.columbia.edu)
    89 points by luu almost 55 years | 53 comments
33. Feeling like a pariah, even when you’re not (statmodeling.stat.columbia.edu)
    80 points by luu almost 55 years | 67 comments
34. Refuted papers continue to be cited more than their failed replications (statmodeling.stat.columbia.edu)
    79 points by nabla9 almost 55 years | 34 comments
35. Only positive reinforcement for researchers in some fields (statmodeling.stat.columbia.edu)
    77 points by thetan almost 55 years | 52 comments
36. A sequence that we see all the time in the world of junk science (statmodeling.stat.columbia.edu)
    74 points by tomrod almost 55 years | 82 comments
37. ChatGPT4 writes Stan code so I don’t have to (statmodeling.stat.columbia.edu)
    70 points by luu almost 55 years | 61 comments
38. Using predictions from arbitrary models to get tighter confidence intervals (statmodeling.stat.columbia.edu)
    69 points by luu almost 55 years | 3 comments
39. Non-Standard Errors [pdf] (statmodeling.stat.columbia.edu)
    67 points by luu almost 55 years | 8 comments
40. The Bayesian Cringe (2021) (statmodeling.stat.columbia.edu)
    65 points by EndXA almost 55 years | 24 comments
41. Thinking fast, slow, and not at all: System 3 jumps the shark (statmodeling.stat.columbia.edu)
    64 points by martingoodson almost 55 years | 13 comments
42. Why I’m skeptical of “steelmanning” (statmodeling.stat.columbia.edu)
    61 points by luu almost 55 years | 122 comments
43. The piranha problem in social psychology / behavioral economics (2017) (statmodeling.stat.columbia.edu)
    60 points by Tomte almost 55 years | 16 comments
44. The formation and revision of intuitions (2023) [pdf] (statmodeling.stat.columbia.edu)
    53 points by luu almost 55 years | 8 comments
45. Pinker was right, I was wrong (statmodeling.stat.columbia.edu)
    52 points by Tomte almost 55 years | 55 comments
46. Analog computing and hybrid computing: The view from 1962 (statmodeling.stat.columbia.edu)
    50 points by luu almost 55 years | 3 comments
47. Relating t-statistics and the relative width of confidence intervals (statmodeling.stat.columbia.edu)
    50 points by luu almost 55 years | 13 comments