Ben Alfred (@Ben_Alfred_) header
Ben Alfred (@Ben_Alfred_) profile picture

Quant sociobiology ∨ Psychometrics ∨ Stats || Entropy observer || Mischl. I. Gr.

South Africa
Joined July 2021
Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)Photo by Ben Alfred (@Ben_Alfred_)
Actually, the low IQ, beastly act was carried out by the low IQ, criminal janitor. In fact, it was one of the least convincing ‘dindu nuffin’ pleas of all time, and merely skimming Herr Wittmayer’s article can convince anyone not ideologically committed to denying Conley’s guilt. mischrev.substack.com/p/so-w…
How the ADL was founded to protect a Jewish-American child rapist and murder:
23
7
4
73
5,821
“Appeared to try to”???? This is DISGUSTING false reporting from the BBC. You had a reporter on the ground. I and many others shot footage of antifa holding the picket lines ensuring nobody could get in past 5PM. How dare you lie to the public!
23
503
11
3,357
143,040
Fable, Grok, and Astra are all Ashkenazi Jews.
60
147
32
1,420
204,822
Human slop. Trying to find an alternative explanation to dark matter/energy is important, but their theory is ultimately bad and looks worse the closer you inspect it. 1. Its out-of-sample predictive performance was inferior to the currently accepted model. 2. They make post hoc additions or alterations to the model for every observation they need to explain, which is why it “passes tests”. Basically, it’s overfit, which explains point 1. For example, the model produced a 6σ discrepancy between estimates of the same parameter, and Gupta’s response was to make a correction whose only justification was to resolve the discrepancy (refer to red square). Like string theorists, their approach is antithetical to science.
5
1
16
1,540
Ben Alfred retweeted
It's crazy, harmful, and it'll surrender the future to China But it does seem to be what the electorate wants
Top of the Democratic agenda on day 1 needs to be a pause on advanced AI development, regulation of this out of control industry, and a ban on artificial superintelligence.
134
153
18
3,013
68,362
I have a new substack post on predicting opinions, politics and IQ with AI. Across a suite of about ~100 batteries, AI has near perfect signal extraction capabilities.
2
7
55
2,583
Lyman Stone pointed out the engagement variable here is g-loaded, so I decided to check how strongly support for political violence relates to IQ. It probably explains some of the effect: a 130 IQ person is 10x less likely to think political violence is ok than an 80 IQ person.
Reminder: When you ask people who aren't engaged with a survey if they think political violence is justified, large numbers say yes. If you give them incentives to engage with the questions, much smaller percentages support political violence. MOST surveys are not engaging.
14
12
5
143
14,212
Now to be fair, the total proportion who thought hitting a protestor was acceptable was lower than recent polling for political violence support, and reactive 'hitting' is not the same as organized political violence as a means to some end. Still, I thought it interesting enough to share.
7
465
Now to be fair, the total proportion who thought hitting a protestor was acceptable was lower than recent polling for political violence support, and reactive 'hitting' is not the same as organized political violence as a means to some end. Still, I thought it interesting enough to share.
1
1
47
*Lyman. Is there no editing option for posts anymore?
1
52
4 million years of progress on the hominid frontier: 1) Humans got spikier, not rounder. 2) Vibe charts, maths, verbal spew: way past Australopithecus level. 3) Climbing trees, having a penis bone, nose sensitivity: far inferior. Basically no middle ground guys, “generalisation” is a complete meme.
12 months of progress on the jagged frontier: 1) AI got spikier, not rounder. 2) Maths, fluid reasoning, expert work: past human level. 3) There is almost no middle ground. Once a capability crosses human level, it often quickly becomes 100x+ cheaper than a human. 4) Cost keeps decreasing after capability increases stop. Language and knowledge look saturated but price keeps falling. 5) The gaps are not a budget issue. Human-level memory, self-monitoring and long-run planning are not for sale at any price. AI is now a superhuman specialist that is sub-human at running itself.
21
33
5
538
25,599
My hypothesis seems to have been correct. Mechanics knowledge remains g-loaded, but its intercept has shifted by about two standard deviations. Despite my test-taker sample being above average, so far no one has scored above the 1960 17yro male average. Also, I bought a real domain: projecttalenttests.com/
Curious what 'creativity' measured in Project Talent? Or what the 2D, 3D, and abstract reasoning subtests were? I digitized the full battery, with automated IRT scoring, so you can check them out and also see where you stand relative to 1960s teenagers: project-talent-questionnaire…
7
10
1
126
10,562
People just don't directly interact with mechanical systems anymore. Things are now too complex to mess with, and when they break down repair is left to experts. So whilst a modern person who knows how to change an engine's compression ratio is smarter than someone who doesn't on average, the average smart person does not know how how to change the compression ratio of an engine.
2
3
38
1,074
The opposite is true for abstract knowledge. Familiarity with complex physical subjects has declined, but mathematical information has improved (modern people score 10 points higher than similarly able 1960s men).
2
32
720
If we plot Jews, White gentiles, and Blacks separately, the result looks even more robust because we see essentially the same pattern in all six groups. Also cool to see how group g differences and ability profiles shift the intercepts in all cases.
Here's a curious one: processing speed is nonlinearly related to intelligence in Project Talent. This persists whether or not you penalize wrong answers/punish inaccuracy.
4
1
23
1,413
Could some of this be Spearman's law of diminishing returns? If the curve merely flattened then that might be plausible, but as you can see there is more than just flattening going on. There is also the fact that the other g-loaded factors are all linearly related:
1
9
623
If you’d like to see each of the subtests each factor comprises, they are all on here.
Curious what 'creativity' measured in Project Talent? Or what the 2D, 3D, and abstract reasoning subtests were? I digitized the full battery, with automated IRT scoring, so you can check them out and also see where you stand relative to 1960s teenagers: project-talent-questionnaire…
6
508
Here's a curious one: processing speed is nonlinearly related to intelligence in Project Talent. This persists whether or not you penalize wrong answers/punish inaccuracy.
It's genuinely difficult to find nonlinear patterns in psychology. Looking around at the various correlations, these are the first 4 I found: 1) English vocab and age 2) US civis and age (very slight) 3) US general knowledge and age 4) ??? taketest.xyz/correlations
3
6
1
75
6,774