Philip E. Tetlock, the Canadian-American politologist, is someone who, as they say, is "widely-known in narrow circles." Unlike many other figures typed here, he's a professor without an exciting up-and-downs biography, working on political judgement and forecasting. All his professional life, he's been working on a research program about the conditions under which knowledge of complex systems is possible at all, attempting to simplify it into a useful form (as far as the tertiary dichotomy of process-result, aka complication-simplification is concerned, LII tends to invert to complication in many ways, but keeps the desire to untangle a complicated mess to relatively simple principles).
He didn't seem to have any inherent deflationary impulse of "uncovering" experts. His conclusion in Expert Political Judgment: How Good Is It? How Can We Know?, that the average expert is roughly as accurate as a smart second-year university student (and some well-regarded experts are only slightly better than a dart-throwing chimpanzee), is stated without much schadenfreude, and the book still insists that good forecasting, while difficult, is still possible, rejecting total nihilism in regard to predicting the future. His insistence on probabilistic thinking as the core competence rules out LSI as a type — nobody with hard-boundary thinking writes the book which rejects the "hedgehog" approach of not tolerating any vagueness, applying only one approach, and advocates the "fox" method of being modest, self-critical, drawing on many perspectives and sceptical of grand theories. The last part may point to a Gamma NT accent, but the ultimate project of building a grand theory of prediction itself is still more Ti. There's also a small LIE accent — Good Judgment Inc. is a commercial spinout, which is a touch more appetite for the arena than pure LII would predict. Note that his "foxes" are actually better at predicting events, they don't only come out as better because of their refusal to commit!
His alternate history impulse is also on the border between LII and ILI, but leans more towards LII, since the general object of his interest here are is the general historical laws. A counterfactual is a possibility raised in order to test what was necessary — the branch illuminates the structure of the actual world. While not a totally reliable way of distinguishing these closest "quasis", your canonical ILI is still a touch more interested in dynamic storytelling and Game of Thrones-like "intrigues in royal chambers" than in global laws as such (at least, someone with very strong ILI traits, but with a distinct lack of LII traits is like that). In fact, his Counterfactual Thought Experiments in World Politics preceded his book on prediction judgement, so the general question about historical necessity was the starting point and the pundit study was an application of it. Later he wrote Unmaking the West: What-if Scenarios that Rewrite World History, where he defended counterfactual history as a valuable research tool.
He's still doing the same intellectual adversarial collaboration of many perspectives at on AI risk, running structured dialogue between doomers and sceptics to find cruxes resolvable by 2030. In fact, he is well-liked by the AI-risk obsessed Rationalists sub-culture, which is harder to type (there are ILE's and LII's there, but also EIE's, ILI's, and who knows who Yudkovsky himself is, might be a rare case of "ILE/LSI"?) He himself isn't really a participant in it, however. They'd be better served by his findings that long-range forecasting on complex systems is far less certain than these forecasters believe and degrates towards chance far quicker than you'd think. Ultimately, he's an academic answering a general question who happens to have produced instruments other people found useful, not a member of a movement.
Comments
Post a Comment