• Assistant professor, London School of Economics and Political Science • Persuasion, technology, experiments • benmtappin.comhttps://bsky.app/profile/benmtappin.bsky.socialhttps://bsky.app/profile/benmtappin.bsky.social/post/3mi2vp6wnxs2tI wonder how these rates of sycophancy (and their effects) compare against realistic counterfactuals like talking with one’s close friends or spouse. [contains quote post or other embedded content]27 Mar 2026 19:57 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mi2vp6wnxs2thttps://bsky.app/profile/benmtappin.bsky.social/post/3mhn7mv65qs2vThe synth backing track takes the vibe to an unexpected place 🥲 [contains quote post or other embedded content]22 Mar 2026 09:18 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mhn7mv65qs2vhttps://bsky.app/profile/benmtappin.bsky.social/post/3mhljflzg3s24Closing out the teaching semester this week with one of my favourite topics. Sadly its pedagogy was forever and unforgivably mar'd by the truly worst naming convention of all time...21 Mar 2026 17:07 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mhljflzg3s24https://bsky.app/profile/benmtappin.bsky.social/post/3mhkpsizogk2nYes. Then if you push on this the claim becomes “okay it may not cause but it’s still useful because it predicts”. But then it’s like if predictive accuracy was your goal the design and analysis should be different e.g., CV + more predictors. I’ve been fully Westfall & Yarkoni-pilled on this point. [contains quote post or other embedded content]21 Mar 2026 09:29 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mhkpsizogk2nhttps://bsky.app/profile/benmtappin.bsky.social/post/3mgoyeawv4k2fThe character of Per’s posts increasingly remind me of when an LLM talks to itself (not derogatory) [contains quote post or other embedded content]10 Mar 2026 08:48 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mgoyeawv4k2fhttps://bsky.app/profile/benmtappin.bsky.social/post/3mftcnsmivk2bI too was very glad to see this! But I feel like the whole episode bodes badly for the future. It’s not sustainable to rely on the CEO of a private company to act against their financial self-interest in order to curtail high-risk AI deployment (here mass surveillance and fully autonomous weapons). [contains quote post or other embedded content]27 Feb 2026 08:37 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mftcnsmivk2bhttps://bsky.app/profile/benmtappin.bsky.social/post/3mfqselalv22vReposting for visibility. Many researchers still appear oblivious to this fact, which is terrifying! It should be included in every experiment design 101. [contains quote post or other embedded content]26 Feb 2026 08:41 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mfqselalv22vhttps://bsky.app/profile/benmtappin.bsky.social/post/3melaeuaq2c22Felix and friends looking closely at the details so you don’t have to 👌👇 [contains quote post or other embedded content]11 Feb 2026 10:10 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3melaeuaq2c22https://bsky.app/profile/benmtappin.bsky.social/post/3mebw72dfok2dIt’s that time of year again07 Feb 2026 17:14 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mebw72dfok2dhttps://bsky.app/profile/benmtappin.bsky.social/post/3mddzqacuu22eI land somewhere between this and the OP. Evaluating the quality of the methods often requires fully understanding the research question and estimand. And that usually requires reading the intro. (Disclaimer: but even then it’s no guarantee😭 cf. https://www.the100.ci/2024/08/27/lets-do-statistics-the-other-way-around/ @dingdingpeng.the100.ci) [contains quote post or other embedded content]26 Jan 2026 19:57 +0000at://did:plc:k3fgjtb4ooplcsfo5wh3j574/app.bsky.feed.post/3mddzqacuu22e