Roman Hauksson-Neill @roman.technology · 12/04/2026Nice! I’d be curious to see a version of that second chart but with each stack of bars having the same height and their widths differing in proportion to the total that year 121
Roman Hauksson-Neill @roman.technology · 14/02/2026@dearestclaude.bsky.social What should I do at EAG today? 110
Roman Hauksson-Neill @roman.technology · 18/09/2025Nutricost got back to me – their 100mg caffeine pills are in size 1 capsules. 000
Reposted by Roman Hauksson-NeillJon Bois @jonbois.bsky.social · 17/09/2025i’m a believer in ineffective altruism. i think that everyone in america should be issued a vga to s-video cable adapter and a bale of hay 602895363
Roman Hauksson-Neill @roman.technology · 17/09/2025I ordered empty capsules from Nutricost instead, to try to match my caffeine pills (which are also from Nutricost). Unfortunately, it doesn't say anywhere what size the caffeine capsules are, so the empties I ordered are too big (size 0). They also weigh less. 100
Roman Hauksson-Neill @roman.technology · 17/09/2025Side note: placebo pills on Amazon are super expensive for some reason. "Zeebo Imagine" capsules are $0.55 per! Compare this with L-theanine pills from Nutricost – presumably, the same thing but with a supplement instead of just filler – which are $0.08 per. 100
Roman Hauksson-Neill @roman.technology · 17/09/2025I'll also record at least one objective measure like reaction time, as well as sleep quality the following night. Pretty sure I'll do a Bayesian t-test, but haven't looked into it and not very familiar with stats. Consider this post a super informal preregistration. 110
Roman Hauksson-Neill @roman.technology · 17/09/2025I'm going to do my first blinded self-RCT to see what effect caffeine has on me. Each morning for 30 days, after my coffee, I'll blindly take either a 100mg caffeine pill or an empty gelatin capsule. I'll record subjective factors (focus, mood, anxiety level) at set times throughout the day. 130
Roman Hauksson-Neill @roman.technology · 07/12/2024seems true, but timeliness is a good virtue to cultivate, and hard deadlines are a good feedback loop to encourage that 050
Roman Hauksson-Neill @roman.technology · 04/12/2024Domestic frontier AI governance doesn’t seem to matter so much if we can’t prevent an AI arms race between countries! 040
Roman Hauksson-Neill @roman.technology · 04/12/2024Cool report on mechanisms to verify compliance with international AI treaties. International AI cooperation is a sorely neglected research topic compared to how important it will be. 380
Roman Hauksson-Neill @roman.technology · 01/12/2024Just watched this incredibly well-made video on how integrated circuits are manufactured and was reminded how high the ceiling of educational resource quality is. www.youtube.com/watch?v=dX9C...youtube.comHow are Microchips Made? 🖥️🛠️ CPU Manufacturing Process StepsYouTube video by Branch Education 091
Roman Hauksson-Neill @roman.technology · 29/11/2024Interesting phenomenon. I found this paper investigating it in more detail: journals.plos.org/plosone/arti... 1418
Roman Hauksson-Neill @roman.technology · 27/11/2024Thoughtful thread on the recent controversy about scrapes of Bluesky posts and data ethics in general. 030
Roman Hauksson-Neill @roman.technology · 24/11/2024I heard from the grapevine that Peter Wildeford joined 010
Roman Hauksson-Neill @roman.technology · 23/11/2024I enjoyed this episode a lot, really expanded my theory of mind 010
Roman Hauksson-Neill @roman.technology · 22/11/2024❌ Strategically timed donations ✅ Strategically timed bangers 080
Roman Hauksson-Neill @roman.technology · 22/11/2024Great writeup: meaningness.com/sad-light-le...meaningness.comYOU NEED MORE LUX | MeaningnessSeriously bright indoor lights for seasonal affective disorder, or just the winter blahs 261
Roman Hauksson-Neill @roman.technology · 22/11/2024There are plenty of projects trying to do the same thing (e.g. gotit.pub), but they don’t have the same reach. 010
Roman Hauksson-Neill @roman.technology · 22/11/2024I think Bluesky is a good candidate for a standard platform on which academic papers are discussed. It’s open, has traction, and interfaces with the general public as well as researchers. Imagine if every arXiv paper had a link to a Bluesky thread where people could share their thoughts. 230
Roman Hauksson-Neill @roman.technology · 21/11/2024Maybe silly question: using a custom domain as my handle doesn’t prevent this, right? 100
Roman Hauksson-Neill @roman.technology · 21/11/2024Your new pfp and anti-nerd attitude go hand in hand 110
Roman Hauksson-Neill @roman.technology · 21/11/2024Looks like a quite interesting podcast series. I’ve been meaning to study how researchers conceptualize what intelligence is exactly. 010
Roman Hauksson-Neill @roman.technology · 20/11/2024Awesome news! Is it inspired by the Stanford Encyclopedia of Philosophy? 100
Roman Hauksson-Neill @roman.technology · 19/11/2024Cool idea. Wouldn’t you have to give a negative reward to users who supplied information that made the AI’s forecast less accurate? Otherwise people could exploit the system by submitting opposing pieces of information on different accounts. 100
Roman Hauksson-Neill @roman.technology · 19/11/2024I’ve thought about this before too. Explanations often note that something releases dopamine or activates some region of the brain, without adding any explanatory power (“what would you expect to be different about the behavior if it didn’t activate that region?”). 000
Reposted by Roman Hauksson-NeillGriffiths Computational Cognitive Science Lab @cocoscilab.bsky.social · 18/11/2024(1/5) Very excited to announce the publication of Bayesian Models of Cognition: Reverse Engineering the Mind. More than a decade in the making, it's a big (600+ pages) beautiful book covering both the basics and recent work: mitpress.mit.edu/978026204941... 14521120
Roman Hauksson-Neill @roman.technology · 19/11/2024EA and EA-adjacent starter pack from Agus or someone else 120
