Short posts from Andrew Jaffe
Thomas Bayes proved we can never be 100% certain of anything — but that probability tells us exactly how to learn anyway. That’s the core of my SXSW 2027 proposal, from the physics of the cosmos to the beliefs we live by.
Vote here (free, 30 sec) — closes Aug 23.
Exactly two years ago, we were on the first day of our sabbatical year, starting in Indonesia but with extended stays in Tsukuba, Japan; New York City; and Leiden, Netherlands. We’ve been back almost a year but I still think about our year abroad every day.
Clafoutis
Peter Coles points out, I fear with some justification, that the recent STFC cuts to high-profile fields of physics like astronomy, cosmology, and particle physics might have as much (or more) to do with internal conflicts and jealousy within physics as with financial issues.
This is only of niche interest, but it’s great to be at the Sterrewacht group at the University of Leiden, drinking from an old-school early-90s vintage CITA (Canadian Institute for Theoretical Astrophysics) mug.
This White House paper marks a huge change to US science funding, trying to pick tech winners, alongside its efforts to align science with its right-wing ideology.
For a party supposedly against communism, they are doing their best to reproduce some of the worst excesses of the soviet state.
UK Science moves from DSIT to DBIST. Should we be worried? (I have no idea.)
Good (depressing) discussion of the effect of the large cuts to the UK’s astrophysics and particle physics programs. (The government’s own “explainer” is not very good…).
Presented at a meeting today: “Remaining challenges are systemic rather than structural.”
What does this mean?
Goodyear Blimp over the Royal Albert Hall, London.
Fun paper “The Squealer: Sensification of model exploration and model misfit” with new co-authors Andrew Gelman, Eliot Carlson and Philip Greengard, all computational statisticians in the New York area.
Heading off on a wild suit chase, somehow involving the exchange of bank details and morning dress in an English pub.
Hey Peter, I’m not planning on retiring any time soon!
This meeting has been going on here at Imperial College for a few days. I am humbled and honoured by my friends and colleagues who have come to London.
An intersection of my obsessions: @andrewhickey.500songs.com’s “500 Songs” episode on “Hey Jude” opens with a discussion of how we (our brains/minds) build and refine models of the world that could have come straight from my book The Random Universe.
Five-star review by Brian Clegg of The Random Universe:
Probably the best book I’ve ever read on what the scientific method really is, how scientists build models, how they deal with randomness, uncertainty and the pitfalls of induction. Engaging stuff.
There are still tickets available to hear me answer the question “Is Everything We Know About The Big Bang Wrong?” and talk about The Random Universe tomorrow, Tuesday, 16 June at The Blues Kitchen in Brixton (London). The late show is sold out, but you can still get in for some early Pints of Knowledge at 5:45.
Newest PhD student giving his first science talk to our group… (They are all much better at this stuff than we were at the same stage in our careers!)
Shocked but not surprised.
Trump Keeps Immunity from I.R.S., a Victory in a Long-Running Feud
Of course he doesn’t really care about the $1.8bn slush fund — that’s about helping other people (albeit maga criminals). Let’s see about the IRS deal.
The heat has broken just in time for us to have forgotten two of our sleeping bags on our camping trip.
Articles like these arguing that LLMs can’t be conscious simply because they are “mere formulas” are just a version of Searle’s (debatable!) “Chinese Room” argument. (I don’t think they’re conscious either, but that’s not why.)
The Mekons on how they made “Where Were You?” | The Guardian
Well, my Pints of Knowledge talk next month in Brixton on 16 June, is sold out! — so they’ve added an earlier (5:45pm) show. Please come if you want to hear — or ask — about The Random Universe, how we know what we know, and the state of Big Bang as a model for the cosmos.
@andrewhickey.500songs.com, only 3/4 of the way through the Dark Star episode, you’ve managed what none of my Deadhead friends, nor years of reading Xgau, have done: made me wish I had seen the Grateful Dead (possibly with appropriate psychedelic enhancement) when I had the chance 40 years ago.
@siracusa@mastodon.social @atpfm@mastodon.social For me, terminal session restore works — it keeps scrollback & restores directories. Sometimes the tab/window title stops tracking. And there’s a (longstanding, unfixed, race condition) bug in the interaction between the shell and the terminal.
Happy to find myself on a bookshelf with Damian Hirst, Nick Cave and Manolo Blahnik.
Claude is helping me run some code, preparing inputs, monitoring outputs, &c. It kept making badly wrong predictions about how long the run would take, and told me that it was “doing it in my head”, by which it meant it “was computing the estimate inline as part of generating text”.