Andrej Bauer

@dearmadisonblue Found it! #comment-73703" target="_blank" rel="nofollow noopener">lambda-the-ultimate.org/node/4655#comm…

@andrejbauer I thought it was a great post as soon as you wrote it back in 2012 and saved it right away; Lambda the Ultimate went down a couple years ago but I still had a copy. I don't think it's available other than this Twitter thread anymore

They will dictate trends and fashions in programming, not those few specialists who know how to use keywords and who actually understand what is under the hood. END 🧵

@dearmadisonblue By the way, where is this available? It looks like lambda-the-ultimate isn't working anymore.

I didn't write this!! This is from Andrej Bauer on Lambda the Ultimate back in like 2012, predicting what programming would be like in 2020. seemed pretty accurate to me, although data science kinda flopped and isn't a thing anymore afaik

@dearmadisonblue I have no recollection of writing this. If I really did it, I used to be a lot wiser than I am today. (Or perhaps you sifted through all of Lambda the ultimate until you found a prediction that happened to become true.)

hax: Verifying Security-Critical Rust Software using Multiple Provers eprint.iacr.org/2025/142 Hax is a verification toolchain for Rust targeted at security-critical software with backends in ProVerif and SSProve, as well as Coq and F*.

@tailcalled @JDHamkins What rules are those (in Lean specifically)?

People are formalizing @JDHamkins's infinite-time Turing machines and wondering whether ordinals up to ω₁ are the same in all type-theoretic universes. (The universe Type_n can prove the existence of n-1 inaccessible cardinals, if that matters.) proofassistants.stackexchange.com/q/4626/169

@JasonRute @JDHamkins Yes, see this discussion on Zulip. Thanks @JasonRute. Now I wonder what extra assumption about type-theoretic universes in Lean allows us this. #narrow/channel/217875-Is-there-code-for-X.3F/topic/Countable.20.28or.20omega1.29.20is.20the.20same.20when.20lifted" target="_blank" rel="nofollow noopener">leanprover.zulipchat.com/#narrow/channe…

@JDHamkins @andrejbauer "If each model of ZFC gives rise to a type-theoretic universe". IIUC, they don't. Lean universes are just Grothediek universes, which in set theory is V_kappa for strongly inaccessible kappa. One can actually prove in Lean that omega_1 is the same in all universes Type u.

@JDHamkins Ah, we could have Type_1 belive in one kind of countability and Type_2 (which contains Type_1) in another, as there is no requirement what these are, other than models of Zermelo set theory, more or less.

Wouldn't there be type-theoretic universes that correspond to various set-theoretic inner models, such as the constructible universe L, which (consistently) has a different concept of countable ordinal? In set theory, being a countable ordinal is not absolute between transitive models of set theory, even when they have the ordinals overall.

OCaml 5.3 is now released discuss.ocaml.org/t/ocaml-5-3-0-…. Also in the box -- MSVC support, statistical memory profiler (remastered for multicore). With OCaml 5.3, we're at feature parity with OCaml 4.x series, except for deprecated features (native 32-bit backends).

Who in industry is offering PhD-level internships related to Lean, Proof Automation, or Constraint Programming?

@JDHamkins I honestly fail to see what this tension between set theory and category theory/topos theory is supposed to be. There are people (some of which quite influential) who say silly things, that much is clear. Perhaps that's the problem. P.S. Mastodon lets you edit posts for free.

James Hanson, arguing that the dispute between category theory and set theory is essentially over and has been for some time. He calls out the rhetoric of Lawvere as frustrating and unhelpful. mathoverflow.net/a/484638/1946

My hand made Möbius strips are on the cover of the award winning book The Math Book by Dr. Cliff @Pickover. Created especially for it, 8 artworks of mine are in it, too. Upcoming updated version will have a different. cover. See sprott.physics.wisc.edu/pickover/math-…🥰

S kamerami so posneli lučke in izračunali njihove koordinate v prostoru. Napisali so 3 simulatorje in spletno stran, ki prikazuje trenutni vzorec. Učitelje vabimo, da se nam z učenci in dijaki pridružijo. Pošljite svoje vzorce! Podrobnosti so na spletni strani. 2/2

V predverju @FMF_UL stoji novoletna jelka, ki so jo postavili naši študenti. Njenih 500 barvnih lučk lahko programira vsak na daljavo in ustvari svoj vzorec. Študenti so ustvarili tudi vizualni programski jezik Jelkly za ne-programerje. Več na jelka.fmf.uni-lj.si 1/2

Ministrstvo za visoko šolstvo, znanost in inovacije (MVZI) je včeraj organiziralo posvet na temo urejanja področja komuniciranja znanosti. Tole je bil moj uvodni dvajsetminutni prispevek.

Prof. dr. Maruša Bradač, FMF UL je članica raziskovalne skupine, ki je razvila enega od 4 najpomembnejših znanstvenih instrumentov na vesoljskem teleskopu James Webb, kamero NIRISS, ki preučuje tudi eksoplanete. Ugriznimo znanost: Iskanje eksoplanetov jutri ob 17.25 na TV SLO 1.

Na Fakulteti za matematiko in fiziko Univerze v Ljubljani je dopoldne potekala okrogla miza z naslovom »Slovenija v CERN-u – v centru prebojne znanosti in velikih odkritij«. Presenetljivo veliko zanimivih informacij je zbranim predstavil Jure Gašparič, državni sekretar na Ministrstvu za visoko šolstvo, znanost in inovacije, ki je po izobrazbi sicer zgodovinar. Gašparič se že dlje časa udeležuje sej upravnega odbora CERN-a (CERN Council) kot predstavnik države in je zato zelo dobro seznanjen z razpravami na najvišji ravni delovanja te mednarodne znanstvene inštitucije, obenem pa zna svoje znanje in izkušnje posredovati na razumljiv in zanimiv način. Med znanstveniki je še posebej izstopal Jernej Fesel Kamenik z Instituta Jožef Stefan (IJS), ki mu je prav tako uspelo kompleksne znanstvene teme ubesediti tako, da so jih vsaj malo razumeli tudi tisti, ki z njimi niso bili že predhodno seznanjeni. Posnetek okrogle mize si lahko ogledate na povezavi: youtube.com/watch?v=WoNPX3…

We added a graphic library explorer to the agda-unimath library just now: unimath.github.io/agda-unimath/V… Thanks to Job Petrovcic and @andrejbauer for creating it!