Enrico Borba

@kerckhove_ts huh, i always thought "bare" paths had to begin with ./ or /

WAT

@sampocino The QPF paper is pretty good: contrib.andrew.cmu.edu/~avigad/Papers…

banging my head against the wall for this

@mitsuhiko @jarredsumner @Nekrolm This sounds like more the Rust crowd being concerned that Bun (in Rust) will be used as a token example of "see how Rust doesn't solve memory issues".

@jarredsumner @Nekrolm I will say that I find the criticism about the unsafe usage by the zig crowd quite amusing given how dismissive the entire community has been towards memory safety. (And to be clear I do think the rust community is over-doing it simultaneously)

Не знаю, какую там надежность слопотрансляция на Rust привнесет в Bun, но вот с такой отвратительной фигней, ни от каких use-after-free Rust не защитит. Option> : Copy — ВО ВСЕХ этих safe методах можно получить use-after-free, подав дважды тот же указатель и его копию. Пока AI не перепишет подобное на нормальный безопасный Rust, без спрятанного unsafe с сомнительными доказательствами, я б не рекомендовал - Реализовывать новые фичи - Вообще этим пользоваться

FR8 is a 12,000 m² palace, filled with geniuses researching or building startups, it charges 0% equity and even pays for your food, living, and flights. One of their founders drank actual poison on stage to demo their tech. Welcome to FR8. Nothing about FR8 makes sense because it’s so over the top in their ambition, but they might eventually become the biggest thing for young founders globally. And it’s happening right here in Europe. They are neither a hackerhouse, nor a startup accelerator, nor a classic research lab. Instead they think of themselves as a university-like institution for the post AGI world that pushes you towards building companies, ambition, obsession, and bias-to-action. Think YCombinator, Stanford and Bell Labs all wrapped into one thing for the most ambitious 20-somethings in the world to work, run by 20-somethings. They just came out of stealth. Until recently people didn’t even know where their latest cohort is based. Because additionally on top FR8 is absurdly secretive. Their target group knows them and that’s about all they care for. We visited last week to join them behind-the-scenes as they prepare for their first demo day in their new building - a 5 floor university building in the middle of Helsinki. We knew them for quite some time so we were allowed to film them as the first team worldwide. But even we couldn’t film multiple floors and rooms of their building. This video gives you an insight into the ambitious craziness that FR8 is – but trust me there’s more to come in the near future. The biggest new thing in startups – isn’t in SF – it’s in the north of Europe and attracts young geniuses worldwide. Welcome to FR8!

@TheEduardoRFS @KleeneAlgebra Just because it won't happen if you're writing sane code, doesn't mean it "doesn't matter" or "isn't real".

@TheEduardoRFS @KleeneAlgebra Yeah but you can use this kind of thing adversarially. E.g. people could introduce supply chain attacks into Rust libs that still compile but cause memory errors that Rust allegedly always protects against, even with #![forbid(unsafe_code)], as it has an unsound type system.

I still don't believe in the incompleteness theorems or the halting problem. I understand them considerably better, but I think they're not real.

@TheEduardoRFS @KleeneAlgebra > It's okay if type checkers don't halt as long as it doesn't happen. This OCaml program loops the type checker: caml.inria.fr/pub/old_caml_s… > Also even if your type checker halts, ... Are you saying that guaranteed termination does not guarantee utility? If so, agreed.

@enricozb @KleeneAlgebra It's okay if type checkers don't halt as long as it doesn't happen. Also even if your type checker halts, it's still likely that in the worst case it takes more than the age of the universe to check.

@TheEduardoRFS @KleeneAlgebra Type checkers being undecidable means that they might not halt, so in many typed languages I can make the compiler loop forever. Some languages have a "step count" where they only perform N type checking steps before giving up as a result of this. This feels real to me, no?

@KleeneAlgebra There is no concrete problem that you will ever hit them, they only happen because we assume some infinities. Infinites are useful, but they're "processes that never terminate" and it makes sense that if you assume some "processes that never terminate", you get to weird spots.

@HSVSphere You on the pro?

I actually can't use my macbook anymore. The framework 13 cucklinux dot com laptop is genuinely that much better of an experience for me. I'm not even kidding. I just tried because I had some software on it that I needed. I'd just rather not. Wow. My macbook is now a server

@profleonn If you're real you'll bind ctrl-z in your shell to fg. Then it's just a toggle.

if you are not constantly using ctrl-z and fg, what are you doing?

@MakerInParadise @nilscmr Yes, and no cache invalidation/coherency despite the independent workloads cores have, bc of the computational model.

@enricozb @nilscmr Shared memory and if so, bandwidth? If not, advantage over traditional clustering?

CPUs suck. We're building a new general-purpose chip that scales to thousands of cores while being more energy-efficient. We're hiring hardware design engineers, consider joining us tendrils.co/jobs What we do differently ...

@MakerInParadise @nilscmr GPUs are SIMD, where many cores perform the same instructions in lock-step. What we're building has all cores concurrently performing completely independent instructions.

@nilscmr You basically just described a GPU network. What are you doing different than what already exists?

@profleonn another day another CH-correspondence

its been roughly 1 week since I realize that proof by induction is simply just recursion. life has been pretty good since then.

@profleonn Wireguard for personal VPN, configured with nix so it's easy to set up new machines, laptop / mobile device is mostly stateless; it just pushes git repos, personal data lives on the server.

how do you guys daily more than 1 computer at a time. I basically have my entire life on 1 ssd at a time and have never been happy enough with all the syncing solutions etc., idk

@profleonn What's the benefit over alacritty?

does anybody actually use a terminal other than ghostty ?

@AcerFur Is "Lean Formalized Problem" just stating the problem in Lean?

it is honestly quite remarkable to see the progress on the Erdos problems in less than a year

@schweingehtbree @AcerFur Fair. Vague thoughts: In this blog post dnhkng.github.io/posts/rys/, we see large benefits from duplicating a block of weights of a trained model. These harnesses put LLM's in loops. I'm wondering if there's some reasoning kernel which consistently increases truth buried in there.

@enricozb @AcerFur (Maybe a pedantic point, but: that it lies *somewhere* isn’t actually clear, because if AI actually is intelligent then its intelligence is likely what’s called an emergent property, and emergent properties don’t really “come from” anywhere)

LLMs solving previously open Erdős problems btw

@spicey_lemonade It was really interesting to see @GMHikaru play this. I am probably elo 900 in chess, but there were times where I saw him not grasp possible moves which I was able to see, I'm assuming bc of my CS/Math background. Weird how brains work.

A game between GPT-5.4 xhigh and Claude 4.6 Opus in 5D chess on the eval I’m creating (GPT-5.4 won after ~5 hours). 5D chess is a great eval because it tests long-horizon planning and generalization. There isn’t much data on this game, so any strategy a model develops is likely novel. It’s also exponentially harder than standard chess. The eval scales to ASI. This is a game where you can move pieces back in time and checkmate a player from 5 parallel universes in its futures past. Even @GMHikaru struggled.