HN user

matesz

1,160 karma
Posts90
Comments203
View on HN
hypr.land 2d ago

Hyprland 0.55 announced the switch to Lua for its config files

matesz
140pts141
jacek.migdal.pl 11d ago

Make Europe Cool Again

matesz
5pts0
github.com 2mo ago

Talos OS images are now bit-by-bit reproducible

matesz
1pts0
luxmentis.org 4mo ago

Build, run and debug iOS and Mac apps in Zed instead of Xcode

matesz
2pts0
www.eu-inc.org 6mo ago

EU-Inc: An API-first corporate structure as alternative to Delaware C-Corps

matesz
4pts2
www.uber.com 6mo ago

Uber Conquered Database Overload

matesz
2pts0
bchess.github.io 7mo ago

K8s-1M: Unintentionally reinventing Google Borg to scale Kubernetes

matesz
2pts0
iwriteiam.nl 7mo ago

Reviewing live-bootstrap and reflections on trusting trust

matesz
1pts0
guix.gnu.org 7mo ago

The Full-Source Bootstrap: Building from source all the way down

matesz
2pts3
link.springer.com 7mo ago

Busy doing nothing: why politicians implement inefficient policies

matesz
2pts0
turso.tech 8mo ago

Turso in the Browser

matesz
2pts0
arxiv.org 8mo ago

Reliable Confidence Intervals for Information Retrieval Evaluation

matesz
1pts0
www.ingonyama.com 9mo ago

Solving Reproducibility Challenges in Deep Learning and LLMs: Our Journey

matesz
1pts0
poolside.ai 10mo ago

Building Towards AGI

matesz
2pts0
www.youtube.com 10mo ago

How EUV machines work [video]

matesz
3pts0
news.ycombinator.com 10mo ago

Ask HN: Why does Google word privacy settings like you agree even when off?

matesz
29pts21
debu.gs 10mo ago

Inferno Part 2: Let's Make a Cluster

matesz
1pts0
swtch.com 10mo ago

9vx and how to run Plan 9 on FreeBSD, Linux, and OS X

matesz
2pts0
www.anthropic.com 10mo ago

Anthropic Changes Training Data Policy from Opt-In to Opt-Out

matesz
8pts2
justine.lol 11mo ago

Cosmopolitan Libc – your build-anywhere run-anywhere C library

matesz
1pts0
blog.cleancoder.com 11mo ago

Why Clojure?

matesz
21pts0
temporal.io 11mo ago

The Curse of the A-Word

matesz
3pts0
turso.tech 11mo ago

How Turso made connections to SQLite databases 575x faster

matesz
2pts1
prg.is.titech.ac.jp 1y ago

BatakJava: An Object-Oriented Programming Language with Versions [pdf]

matesz
1pts0
blog.thelifeofkenneth.com 1y ago

Creating an Autonomous System for Fun and Profit

matesz
1pts0
temporal.io 1y ago

Temporal raises $146M Series C

matesz
3pts0
antithesis.com 1y ago

Our own worst best customer

matesz
12pts0
bdow.com 1y ago

80-20 Business Idea Validation (Fastest Way to Start a Business)

matesz
1pts0
adobe.design 1y ago

Spectrum 2: Our vision for the future of Adobe experience design

matesz
1pts1
spectrum.adobe.com 1y ago

The Basics of Color Fundamentals in Adobe Spectrum

matesz
1pts0
Kimi Work 2 days ago

That is an awful take. Although Silicon Valley's tech bros are shameless in their own way, there's no justification for symmetrizing the two.

I have to ask…why do people want a Turing complete language to configure stuff?

I don't understand this either. IMO, the current SOTA tool for configuration is CUE, because it is the only one that lets you write schemas for arbitrary refinements. I haven't used it myself, but my understanding is that everybody should be writing configurations in something like this instead of YAML, TOML, HCL, or a custom DSL with no tooling. And yes, definitely not in a Turing complete programing language.

Apple WWDC 2026 1 month ago

I bet they planned this before the initial release and actually had this capability then and there. Just needed some guinea pigs (aka their users) to learn more and establish the trend.

These days, when I think of empire, countries like the US, China or Russia fit the bill much more than the EU, which is struggling to reach that level of integration and influence.

Of course! Especially because there is no unified army control.

But this requires giving more context. We can't forget that there are ways, especially ways made by empires, to force other nations to go to war not only as an ally but also to make them less relevant and take a hit also.

One of the main factors which makes this more probable, is what op mentioned, the raise of fascism and combatant militaristic attitudes exacerbated by the fact that their own nation / empire is a falling empire. And EU didn't fell yet, it is huge economy with more people than the US.

It isn't prophecy. The Harvard Belfer Center study (Thucydides's Trap) analyzed 16 cases in the last 500 years where a rising power challenged a ruling empire: 12 of them ended in war.

The UK is actually the perfect example of this danger. The British Empire didn't dissolve peacefully - it was effectively destroyed by WWI and WWII while trying to suppress a rising Germany.

The subsequent transfer of hegemony to the US was a rare statistical anomaly (a "special case" driven by shared culture and total British exhaustion), but the Empire’s fall itself was catastrophic.

The pattern is violence, not peace. And remember that other aspiring nations to maintain it's position as Empire actively acting to destabilize situation in other states. The reason is simple - it is easiest way to maintain their status.

Brexit for instance was a boon for everybody but UK and EU. There is clear data already about Russian intervention. Recent overt US intervention into ensuring UK remaining separate and EU becoming separate. Think about it.

Its important to understand the context. Germany, and Europe in general, is basically like a falling empire. It will be less and less significant and life won't get any better than it is relatively speaking. The same will almost certainly happen with the US, but first goes Europe.

Unfortunately citizens and therefore ruling elites of empires fueled by relatively extremely high standard of living for decades in comparison to the rest of the World always have very hard time swallowing their national pride. They have built very elaborate conceptual framework of linking their nationality to the level of relative success, fueled by politicians who want to make people feel good again about their nationality.

Just look at the news, almost everything directly or indirectly is linked to the concept "nation".

And in almost all cases of empires a natural consequence of their fall is war. So, it is very important to set expectations right.

This idea was made by French and German government

It is true, but lets face it - it is just a high-level proposal. Better than previous ones for sure, but still just a proposal.

And there is a massive gap between the PR and the legal reality. Germany and Austria already torpedoed the previous attempt (the SUP) specifically because it allowed online formation without a notary.

So I would advise against certainty in that the fight is over. I also don't want to wait another 10 years. Honestly by that time just setup topco in UK and if series A will come just do the flip later if you don't want to go to SV straight away.

I don't understand the tax objection. The 28th regime changes corporate law, not fiscal law. A Berlin-based "EU-Inc" still pays full taxes (Körperschaftsteuer, Gewerbesteuer, etc.) to the local Finanzamt. It isn't a disembodied tax haven - it just standardizes the legal wrapper.

You are right that this threatens the status quo, though. If this works, founders based in Germany will likely abandon the GmbH. That will require swallowing some national pride and admitting that the current system is simply a less efficient, less competitive legal form for high-growth companies.

One thing I think is also worth mentioning are labor rights. I am not arguing against the German model of employee protection. Mitbestimmung could be viewed as a good thing, even if it will mean less power to the VC and / or founders. And frankly, I don't care if the consensus forces strict, German-style Mitbestimmung on the EU-Inc. Stricter form of EU-INC is still vastly better than nothing at all.

Asianometry has a great video on the labor rights in germany btw - https://www.youtube.com/watch?v=K0teMtLT9XI).

My Dear German friends, please accept that EU-inc is a massive boost for EU competitiveness. Continuing to block progress here only proves that the EU is good at making promises on paper but fails to deliver real change.

The reason this "28th regime" actually works—and bypasses the previous vetos — is that it’s optional. It doesn’t try to force Germany to change its local GmbH laws or kill the notary system overnight. Instead, it creates a parallel, voluntary path. Berlin doesn't have to give up its red tape for local shops, but they have to let a "Unified European Company" exist alongside them.

If we let national pride or local bureaucracy stop this again, we are essentially telling our best founders to leave for the US. Let’s stop protecting red tape and start protecting our future.

In general, this is good effort and its a shame they only come up with this now. This should have been done some 20 years ago.

I am wondering whether it what are the chances this is going to move forward and when. These are big questions.

Also, is it possible to have all economies unified in the future? No tariffs is one thing, but if we have free trade one not just unified entire legal framework?

Unfortunately it doesn't - even if our package definition doesn’t change output may, since software referenced by the package’s build function is not bit by bit reproducible.

The biggest issue is with Linux Kernel itself. Once kernel reproducibility will be handled, we can practically claim Guix, NixOS, Debian or any other Linux distro is reproducible as well, since with little effort we can avoid installing packages which will colour our reproducibility.

For desktop environments it’s a bit worse, but there are huge efforts to make this happen for all packages in package registries of aforementioned distros.

As a side note NetBSD is THE ONLY operating system of which binaries are fully bit by bit reproducible.

Moreover it vendores patched gcc compiler, so the entire toolchain is reproducible too as well.

How cool is that?! For the record golang brings it further - its entire package registry containing +40mln packages is bit by bit reproducible.

When I discovered Clojure, apart from the functional language properties and Java integration it brings with it, I was completely struck by how elegant its codebase is.

From what I remember there is around 60k lines of Clojure itself and pretty much all files were edited like minimum 8 years ago, apart of main file with most of the function utilities.

Mind you I am hosting this just for about a week now - +100GB in total for all inboxes. Also I removed automatic daily purging so all spam and deleted items stay just to be safe.

Haven't looked into spam more closely yet. After first glance on most publicly shared email address - there is around 2 spam messages per hour.

Here is report prepared by llm which looked through the last 20 email headers found in spam. All of them were categorized correctly, however there were few emails in the past few days which went to spam where they shouldn't but I think this is fixable.

- Critical Authentication Failures: A large number of the messages failed basic email authentication. We see many instances of SPF_FAIL and VIOLATED_DIRECT_SPF, meaning the sending IP address was not authorized to send emails for that domain. This is a major red flag for spoofing.

- Poor Sender IP Reputation: Many senders were listed on well-known Real-time Blackhole Lists (RBLs). Rules like RBL_SPAMCOP, RBL_MAILSPIKE_VERYBAD, and RBL_VIRUSFREE_BOTNET indicate the sending IPs are known sources of spam or are part of botnets.

- Suspicious Content and Links: The spam filter identified content patterns statistically similar to known spam (BAYES_SPAM) and found links to malicious websites (ABUSE_SURBL, PHISHING).

- Fundamental Technical Misconfigurations: Many sending servers had no Reverse DNS (RDNS_NONE), a common trait of compromised machines used for spam.

There have been few messages which went to spam which didn't meet any of this spam criteria but actually they were cold marketing emails, so it's good too. In addition to this stalwart emits info log for each possible spam message ingested. Not sure if this can get any better than this.

Downsides: some features are gated behind an enterprise version

I treat this as an insurance policy. Even in this thread people mentioned how Maddy, which is an alternative modern full stack email solution in a single binary, lacks development efforts.

This is why we have this fantastic release for Stalwart - free shit.

Also as of now enterprise is for $0.2 per account per month which is extremely cheap unless somebody wants to build a big spam farm, of which as civilized Internet user I don't support. Obviously this might change, but even if you can always built multi-tenancy layer by yourself if you really need it - rest of the codebase is AGPL.

I just run single instance for now with RocksDB backend for internal / search and S3 for blobs - that is what made me think it’s so flexible.

Never hosted Postfix / Dovecot stack, in fact this is the first time I host emails, but from what I understand Stalwart is designed to handle inbound directly.

For very high throughput inbound you could check out KumaMTA - it was designed specifically for that, but I think Stalwart doesn’t have bottlenecks in it’s clustered topologies which would require it unless you are doing something crazy.

They have very good docs in general IMO, here are docs on how to cluster - https://stalw.art/docs/cluster/configuration

Reminds me how Alibaba innovated in database space with a cluster of FPGAs doing LSM compaction with custom MySQL storage engine - https://dl.acm.org/doi/pdf/10.1145/3299869.3314041

Basically they achieved massive improvements in throughput and latencies vs RocksDB, for some operations even an order of magnitude.

They even offered this as a service, but eventually pulled out. Not sure if they are still using it, but I was surprised that pretty much nobody else, as far as I am aware, is using any hardware acceleration for repetitive database workloads.

FB is a ghost town

I've done surveys in cities about what social media people use and came to the same conclusion. However, I was completely wrong.

Facebook is so alive and well it's hard to believe. Besides that they skillfully connected two ecosystems together and there is much more people having FB than IG. Stories show up in messenger and quietly lead back to facebook just as links to fb videos people send to each other frequently.

It's just that people simply lie in their actual usage patterns because it's really uncool.

Primary people's identity online is still their Facebook profile.

Unfortunately it's an incredible tool, especially for industries which pray on people's insecurities like beauty - botox, fillers etc. This person I know puts instagram story and gets instantly booked for all free slots she has for the entire week.

She talked about some people from her industry doing billboard ads and laughed how inefficient they must be, knowing that people are so hooked on "insta".

I believe it takes maturity and wisdom to unhook from social media - facebook, youtube, linkedin, instagram etc. Especially reactive use, not the one which comes from internal pause / response.

I tried to unhook pretty much for the past 15 years as I sensed that it basically doesn't serve me. If I would summarize the one primary cause for my inability to do it is the following - the belief that consuming content online is better for my own being than learning to manage my monkey mind.

I mean any content - from scrolling dumb instagram and facebook feeds to factory making process videos on youtube and streamers playing online games, political debates etc.

The problem is not consuming content on social media, but doing it reactively, excessively.

What helps with unhooking is basically wisdom and experience because how to do it when pretty much everybody is doing it?

Realizing that entire social media world is just incredibly fucking corrupt. Like omg corrupt. It's the epitome of corruption, starting with CEOs themselves.

Last week I've had situation where the person I knew who has professional instagram profile with +10k and runs business there just went fucking nuts. Instead of focusing on working on herself she decided to double down on her narcisism and went mental. Episode, however this is where it leads.

I am just happy that I can see it better and better and step into the right direction - away from social media.

PS. I removed X account few months ago, oh my, what a relief!

I’m still using iPhone 12 mini running iOS 26 beta and it’s good enough. OS is definitely not polished, some of the design choices don’t even make sense but in general I believe it’s the right direction - spatial + maxing out visual looks.

Being able to turn Liquid Glass off to sth like flat design would be nice but this probably won’t happen.

Now when it comes to the event itself, it felt so cartoonish.