Hacker News: lorenzhs

New comment by lorenzhs in "College instructor turns to typewriters to curb AI-written work"

lorenzhs — Sun, 19 Apr 2026 10:45:20 +0000

If you can't transfer the course's material onto new, unseen questions, then you might have memorised it, but you didn't understand it. Getting an A requires understanding, not rote memorisation.

New comment by lorenzhs in "Report: Microsoft kills official way to activate Windows 11/10 without internet"

lorenzhs — Sat, 03 Jan 2026 21:28:50 +0000

Even if it was still a thing (and it really isn’t, imo), libimobiledevice does a decent job already, and given a little funding it shouldn’t be super hard to close the gaps and build a nice UI on top of it. But that’s not happening because very few people care about it at all.

Now, AirDrop support is a completely different beast. But it requires hardware support (promiscuous mode, iirc) that many common chipsets simply lack.

New comment by lorenzhs in "Kidnapped by Deutsche Bahn"

lorenzhs — Tue, 30 Dec 2025 11:25:39 +0000

Good news for you! 1.5 is in O(50) since they’re both constants! (It’s 1.50 because the author has the 58€-a-month flat rate ticket for all local and regional services in the entire country. If you buy a regular ticket, you get back 25% of the ticket price for delays over an hour and 50% for delays over two hours)

Show HN: Firebolt Core – #1 in ClickBench – Free Scale-Out Analytical SQL Engine

lorenzhs — Wed, 25 Jun 2025 14:00:34 +0000

Hi HN, there’s a distinct lack of modern self-hosted scale-out query engines. A lot of the innovation in the last 10 years has been in SaaS-only systems. That’s also been true for Firebolt until now. We’re now taking the radical step of offering our query engine as a Docker image that’s free for commercial use without any real restrictions on what you can use it for (basically everything except competing with our SaaS offering). There are helm charts and docker compose files in the repo as well to help you get started.

The focus of Firebolt is on low-latency, high-concurrency analytics like you might have them in user-facing workloads (dashboards, data-heavy apps, ...). To show just how fast it is, we submitted Core to ClickBench and instantly took the top spot: https://benchmark.clickhouse.com/. But Firebolt isn’t just a query accelerator, it can also handle scale-out ETL/ELT and all the other things you might want to use a data warehouse for.

I’ve been working on Core for the last few months, and moshap (our CTO), wagjamin (VP Eng) and I are here for your questions. Mosha and Benjamin have written about our motivations to offer Core on the company blog: https://www.firebolt.io/blog/introducing-firebolt-core

Comments URL: https://news.ycombinator.com/item?id=44377439

Points: 6

# Comments: 0

New comment by lorenzhs in "Rusty.hpp: A Borrow Checker and Memory Ownership System for C++20"

lorenzhs — Mon, 06 May 2024 11:07:13 +0000

Indeed: https://clang.llvm.org/extra/clang-tidy/checks/bugprone/unch...

New comment by lorenzhs in "Rusty.hpp: A Borrow Checker and Memory Ownership System for C++20"

lorenzhs — Mon, 06 May 2024 11:06:25 +0000

clang-tidy has a check for this -- it's not a compiler check but with clangd and LSP, almost every code editor can show an inline warning: https://clang.llvm.org/extra/clang-tidy/checks/bugprone/unch...

New comment by lorenzhs in "Rusty.hpp: A Borrow Checker and Memory Ownership System for C++20"

lorenzhs — Sun, 05 May 2024 19:18:52 +0000

Should’ve checked my own link instead of relying on memory — I might have some code to revisit on Monday. That’s insane, thanks for correcting me!

New comment by lorenzhs in "Rusty.hpp: A Borrow Checker and Memory Ownership System for C++20"

lorenzhs — Sun, 05 May 2024 11:20:31 +0000

There is no implicit conversion (except to bool, but that tells you whether the optional contains a value), and operator* / operator-> throw std::bad_optional_access if it’s empty. See https://en.cppreference.com/w/cpp/utility/optional

New comment by lorenzhs in "Germany explores 4-day workweek amid labor shortage"

lorenzhs — Thu, 01 Feb 2024 22:02:02 +0000

Making an argument about how people may want to work less based on a number most have never seen or care about is a flawed argument no matter how you approach it.

New comment by lorenzhs in "Germany explores 4-day workweek amid labor shortage"

lorenzhs — Thu, 01 Feb 2024 21:22:46 +0000

As an employee, the Arbeitgeberbrutto is a number you never see. You’re not saying that employing people isn’t worth it, you’re saying that working in Germany isn’t worth it. Then you have to take the employee’s perspective, not the employer’s. Nobody advertises the true cost to the employer in other countries either. How much do all those benefits to US employees cost? Do you include that when you list US salaries, or do those get magically excluded because health insurance isn’t mandatory?

Your premise just seems fundamentally flawed.

New comment by lorenzhs in "Germany explores 4-day workweek amid labor shortage"

lorenzhs — Thu, 01 Feb 2024 20:52:15 +0000

This is not the norm, as the employment modalities in that thread are super strange. If you’re employed by a German entity and your salary is 91.1k€, your take-home income would be around 53.8k€. Put another way - the $100k / 91.1k€ are the Arbeitgeberbrutto, they correspond to 75.5k€ Arbeitnehmerbrutto (the number you’d normally see on your contract). The difference exists because there are employer deductions and employee deductions, but the employer deductions don’t show up on the pay slip. Only in this constellation the employee has to pay both. It’s very much the exception, not the norm.

New comment by lorenzhs in "The Hyperloop was always a scam"

lorenzhs — Fri, 22 Dec 2023 16:16:57 +0000

Something like this? https://en.wikipedia.org/wiki/London_Post_Office_Railway https://de.wikipedia.org/wiki/Post-U-Bahn_M%C3%BCnchen https://de.wikipedia.org/wiki/Sihlpost#Post-U-Bahn

New comment by lorenzhs in "California PG&E joins proposal to set utility rates based on income"

lorenzhs — Mon, 14 Aug 2023 18:40:27 +0000

Your information is quite a few years out of date. The solar learning curve has continued in the meantime, reducing the cost every year. Especially in sunny places like California, unsubsidised solar pays for itself surprisingly quickly (I don’t have the numbers for CA, but I wouldn’t be surprised if the number was 10 years).

New comment by lorenzhs in "Ask HN: What are some cool but obscure data structures you know about?"

lorenzhs — Fri, 22 Jul 2022 22:44:27 +0000

If you enjoyed XOR filters, you might also like ribbon filters, something that I had the pleasure of working on last year. They share the basic idea of using a system of linear equations, but instead of considering 3 random positions per key, the positions to probe are narrowly concentrated along a ribbon with a typical width of 64. This makes them far more cache-efficient to construct and query.

By purposefully overloading the data structure by a few per cent and bumping those items that cannot be inserted as a result of this overloading to the next layer (making this a recursive data structure), we can achieve almost arbitrarily small space overheads: <1% is no problem for the fast configurations, and <0.1% overhead with around 50% extra runtime cost. This compares to around 10% for XOR filters and ≥ 44% for Bloom filters.

In fact, I'm going to present them at a conference on Monday - the paper is already out: https://drops.dagstuhl.de/opus/volltexte/2022/16538/pdf/LIPI... and the implementation is at https://github.com/lorenzhs/BuRR/. I hope this isn't too much self-promotion for HN, but I'm super hyped about this :)

New comment by lorenzhs in "Ask HN: What are some cool but obscure data structures you know about?"

lorenzhs — Fri, 22 Jul 2022 22:26:14 +0000

Indeed! Funny enough, I rewrote that article a few years ago, it previously contained an approximation of something like Algorithm L that some person with a blog came up with, having no idea that an even simpler and provably correct algorithm was published as early as 1994 :) Though others have improved the article a lot since then, adding explanations for how/why it works. Couldn't be happier to see it cited in this list!

New comment by lorenzhs in "Turn Photos into 3D Models"

lorenzhs — Sun, 10 Apr 2022 12:10:42 +0000

The App Store page says "Simply take photos of an object from all angles and upload them to our service, and we will create and send you a 3D model ready for 3D printing, Augmented Reality and our web app r3Dent." so not it appears to use some kind of server-side processing. So it's not even on-device.

New comment by lorenzhs in "The definitive guide that you never wanted: Backpack fabrics (2017)"

lorenzhs — Sat, 26 Feb 2022 08:02:07 +0000

Often it’s just the thing that goes up and down that needs to be replaced, that’s a 5$ fix

New comment by lorenzhs in "Google says iMessage is too powerful"

lorenzhs — Wed, 12 Jan 2022 11:59:21 +0000

https://signal.org/android/apk/ and it appears to work without Google Play Services (though push notifications drain more battery iirc because they can't use the google push notification service)

New comment by lorenzhs in "Nuclear fallout is showing up in U.S. honey, decades after bomb tests"

lorenzhs — Sat, 01 Jan 2022 18:25:47 +0000

Cosmic background radiation is a thing and water is really good at blocking it, so it's not complete bullshit. Of course, without knowing what dose the GP received, it's impossible to say whether it's a comparable dose to what they would have received on the surface - but it's plausible.

New comment by lorenzhs in "We will be retiring Alexa.com"

lorenzhs — Thu, 09 Dec 2021 08:52:13 +0000

Other companies do that, too - I think it's due to trademark law and how it's much harder to establish a new name than reuse an existing one. Cortana comes to mind.