Hendrikto

9 ms·

151.
by Hendrikto 6mo ago
> When I am working with a small team, I do not care if my commits are ugly or repetitive Your team cares though. Probably including yourself later. Maintaining proper commit history is always worth it.
152.
by Hendrikto 6mo ago
> rebuilding Coinbase as an intelligence, with humans around the edge aligning it Wow. That’s my cue to never use Coinbase again.
153.
by Hendrikto 6mo ago
The most promiment example is probably Go with its goroutines, but there are so many more. You can easily spawn tens of thousands of goroutines, with low overhead and great performance.
154.
by Hendrikto 6mo ago
Very likely not, if they are of similarly low quality.
155.
by Hendrikto 6mo ago
The Zig maintainers did a pretty in-depth review of the PR, and laid out multiple technical reasons for why it would not get merged. They did not reject it simply for being vibe-coded (though that is likely the cause of it sucking).
156.
by Hendrikto 6mo ago
> threads are inherently and inescapably too heavy weight to express concurrency in an efficient way Your premise is wrong. There are many counterexamples to this.
157.
by Hendrikto 6mo ago
If people just followed the XDG Base Directory Specification, config file littering would be a non-issue. More and more projects adopt it, even holdouts like Firefox.
158.
by Hendrikto 6mo ago
Calling people idiots for having different taste? You truly are a muppetman.
159.
by Hendrikto 6mo ago
That 6 months is considered “quick and easy” says a lot.
160.
by Hendrikto 6mo ago
Ironically, I think the examples given in the post validate Dijkstra’s points, instead of disproving them, as the author intended.
161.
by Hendrikto 6mo ago
> HN is still worth visiting, but it's getting harder to justify spending time here I feel the same. Quality of both submissions and discussions have considerable decreased. It is still the best general purpose “aggregator” I know o
162.
by Hendrikto 6mo ago
I would like to read more about this. Do you have a source?
163.
by Hendrikto 6mo ago
> managing an on-prem instance is (literally) a full time job. Hosting a Docker container is a full-time job? I have worked at several employers self-hosting their own instances without issues or a lot of effort. Many FOSS projects do, t
164.
by Hendrikto 6mo ago
But how much of that do people actually need? Most users don’t use most features. The core MVP is not that big.
165.
by Hendrikto 6mo ago
They can clone the repo, make changes, and then push. On the server, you can have a hook that checks if the commit only contains appropriate issue changes, and apply just those. Sure, a little more complicated than “Create Issue”, but not t
166.
by Hendrikto 6mo ago
Regardless of your opinion on gender and identity politics, surely people can agree that only biological women have periods.
167.
by Hendrikto 6mo ago
As you said yourself: Quantitative easing did not solve anything. We keep kicking the can down the road, and the problems grow exponentially every time. This approach won’t work forever. In fact, we may be past the tipping point already.
168.
by Hendrikto 6mo ago
Well, it is currently cheaper because it is massively subsidized. That will change when subsidies stop. I don’t think it is a good argument.
169.
by Hendrikto 6mo ago
> In general, [managers] prefer a meeting with 20+ or 50+ engineers in one place, each giving 1 minute or longer feedback, because they can do that every day and in an hour, they know what everybody is doing. Most of the time they don’t
170.
by Hendrikto 6mo ago
> A short weekly standing meeting […] is actually a powerful tool. The problem is they never stay short, they never stay on topic, they always expand beyond one a week. Then managers want everybody to say something, so they feel in the k
171.
by Hendrikto 6mo ago
> Stacking more and more complexity into sanitization is clearly a doomed approach. We are more than 5 major revisions deep and yet there are still known holes. People are actively sharing projects on the Scratch website bypassing SVG sa
172.
by Hendrikto 6mo ago
Both have their advantages and disadvantages. git-annex is not strictly better, LFS just chose different tradeoffs.
173.
by Hendrikto 6mo ago
The series of articles linked at the end (troubles.md/posts/wasm-is-not-a-stack-machine/) is even more interesting, imo. Very well articulated and concise critique by somebody who seems to have a great amount of knowledge and
174.
by Hendrikto 6mo ago
What is worse is that for about half a year or so, I now have to authenticate my ed25519-sk key with my Yubikey thrice (!) when using LFS. On every push.
175.
by Hendrikto 6mo ago
Afaik, TPUs are only used for inference, not training. Maybe that was also what the quote referred to.
176.
by Hendrikto 6mo ago
> source: i talked to a GnuPG developer Of course they would say that. See also the responses to your comment: > My honest first reaction to this statement would get me permabanned from this site, so here’s the polite version: > Th
177.
by Hendrikto 6mo ago
Play stupid games, win stupid prices. If you give an agent full reign over your system, do not be surprised when it fucks up.
178.
by Hendrikto 6mo ago
> There is a 1-1 mapping It is strong, but not 1 to 1: > Tubbenhauer computed, for instance, that the invariant uniquely identifies more than 97% of the knots with 18 crossings.
179.
by Hendrikto 6mo ago
I have had that feeling for a while and Ars had some high-profile slipups over the past few months. This confirms that it is just slop now :(
180.
by Hendrikto 6mo ago
> In order to use this, you need to both install the extension in your browser and install a small program (separate from the browser) on your computer. This extra program is called the "native stub". Idk. Sure it has security

More ›