@badlogicgames @andreasklinger "There are many shitty robots, but this one is mine" :)
English
Jon Åslund
3.1K posts
omg that's exactly it. one direction rotates, the other drives forward.
this €9 robot is the perfect leg base for my shitty robot! thank you, china! much cheaper than a custom build.
gonna buy one of these for every kid in the hood, turn them into custom phone based robots. i'm sure the other parents will sacrifice a phone each :D
English
had to use violence. but the legs will be a great base for an old Android phone.
Mario Zechner@badlogicgames
new electronics project. don't uave the time to design my own littlr AI robot. still waiting for my HF robot. so i'm going to frankenstein this €9 one and give it a brain. wish me reversing luck!
English
the most used skill internally at cursor right now
/thermo-nuclear-code-quality-review
- deletes complexity instead of moving it
- blocks files over 1k lines
- flags thin wrappers and leaked logic
- rejects PRs that work but make code messier
English
@DNAutics Are these just set to different refresh intervals and it was a bit stuffy 5 minutes ago, and then you opened a window? Or are they stable at those values? I want to buy one, but I don't know which brand. I guess they all suck? :)
English
@jorandirkgreef @pavan4820 When I read DST in your posts my brain sometimes interprets it as daylight savings time before I realize my mistake. When you run simulations, you obviously speed up time, but is "simulated time" an actual input? I guess it is used for some local metadata somewhere.
English
To start, the more “all in” on TigerStyle you go, the easier DST gets:
- static allocation (often skipped)
- assertions (sometimes skipped)
- explicit fault models (including gray failure)
- narrow interfaces (with weak expectations)
- zero dependencies
But to really do DST your simulator needs to be “Protocol-Aware”.
For example, Determinism, say in a hypervisor is powerful and important, but you really want the simulator not only to be deterministic but also to “understand” your world, your protocol, to be able to reach in and know how far it can push things, where the breaking limit is, even w.r.t. things like optimal erasure coding or strict serializability or minimizing the number of messages for a recovery path.
Then you start to be able to do things like Deterministic Performance Testing (DPT) which is where we’re exploring now.
English
yeah, this has been my experience with turso's simulator too,
it's awesome when it works but chasing false positives is annoying. i have no idea how folks at tigerbeetle managed to do it so well lol.
Chris@criccomini
Spent the past week on SlateDB's DST harness. It was a bit of a slog. The more state I explored, the more false positives I encountered. rng.md/posts/determin…
English
@rough__sea Can a desync attack really come from two valid implementations disagreeing about boundary? Isn't it always at least one implementation is broken?
English
Regarding agwa.name/blog/post/fast…
HTTP is hard to parse - but it's not that hard.
Maybe this is little known, but one of the critical pieces that allowed Node to succeed was a little http 1.1 parser I wrote painstakingly by hand (with heavy inspo from Mongrel and NGINX). I am still quite proud of it.
You can see the first version here: github.com/nodejs/node/tr…
But HTTP parsing is a solved problem now. Even if your language didn't have an HTTP parser - which is very unlikely - you could vibe it up quickly. We really don't need to re-serialize it to FastCGI.
English
Jon Åslund 已转推
Contributor Poker and Zig's AI Ban
kristoff.it/blog/contribut…
Română
@jedisct1 Was it literally just /audit in swival that found all of these?
English
Okay, we have quite a few serious issues to fix in the Zig standard library. Some of these bugs are a little embarrassing. Let’s get to it. The good news is that they’ve now been identified, and they’ll all be fixed in the next release. #zig-standard-library-audit-findings" target="_blank" rel="nofollow noopener">github.com/Swival/securit…
English
Swedencore should be a thing for tech lingo.
JWT = Yi Ve Te
Yavaskript
English
Interesting way to manage new issues and pull requests in the age of agents. I hope it works out.
Mario Zechner@badlogicgames
People of pi. The great @steipete has graced our repository with a bespoke slop PR to fix cache affinity in the OpenAI Responses provider, which should lead to better prompt caching. And the new "pi contribution model (tm)" is now live. Here's how it works: - If you send a PR, it gets autoclosed, unless you've previously been approved by a maintainer. - If you send an issue, it gets autoclosed, unless you've previously been approved by a maintainer. All auto-closed issues are triaged daily. - Issues that follow CONTRIBUTING.md and are worthwhile will be reopened. - Issues that are exceptionally well written get an "lgtmi" comment from me or @mitsuhiko, which will approve all your future issues automatically. No more auto-closing. - Issues that are well written AND offer a PR get an "lgtm" comment from me or @mitsuhiko, which will approve all your future issues and PRs automatically. No more auto-closing. I, the idiot who has to go through all the closed slop daily, mark the last issue I processed with the "last read" label. If your issue is below that and hasn't been opened, then it did not meet the quality standard. You may or may not receive a reply on why the issue was not opened, depending on my time and mood. Accounts that: - Let their agents slop a book into the issue tracker repeatedly - Otherwise behave badly will get their account blocked across all my repositories. no exceptions. not takesies backsies. I get anywhere between 30-50 issues per day. Most of them are agent garbage. This is the only way to keep me sane and ensure the issue and PR trackers have actual good signal.
English
@ID_AA_Carmack Elo is not an acronym. It is named after its creator Arpad Elo. en.wikipedia.org/wiki/Arpad_Elo
English
So many judging tasks could be improved by aggregating partial orderings, and in the limit, just ordering pairs.
The annual Libertarian Futurist Society novel awards discussion is starting, and while I would like to participate on some level, there is no way I have time to read an entire slate of novels. However, I will likely read at least two from the list, and I could give a relative assessment.
This cries out for the use of something like ELO ranking, as in chess competition, perhaps with some suggestions to get sufficient coverage.
Peer and out-of-chain employee performance calibrations could probably also benefit from a greater quantity of sparse pairwise comparisons
English
Jon Åslund 已转推
I think this is the best post I've ever made.
ngrok@ngrokHQ
Quantization can make an LLM 4x smaller and 2x faster, with barely any quality loss. But what *is* it? @samwhoo crafted a beautiful interactive essay explaining it from first principles, aimed at coders, not mathematicians. ngrok.com/blog/quantizat…
English
I don't want any food to expire until 2032 and definitely never during the first 12 days of the month.
English
@eatonphil @clattner_llvm Those articles also made me more understand the how in "Let's not work forward from Python and try to make Python a little bit better. We're saying, let's work backwards from the speed of light of hardware." x.com/sw_unscripted/…
Software Unscripted@sw_unscripted
It's the 💯th episode of Software Unscripted!!! 🎉🎉🎉 I talk with @clattner_llvm about his new language Mojo, @roc_lang, and lots more! 😃 YouTube video: youtu.be/ENviIxDTmUA Ad-free video for Patreon members! patreon.com/posts/mojo-wit… Audio version: pod.link/1602572955/epi…
English
@eatonphil I was confused when Mojo was announced. Not yet open source python++? Why? After reading modular.com/democratizing-… by @clattner_llvm it started to make much more sense.
English
I've wanted to look at Mojo for quite some time. It is definitely not yet a superset of Python. But with that seeming to be a goal and with the standard library being open-source there is a very clear area for you to go and contribute to a well-funded major programming language.
English
@jorandirkgreef @Steve1885204 Yes, I agree. I remember I saw this first with Oracle Database over 20 years ago and I didn't understand back then why they would do this.
English
@jooon @Steve1885204 So it’s as easy (or as hard) as writing a filesystem, but growing up on FAST conf, I’ve always personally always seen filesystems and databases as pretty much the same bone structure.
English
Nice to see TigerBeetle becoming an application standard for Linux filesystem benchmarks.
“XFS gained the most out of TigerBeetle with the succeeding Linux kernel releases and closed the gap with EXT4.”
phoronix.com/review/linux-6…
English
@jorandirkgreef @Steve1885204 Is it as easy as just pointing to a device in /dev instead of a data file?
English
@Steve1885204 Yes, TB in fact can already run on raw block devices to bypass the filesystem directly.
English
@mitchellh After nudging, that is what my Claude session did for that fontconfig bug you fixed the other day. It saw some overlap of symbols in the dynamic symbol table and then found the system 2.5.0 source and noticed that the FcConfig struct had removed a field, explaining the segfault.
English
Ahhhh, Codex 5.3 (xhigh) with a vague prompt just solved a bug that I and others have been struggling to fix for over 6 months. Other reasoning levels with Codex failed, Opus 4.6 failed. Cost $4.14 and 45 minutes. Full trace plus includes original issue: ampcode.com/threads/T-019c…
I know this prompt is relatively bad. Honestly, our stable release is in a week, and I was throwing some Hail Marys at the frontier models to see if I could get a clean, understandable fix for some of these bugs. By using `gh`, it grabs much better context from the issue, so its not terrible.
The best thing that Codex did was eventually start reading GTK4 source code. That's where I ended up (see my GH issue), and I knew the answer was somewhere in there, but I didn't have the time or motivation to do it myself. The other models never went there, and lower reasoning efforts with 5.3 didn't go there either. Only xhigh went there. I think that was a critical difference.
The final fix was decent. It was small, all in a single file, and very understandable. It had one bug I identified (you can see in the trace), and then I manually cleaned up some style. But, it did a great job.
Definitely an "it's so over" moment. But at the same time, it feels amazing because now our next stable release will have this fix and I was able to spend the time working on other fixes as it went.
English
@googlephotos Why can't I continue an edit? Until today I could crop an image, save, then edit again, continue edit, save. Why has that ability been removed? I did not add any effects. I cropped an image.
English