billyzs

I already prefer using pipe as separator in logging; now you're telling me there is a chance that my logs can be automatically ingested as tabular data? Sign me up for this branch of the multiverse :)

You might be interested in https://draradech.github.io/jigsaw/index.html

Used its output with a laser cutter and plywood, result was neat

To me the chess AI example he used was perhaps not the most apt. Human players may not be able to reason on as far a horizon as AI and therefore find some of AI's moves perplexing, but they can be more or less sure that a Chess AI is optimizing for the same goal under the same set of rules with them. With Reasoners, alignment is not given. They may be reasoning under an entirely different set of rules and cost functions. On more open ended questions, when Reasoners produce something that human don't understand, we can't easily say whether it's a stroke of genius, or misaligned thoughts.

if kernel is updated the OS would have to be rebooted anyway unless live patching is configured. Rebooting after an update is probably more common & less annoying than one would think.

a2, c = a2+c, c+2 is faster than a2 += c c += 2 My guess is that in the first case the two evaluations and assignments can happen in parallel, and so may happen on different cores

Not sure I follow, isn't Python single threaded by default? Changes to GIL is coming but does it change how the interpreter uses CPU?

I just wish we had something like Matplotlib in R

plotly could be worth a try, i use its python bindings and much prefer it to matplotlib, but i don't know much about the quality of it's R API

There used to a standalone Firefox Password app for mobile but it got the axe.

Yeah, let’s not forget https://www.thebureauinvestigates.com/stories/2021-12-06/swi...

I’m not under any illusion that there’s any “Swiss exceptionalism” when it comes to privacy.

RDI for added sugar itself is 24-36 grams which a single 12oz can of Coke exceeds. Measuring in Calories, it's recommended to limit daily intake from added sugar to below 100-150 Calorie. [0]

% of Caloric RDI isn't the whole story. A can of Coke is less of a blip on the radar as you suggested.

[0] https://www.hsph.harvard.edu/nutritionsource/carbohydrates/a...

I use worktrees, but some builds which uses git hash will fail in worktree checkouts, thinking that "it's not a git directory" (forgot the exact error, it's been a while); haven't found a solution to this

My personal experience is that the situation has improved somewhat but the hardware still plays a large part. A Kaby Lake laptop I had setup hibernation to encrypted swap (not file but partition iirc) worked well enough besides a few times when some peripherals e.g. wifi failed to power on after resuming from hibernation; A Zen3+ laptop from the same manufacturer that I got last week, running Arch (with the latest kernel and packages) _always_ had problem resuming, wifi (and who knows what else) never gets turned back on, the system sometimes enter into a weird state showing just a blank screen, not accepting any keypresses, or otherwise get stuck in some stage in the boot process with no output so it's hard to debug. So far it's been a real crapshoot.

wanted a dictionary app that tracks the words I looked up too. ended up writing an iOS short that queries an online dictionary and posts the word to an ifttt webhook which then adds the word to an airtable. with a bit of manual work, the card view in Airtable can even function as a rudimentary flash card. it's pretty basic and never going to have fancy features such as spaced repetition, but hey not bad for not writing a single line of code

User name checks out

why the double standard?

What would it take to get a self-hosted version of this, preferably using FireFox on the server? How hard would it be to get plugins working as well? Would this be essentially a way of using Blink/Servo based engine on i(Pad)OS?

Hi Peter, if H1B holders who obtained work authorization in 2019 becomes full time student to pursue a MSc for 2 years, graduates in 2023, will that work authorization still be valid for them and allow them to seek employment in the US in 2023? Does it matter if the study happened in our outside of US? Thanks!

afaik there's no native Zoom client for ARM [0]. You'll be stuck with the web version of it, which according to some functions poorly no matter where you run it

[0] https://devforum.zoom.us/t/have-you-considered-building-zoom...

If they did that, then hopefully they learnt a thing or two about networking :)

But I do wonder if bandwidth limiting certain types of applications would be more effective, like reverse qos. There was a post a week or so ago on HN about adding delay to websites that sucks away productivity.

edit: I went and did that on my home router (a ubnt ER-X) and was pleasantly surprised by how granular its DPI and QoS categorizes applications (by protocols, by domain, IM, social network, P2P). It even allows you to make your own categories of apps that are (ab)used often, and rate limit access to them on a range of local IP. I'm hoping that this would be more subtle than an outright block, and the not so instant gratification lead to voluntary reduction of mindless consumption.

I really don't understand this article's high tendency to slip in tangents that have little to do - long sentences enclosed by brackets, or worse, hyphens which made side stories hard to distinguish from main story. Those things have so little to do with the actual subject of review and serve little purpose. Perhaps they're there to show off the author's knowledge on the subject matter?

Perhaps I'm too used to engineering style writing which just cuts to the chase. Is it common for book reviews to include those bloats?

https://warisboring.com/test-pilot-admits-the-f-35-can-t-dog...

felt like I've seen this complaint about the performance of the JSF not too long ago. Maybe things have gotten better?

Gilbert Starng also has an excellent series on Linear Algebra on MIT opencourseware [https://ocw.mit.edu/courses/mathematics/18-06-linear-algebra...]

I'd actually prefer if employers are candid about their preferences, as much as they may disappoint me. At least I know what to expect.

Con: a Bachelor, Masters or PhD in {a specific field}.

Consider rephrasing to "PhD with publication record, or MSc with 3 years of work experience, or BSc with 7 years of work experience, 3 of which at {a specific advanced level}".

I understand if the job requires a certain minimum education background, but how can you possibly consider holders of all three of the above degrees, without stating the specifications for each degree? Can I still apply if I'm just about to get a BSc, or would you require some amount of work experience? Am I competing directly with PhDs? Confuses the hell out of me. I once applied for a SWE I position that had that kind of degree requirement during my senior year. Passed two phone screens, even handled questions about computer vision and deep learning well during the on site round (answered beginner and intermediate level questions well, was honest to admit my limits when asked about advanced topics). Two days after the interview I went back to check the job posting, and it was somehow changed to say Principle Computer Vision Scientist then I was like "oh there goes my chance". Needless to say I wasn't selected. It makes sense for them to select candidates with advanced degrees if they are really looking for a Prin. Scientist. But please, spell out the requirements so that candidates make informed decisions about whether to apply. It saves both parties valuable time.

I agree with you in the sense that if a known and reliable way to map knowledge and information from one domain to another (e.g. from desired trajectory + perceived current position to steering inputs), I'd much prefer that than black box ish neural nets. Neural nets aren't meant to be the silver bullet.

But in this case though, any kind of state space control also requires rather precise knowledge of the physical laws that govern the dynamics of the vehicles. When such information is not available, can neural nets do a decent job at mimicking an analytical control algorithm? I think that's an interesting problem worth exploring.

I see, that makes sense. It'd be indeed worth it if we can apply a model trained on controlled environment to a more challenging one with little to no modification. Good luck with the project and keep us updated!

Not to put down the OP's work (I think it's a great project), but I'm just wondering what advantages might an ML approach have over "traditional" CV algorithms. In a really well controlled environment lanes will be easy to detect, and computing the difference between the current heading and lane direction should be doable; maybe if we're talking about complex outdoor environments and poor sensors then ML would have an advantage? Or if we're teaching the robot what the concept of a lane is?

I think back to the days when I basically implemented lane following with an array of photo resistors, an Arduino, a shitty robot made from Vex parts and some c code. The problem is much simpler than the one presented in this article, but then the computational resource used was order of magnitudes less. At what point then, do you decide that "OK I think the complexity and nature of the problem warrants the use of ML" or "Hmmm I think neural network is an overkill here"?