philipphutterer

I agree to others that the intent of this study could be written more expressively, but honestly, doesn't this show exactly one thing to the people in the tech world? We need better education and communication for people without technical knowledge about what to use which AI models for and what NOT to do with them. For me, quite often I try to give quick help and information on what to expect from an LLM for given input whenever someone non-tech close to me is running into unexpected output. AI just seems so simple and non-complex to most people, it's shocking.

I've never tried LibreTube. I am used to the new pipe interface and had all my subscriptions managed their so the move to Tubular was easy.

I prefer Tubular with Sponsorblock

Or given the number of unscanned books, even just give it the controls for a book scanner, the books and probably some robot arms. Then let it figure out the scanning first in some layers. Shouldn't be that hard.

Not sure about this, but training a model on the website that displays the code is not quite the same as training it specifically on just the code. Moreover, (raw) repo content files might not even be included in crawled datasets (e.g., look at https://gitlab.com/robots.txt). I think there is something specific to GitHub as it being part of Microsoft that makes processing that data much easier.

See also:

https://news.ycombinator.com/item?id=34765093

https://news.ycombinator.com/item?id=35037374

https://news.ycombinator.com/item?id=35219182

https://news.ycombinator.com/item?id=35317576

    the whole planet upgrade their electronics at least every decade is just an obviously stupid thing and not enough people talk about it

This is so true! How can we developers make this a bigger thing? Isn't there a market for this by now?

This! And I thought I'm just too stupid to use this search, ha! Search is hard, for sure, but come on. But you actually reminded me to just use Google Dorks for that with inurl:https://f-droid.org/packages/.

There was some information only recently on HN about reversing those Electron Shelf Labels [0] that was quite interesting, but your comment makes me wonder if you could eavesdrop these wifi price updates in such stores. Also, searching for that term on HN gives some other fun projects.

[0] https://news.ycombinator.com/item?id=34738649

Is that really practical? In the video it looks not very useful to me as it does not seem to be very efficient. The mouse is so integrated into MacOS it looks hard to efficiently replace it. Are you or is anyone else actually using this regularly?

Also, I think the other mentioned alternatives look quite similar to that regard.

I use caret mode, visual selection and P quite often to search for the text in visual selection in a new tab with the preferred search engine. I also like yf for yanking a link to clipboard.

Cool, I like the idea and I actually use this feature a lot in Google Podcasts.

However, I believe there is a way to do this with a single ffmpeg command. I remember watching lecture videos with higher speeds wondering about the huge amount of silence that made it still pretty annoying. I didn’t find anything simple to trim silence on audio and video combined and finally only listened to the audio with trimmed silence using ffmpeg.

Edit: It's literally just an audio filter in ffmpeg, see https://stackoverflow.com/a/29411973

It didn't say so in the post but I suspect 8GB RAM were just not enough and an upgrade was necessary anyway? And future upgrading was mentioned. In my opinion this is a valid point actually and should be very much considered when buying any of the modern all-fixed-no-upgrades machines, I mean its not just the Apple machines.

> I'd prefer my victories to be against the frontier of knowledge, and to win universal truths.

You wouldn't need to tear down barriers if the people that built them thought the same in the first place. Nonetheless, keep up that attitude.

> However you can also get those videos by logging in, which the author doesn't mention.

> Also not mentioned in the article is that you can actually take the throttled URLs as is, and download pieces concurrently for a pretty good result.

The author mentioned both, the login option as well as the chunking mechanism. Sorry, but did you actually read the blog post?

This is really cool to see. I don't have any problems or bugs using it, but it seems like the 3d rendering of the virtual hands is not quite as fast as the tracking itself. I wonder if and how this could be used in a meaningful way. As a feature for things like Google Quick Draw it would be fun.

The idea is just great. Keep in mind what people expect from this type of website (and clear up any misconceptions), as this may lead to unnecessary churn. I would like to see some degree of customizability, e.g. weighting of features or something like that. (Nitpick: wrap the play svgs in a button tag like the "similar" button)

True, from a technical point of view, this should actually be quite simple. But I'm referring more to the administration overhead caused by the large amount of small banks that are considered their own credit institution. In a lecture, we were once shown a chart comparing the costs of financial operations in the European countries. Germany had by far the highest costs. This mess will not survive the next 10 years.

I didn't know about Nordigen, but how are they able to support the absurd amount of 1222 banks in Germany? Looks like a good indicator how broken the digital world of banks here really is.

lmfao, this is so hilarious.

That sounds like this teacher understood what teaching means other than presenting the curriculum. Love it.

Perhaps when the standard of software development shifts towards using AI for everything?

Since the results of each operation can be manually annotated (helpful, not helpful) maybe the intention of this project is more to use the data as a human feedback loop? Therefore such fine-grained data is of course better.

Other than that, it might also be useful for the user when either of both operations is not good enough yet and error propagation might be worse than human correction between the two.

it does not only break your bookmarklet, but even vimium [1] is unable to focus the language selection. this is awful.

[1] https://github.com/gdh1995/vimium-c

How about websites that are browsed http first and then redirected? People might browse for a domain without the https prefix for convenience (or old links) and the browser defaults to http.

What do you think is the reason for that "orders of magnitude" in growth? Can one experience similar without working for a big tech? Is it a mental thing or rather the working environment with many highly qualified people around?

how many legit users are turned away when asking for CC vs how high is the abuse ratio? maybe try to find out when the legit users turn away. maybe they are not satisfied with checkout process?

I didn't have a look at the code yet but I'd love to have repos to learn from that are actually focused on high quality code more than a "good-selling" product.

well summarized.

its easy to make bad SPAs nowadays. imo more time needs to be spent in the tool/environment/framework/language selection and evaluation process.

also, many SPAs cant even handle basic forms correctly anymore, because developers learn more javascript than html5. like the author pointed out, rebuilding features in js that already exist, just because one is unaware of the fact that it exists or what the underlying design choices have been, often is a bad idea.

are there actually still people that care about noscript? i mean, i get that most people dont care and js is enabled by default in most browsers, but the no-js-web-world is also quite interesting.