danbulant

Too fast of a compile time, different versions would recompile the whole dep tree. And even just the end code would take more than .02s

They do have their own index which is used to detect how many ads the website has etc and for small web / blog searches, but they use meta results to have more results.

Also, they do have integration with multiple LLMs - ChatGPT, Claude, Gemini, Mistral and their own

They sort of do that. Next to link, you have information about how many ads were detected in the site, and sites with more ads are lower ranked. They also have an index of their own which only allows sites with few ad detections (they use number of blocked things by ublock, so it also counts tracking links etc)

At least 62% of visitors on my PWA come from Play Store, with 20% being labeled "unknown".

It really depends on your app, but many mobile users just search for stuff in Play Store first before searching on Google, and even Googling it will help as you will get a widget about your app (sometimes). The web version is (for the most popular queries) on average 1.6th in Google results.

Although the dataset may not be very large (10k+ installs shown on Play Store page).

Windows KRunner with slightly improved styles?

That means that multiple arbitration cases cannot be joined together - which is actually worse for them, as they have to pay a fee per case, so more cases means more fees to be paid.

It'a weird that this isn't normal. I bought an "AlzaPower" cable, which is just a rebrand of some chinese manufacturer by a local reseller ("czech amazon"), and they have their wattage as well as speed on them (100W/5Gbps). Then also a samsung usb c cable, which was pure white, has nothing on it and doesn't seem to support any fast charging or higher speeds..

In the post they mentioned that there can only be a single upgradeable read to prevent this.

In which way does svelte(kit) not have first class support for animating between states?

You can animate items being added to page, removed from page or reordered in a keyed each. This also includes pages themselves in layout files. And you can also animate an item being moved from one place to another by being removed in one place and added in another place.

Arch on 2 laptops, 2 VPSes and a minitower desktop (used as a home server).

So far everything good. On the servers, everything runs under docker, I just installed docker, nomad, consul and tailscale, which I'd have to add custom repositories for ubuntu anyway for basically the same end (except now I also have newer kernel).

Tried Debian and Ubuntu in the past, old packages (especially GUI applications, but other stuff too) often broke, and it sucked to just find that there's a fix but not in the repos - and so I had to add custom repositories for every other thing anyway.

This solves the problem of pages full of ads and keywords and irrelevant content. It doesn't solve the problem that's talked about in the article - there are other solutions, usually alternative search engines (kagi[0], or if more AI=more better, the kagi labs with contextual answers with sources[1]. Not sponsored, just like using them. Similar service offered by Hey.com, but I didn't try them much).

[0]: https://kagi.com [1]: https://labs.kagi.com/ai/contextai

Github copilot is pretty good at writing tests

They did for obs[0], didn't see any other projects (I just remember I saw it a while ago, I may have missed other software if they tried).

[0]: https://news.ycombinator.com/item?id=30112595

Not sure about US, but for example in Czechia, you can download any content legally - as long as you don't share it, it's fine. Torrents do share it though, so that is a gray area.

You also have a right to make a personal backup/copy of the content, again, as long as you don't share that copy (projecting for multiple people is also sharing in this context).

I don't see a difference with fonts.bunny.net and Google Fonts? From the perspective of GDPR, they still both receive your IPA and browser info, which was what made it illegal to use G fonts.

You can download fonts from both of them and host them on your own server to avoid processing that information.

I'm all for privacy and such (I don't use Google Fonts or tracking ads/analytics), but I don't really see a difference in here. According to G Fonts privacy policy, they don't store any PII. (This also reminds me that based on what I understand of the ruling, almost any 3rd party requests for assets should be blocked? Including G fonts, bunny fonts and jsdelivr)

This seems to work the same svelte kit does? Except it's react and not svelte.

You can make your own local extension. People will just write easy tool to generate an extension out of userscript(s).

Still not good that they removed that...

But again, GIPHY has the exact same feature, and I can find Giphy just fine

I'd add to try using sveltekit.

Typescript is already supported when you set it up using the documented "npm create svelte@next".

It has cool routing, backend etc but even if you don't care about any of it, you get free hot reloads. And with the new version, you can create it as if it was your average svelte app, just avoiding + in file names and using +page.svelte as your main entry point.

For styles, if you care about styling a bit more than default styles (mvp.css or bulma/bootstrap/similar), I'd recommend using tailwind's better clone windicss, which has great svelte integration (including editor extensions for autocomplete and highlight), and writing styles is quicker than using normal css.

You can also use scss/sass/similar in style blocks, you only need to install node-sass and set "scss"/"sass" as the lang of the style block.

I also think there were some companies that use gpt3 for analyzing text, like reviews or posts (for analytics like do people talk about a product in a positive matter?).

To add, github copilot is a really clever autocomplete that makes some mundane tasks much quicker. Things which are too small for a library but are still fairly often used can be "typed" more quickly.

0xSOMETHING is just a single word so that will usually not find something unless they have an index with all the generated codes, which is something they probably don't want.

How do people find 0x.co to get the data out of the code? Maybe the code should be written like 0x.co/something instead of just 0xSOMETHING?

Fairly sure that's one of the reason for this. Had similar issue with Discord - removed the ability to add my Discord bot to new servers but refused to provide any reason or fix the issue. But unlike here, they said they're eager to see a new bot instead..

Kind of reminds me of sveltekit. For example, I hate writing the same HTML repeatedly, so I just used svelte for it's components and made a fully static, JS-less page. That later evolved and left out the JS-less (some fancy effects, dialogs, cross-page transitions etc), but even for building static pages seems far better than writing normal HTML, when those pages become a bit complex.

As for the last point: everything that logs will eventually run out of space, be sure to have some kind of automation to rotate those logs. Had the same problem with MongoDB and QuestDB as well, and both had logs that were far bigger than the actual data in the database.

> I can accept the pain of using new.reddit.com to login. But as soon as switch to old.reddit.com it shows that I am logged out? What are the product dev teams aiming for?

Cross origin cookies are maybe being blocked by your browser

Hello, author here. I'd like more simple comparisons like these, so if anyone knows about anything useful to compare in a quick way, please do share it as well. Also, I'd really like any reviews about the post and site (what can be improved, what's bad, what's good etc..), thanks!

Can someone compare this post to mine (1)? I made a highly similar post 2 months ago and got just 2 upvotes and no comments.

[1]: https://news.ycombinator.com/item?id=30869137