fiiv

I do this. Ollama makes it very easy - just pull the model you want. The great thing is the ability to test them in the same tasks, there's a huge difference in comparable models.

You can set it up in your editor of choice - I use Zed, and it's just listed as one of the providers you can choose.

In terms of performance - it works decently well.

Or it could be mostly people fed up with working under shitty conditions.

I remember trying out Skeleton many years ago. Skeleton's Github repo has not seen an update since 2014. A fork of it can be seen here:

https://skeleton-framework.github.io/

https://github.com/skeleton-framework/skeleton-framework

But it too has not seen any activity in several years.

Seems to be working for me (firefox, safari, chromium). The SSL cert appears to be valid, expire next year and be issued by Amazon.

It could be a block by your country, ISP or possibly if you're using a VPN it might be affecting it as well.

Yes indeed! I'm not surprised there's some difference. One thing I learned doing this is that there's a lot of choices to make when selecting the best icon.

For example, some sites have SVG icons, which are usually superior in both file size and in scalability (compared to raster formats), but if using them in a mobile app in React Native, it would require an additional step to integrate SVG images.

Whoa, I hadn't thought of a sprite sheet, but that's a great idea actually. It shouldn't be too hard to generate one of these. How many favicons in one sheet are you thinking?

Hi HN,

I was working on an app, and in it, we wanted to render an icon for links from meeting invites. This turned out to be quite a pain!

I found other things that could return icon data, but alas they were all JSON APIs. I didn't relish the idea of fetching an endpoint for each and every website icon I needed to load (and write glue code to render the best icon out of the bunch).

You'd think favicons are as simple as loading "example.com/favicon.ico" but there's a lot of subtleties to it.

First, favicon.ico's are often really low res – there's often higher res icons available via tags in the HTML or through the manifest file.

Second, what if the hostname is unreachable or if there is no favicon at all?

And third, no one wants to parse a whole HTML page just to get one little icon!

So that's why I made Icon Horse – it automatically parses the site's HTML to determine the best possible resolution icon. And if no icon is available, it serves a decent fallback, so you never end up with a broken image. And finally, it's dead-simple to use. You simply use the hostname like so:

https://icon.horse/icon/ycombinator.com

Anyway, hope you also get some use out of it. I'm around if anyone has any questions :)

This is awesome, going to give it a try!

They're kind of different things. Graal is an attempt to create a polyglot environment. For example, using Python and Javascript in the same environment. For example you might be working on an application where the web server managed by the web team is written in Node and your data scientist team uses Python scripts to deal with the data, but they need to run in the same environment. Graal lets you re-use that environment for both.

Node and Deno are dedicated runtimes to run Javascript using the V8 engine. Both provide a standard library of additional functionality to make them worthwhile beyond just browser code.

> Tech companies (Google, Apple) don't custom-tailor their maps because they want to. They do it because the states use leverage to force them into it. And the states do it to promote the idea that the disputed territory is theirs.

So what? Google wants people to use their map software. By showing variations of the map depending on what the official stance of the regional government is, they are ensuring that their map software is available in that region and fits the context of that country.

There is a Russian narrative for Crimea being theirs. There is a Ukrainian narrative for Crimea being theirs. It's not Google's place to decide between them, it's their place to create a map people can find things on.

It does not matter what you, I or Sundar Pichai believe is the truth of whom Crimea should belong to. Their product is a map and it's available in Russia, Ukraine and Crimea. In fact, by presenting all the credible alternatives in this case (for Russia's side, they currently administrate the region and for Ukraine's, it was very recently their territory), they are being apolitical about it.

I'm fiiv, the dude making this. Would love to hear your feedback and if I'm doing the right thing here or not ;)

PS. uh oh, missed the "f" in frameworkless in the title.

"Blockchain" is like "cloud" as a term – it is ambiguous. It means different things to different people. Not everyone understands what a blockchain is – a lot of people probably have about as much understanding of it as they do cloud. I guarantee a lot of people who even give talks and consult on blockchain tech don't know what Byzantine fault tolerance is.

I think the interesting factor here is not political at all. It does not matter if you agree with the sanctions or not. The fact is that clearly Iran sees blockchain tech as a way to circumvent US sanctions on conventional financial networks.

If they were smart they wouldn't just launch their own though. There are plenty of crypto-currencies already out there, and if they need to find a backdoor into sanctioned markets, then using one of those makes more sense. Otherwise, this is going to go the way of the Venezuelan Petro – aka essentially an exit scam as far as we can see at this moment.

I use a variant of Chromium, the Eloston/Ungoogled Chromium build.

I find Chrome just an overall better experience to use. I am kind of a stickler for UI and to me, Brave's and Opera's are a bit clunky even though they're based on the same VM.

As for Firefox, the experience does not have far to go in my opinion but I have had a lot of issues with memory leaks and crashes lately so I switched back to Eloston Chromium.

I support transparency on this topic but come on. At some point, Facebook was going to block automated simulated click events done via browser extension.

> The last thing I want to have to do is use Blogsend to send a "Notice of data breach" email! :)

No one wants to have to send any of those emails, least of all me/Blogsend ;)

> One think you could check out is adding security HTTP response headers - your site is pretty clean and simple so it should be relatively easy to get it locked down tightly. See https://securityheaders.com/.

Thanks for the tip! I've added this to my todos!

Yes indeed, MC does this pretty decently. I tried it myself, but I found the product is very much catering to marketers. Which is fine, of course, but for people that only want this functionality, it seems like overkill.

Blogsend is meant as a much simpler, more focused alternative. This is all it does, and the features to be added in the roadmap are almost exclusively to help bloggers, not marketers or a general audience.

Good luck :)

Hello!

Looking at the spec, it appears to differ from RSS/Atom in very small ways (besides being in a different data format). It would be pretty trivial to build in support for this.

It's difficult to find usage stats since searching for "json feed" unfortunately gives a whole lot of unrelated things!

I think if I can find a paying client that wants it, I will build this in. Wanna be the first? ;)

That's a fantastic idea actually, thanks so much!

You're right, WP does have several plugins that do this, the most popular being from providers like ConstantContact and Mailchimp. I think I'd expect a few WP users but you're right, my target are probably people that are using something like Jekyll.

That makes sense!

Just out of curiosity, what size is your blog?

I spoke to several blog owners of various sizes. The smaller ones tended to not care about this, but indeed the bigger ones took more of an interest in the actual emails they collect. Just wondering if you'd fall into one of these categories.