I'm reading through the documentation and I'm not getting the value prospect from using Crystal.

Is it essentially "Ruby-like" + "easier access to performance"?

Apologies if this is too reductive but https://news.ycombinator.com/item?id=41683815 shows there are multiple concerns about why one would use this language in the current age.

What does Crystal bring to the language decision tree in 2026?

I am mostly using it for hobby projects by now, but for me it is the very transparent yet mighty type system in addition to the aforementioned aspects. Its macro system is a sweet spot in comparison Ruby’s metaprogramming, which can easily be taken way too far. Also, if you do it right, you can build containers from scratch that are small in size and boot really fast. It has a really well curated and extensive standard library that lets you get away with including a lot less dependencies. It comes with a linter, a test runner, docs generator and quite useful debugging features.

I am looking forward to what AI can do to port Ruby gems to the ecosystem.

A very well done type system + native code performance, with the ease of use of Ruby is a massive selling point, IMO.

"like ruby but you can ship a single native binary" is already huge.

Any social space will break down upon reaching a critical point in representation of the general populace.

I have no idea about the development however.

It's been years since I last used `slurm`. Thanks for the blast from the past.

I don't understand why people are so excited for Mojo. I don't get the impression it will replace anything but computational scripting that was done previously in Python.

HPC is a different beast as far as I'm aware.

I don't think that's accurate. Mojo is explicitly compiling tensor graphs that run on accelerators. it's not like PyTorch where python is providing the chassis but not the engine.

I don't think its going to be a good general HPC language just because its targeting a specific set of AI workloads, but they have shown some examples of synthesizing code which is comparable to hand-written kernels.

but its not out of the question from first principles

Additionally, as Intel, AMD and NVidia embrace Python JITs for their GPGPU, across all three major desktop OSes, the Mojo value becomes devalued.

Why rewrite in Mojo, if Intel, AMD and NVidia SDKs give first class support for Python, including writing kernels in a language subset.

The sites seem to be all of the same format-- with different color palettes as the main offering point here.

A little misleading, as I thought they would showcase different types of site design entirely.

I'm completely new to this space, but how are applications using tailscale as a library?

Are they creating their own mesh networks for internal or user use?

Imagine something like writing a server with an /metrics HTTP endpoint that Prometheus can then scrape -- but you bind it on separate port only inside a tailnet, with an ephemeral tailnet key and name it "metrics-service-blahblah".

Now you can simply write a script that uses the tailscale API to find all "metrics-service-*" nodes in your tailnet, and then adds their IP/DNS to your prometheus scraping list. Run it every 60 seconds. Done, now you can just deploy your app anywhere on any cloud and it will get scraped and that route will never be exposed to the outer internet.

This will basically just let you attach bespoke applications and not just "computers" to your network. I suspect I will get a lot of use from it.

Tailscale and Wireguard are great. I'm an OpenZiti maintainer and I've written/spoken about application embedded zero trust for many, many years. Still it seems most devs don't think it's important for whatever reason... It'll make me happy if Tailscale is successful here and can spread the word out to get more devs interested in embedding the secure connectivity directly into the apps instead of relying on the classic underlay network and bolting on security. If that sort of thing interests you, you could check out OpenZiti. It's not Wireguard-based for better or for worse you can decide (if you do end up checking it out)

Just speculating, but that it's an option to open/listen to a port, but that port is on a Tailscale network. So the app is largely unaware of the encryption over the top. Similarly, you could do similar for a client app. Where the Tailscale connectivity options are inside the app, instead of a proxy to the app that lives outside the apps.

Likely more transparent than explicit/implicit TLS.

Agreed. The name collision nowadays is horrible.

Then again, it's frustrating trying to name a product in today's era; too many names are taken.

Here's another explanation. My hypothesis goes deeper than gender imbalance.

Most job interviews are theater nowadays. It's about conformity, performative culture fitting, agreeableness (read, willing to slave away without complaining).

On average, women tend to better suited for such processes. Along with immigrant groups.

> On average, women tend to better suited for such processes. Along with immigrant groups.

This is quite the claim, and I would understand anyone taking deep offence by it.

Can you substantiate it?

Not the commenter, but my take:

Women are taught from a young age of how to engage with their role in a patriarchal society. This includes caretaking, nurturing, and having patience for men. It's not uncommon for young girls to be given household chores that they must "suck up" and do while they're brothers get away with not doing chores.

In addition, we associate politeness and pleasantness with femininity. A woman ought to be kind and softspoken, and must not swear or raise her voice. We associate masculinity with abrasiveness, stubbornness, arrogance, and rage.

In most jobs and, in fact most social processes, the former is more useful to take advantage of. This could be why women, on the whole, would do better in interviews.

In addition, it is well-known that women socialize much more, including more conversing. They are probably much better at being social on the whole because of that. Meaning, women are much more likely to be charming and likable than men.

Please further explain this unabashed racism and sexism.

Schrödinger's immigrant, both actively refusing to conform to the host country's culture and best able to show host country's culture to get jobs

Because it’s based on human slop. It’s simply the student.

I mean, the CAD software, even though visual, has to retain some internal state of the model, right?

It's possible, but I don't think there's enough pressure for these companies to support exposing that state to the user.

It does but the internal state is not going to be very meaningful in text form. Edge 4826 connects to vertex 7264.

Think about SVG - you can do some simple stuff procedurally but you're never going to create something like the SVG tiger in code, nor would it be remotely useful to read the source SVG for the SVG tiger. Most CAD is like that.