skarlso

Hey folks.

I thought I'll share my flow in how I use LLMs in my everyday work. I tried to write down my flows and some of my settings to make it also a bit more fun to work with.

Hope this helps someone. :) Cheers!

Dear HackerNews community!

We are proud to announce that [external-secrets-operator](https://external-secrets.io) is now officially GA! We just released version [v1.0.0](https://github.com/external-secrets/external-secrets/release...).

This version is full of fixes and a couple of new features, however, the point is to go stable. Now, with v1.0.0 we can follow the semver guidelines in breaking changes and patch and minor releases. Hopefully, this provides users and organizations the confidence to invest into ESO and to invest into keeping it up and running. Does this mean the project is on maintenance? No! We are working today to improve ESO in a lot of ways.

One that has been highly requested for a while now and much anticipated. Generic target types for ExternalSecrets (ES) objects. ES objects are now able to create whatever object you want, not just Secrets. ConfigMaps, Deployments, Applications, etc. Anything goes as long as you configure the right RBAC settings. Please read the usage page carefully and note that this feature is in heavy _alpha_. Link to the documentation can be found [here](https://external-secrets.io/latest/guides/targeting-custom-r...).

We are also working hard on [Out-Of-Process](https://github.com/external-secrets/external-secrets/pull/55...) providers. One of the changes involved in that effort is introducing go modules for all providers. This means that eventually, people will be able to build their own ESO binary with _only_ selected provider(s). This will prevent other providers bleeding in their vulnerabilities into your provider. If you would like to have an ESO that only supports AWS you could build it with `go build -tag provider-aws` or something similar. We wanted to make sure these changes are understood by the community as the future of ESO - thus, nothing will change for the current APIs (SecretStores; ExternalSecrets). Expect this to be the v2 version of it ;)

Does this mean all `1.x` versions will be supported? While ESO is on 1.0 - our support policy did not change. We only support the latest minor version (1.x) currently released. Today, that will be 1.0.

How can I upgrade? The upgrade is instant from 0.20.x. There are no expected breaking changes as a part of this update .

Now to organizational things. We also changed our community meeting hours. We will have alternating times where once the meeting is at the same current time it is right now, but the other will be at UTC 21:00. Making sure that people from different continents might be also able to join our community meetings.

Thank you so much for everyone involved in this release and for all our supporters. Without you, this would not have been possible.

Gergely(Skarlso)

By using their Sync subscription feature.

Hello everyone!

My project [crd-to-sample-yaml](https://github.com/Skarlso/crd-to-sample-yaml) just got a neat new feature. It's validation of schemas between versions for a CRD. What the heck do I mean here?

TL;DR: Things like modifying required fields, removing existing fields, increasing minimum etc, all cause breaking changes between versions. That's what this can detect and show.

Longer:

So the project already had a testing feature like helm, but it wasn't this direct. And this command now can more easily be integrated as a CLI checker for between specific versions. Something like this:

``` cty validate schema -c path/to/crd.yaml --from v1alpha1 --to v1beta1 --fail-on-breaking ```

And it, of course, works with all the providers that it supports ( github, url, oci, files, folders... ).

The output is something like this:

``` Schema Validation Report =======================

CRD: TestResource From Version: v1alpha1 To Version: v1beta1

Summary: Total Changes: 5 Breaking Changes: 2 Additions: 2 Removals: 0

Changes: [breaking] spec.required: Field 'version' is now required + [addition] spec.properties.version: Property 'version' added [breaking] spec.properties.count.minimum: Minimum increased Old: 1 New: 5 ```

Or JSON, or YAML for further processing or putting the output into a Release Note.

I hope this is useful, it was very interesting to write it at least. :)

Enjoy it here [v1.2.0 release page](https://github.com/Skarlso/crd-to-sample-yaml/releases/tag/v...) and there is more documentation in [Schema validation doc](https://github.com/Skarlso/crd-to-sample-yaml/blob/main/SCHE...).

Thanks!

WOW this blew up. I didn't expect it at all... Thank you so much for everyone engaging with it. I know it's a difficult topic. And I might have used skeptics term incorrectly? :)

The school LLM thing is absolutely real. And it's not even a cheap school. It's a school in Denmark. I was very disappointed to know this and was not expecting it at all. And it's sadly, not an assignment to learn llms either. I'd wish it was something like that.

To the ones saying we've seen this before with google search, ides, etc. Not on this scale. And even then, you didn't _completely_ outsource you ability to think. I _think_ that is super dangerous. Especially for young people who get into the habit a lot faster. And suddenly, we have people unable to think without asking an LLM for advice. Other disruptive technologies didn't affect thinking on this massive scale.

I'm not saying stop AI booo, that's obviously not going to happen. And I don't want to, maybe the singularity is just an AI away. Who knows? However, I'm asking that we absolutely put that thing behind some kind of oversight especially in schools for young people before they develop critical thinking skills. After all you started to count on paper before you start using a calculator for a reason.

Again, thank you for this discussion. I'm really grateful for the engagement.

The original article didn't talk about any of these things....

Indeed it didn't. That's why I did. :)

Ah. The article's author heard a bad take on LLMs and accidentally attributed it to Ptacek's article.

Not at all. And if that's everything you took from this article, then I'm sorry for the loss of your time.

Hello. Just wanted to share my 2¢ on this post that is entirely missing the point of the skepticism I believe. Take it with a grain of salt. Thank you.

Hey folks. ESO will have an MFA token generator Generator once this[^1] pr is merged and released. I'm looking forward to see what kinds of shenanigans people will come up with to use it. :)

[^1]: https://github.com/external-secrets/external-secrets/pull/47...

Hey folks.

This has been brewing in me for a while now that I wanted to describe how CRDs are registered and actually show some code in Kubernetes to boot. Please enjoy. Maybe it's enough or not enough. Feedback is as always, welcomed. :) Thanks!

Thank you! :) Luckily, I'm now happily employed and hopefully for the rest of my life. :D

Ah envy you. I hate interviewing. I tried to put myself into positions where I present something. I actually did a LOT of presentations by now and yet every time I’m nervous. I just can’t seem to make it work with my brain. :D

I feel you. I have adhd as well. I have routines when I code which some people don’t like or can’t follow and that caused problems during interviews. When I do pair coding today with people I worked with for a while now that’s not a problem because they understand my style and they don’t mind because they are good people.

But most people don’t understand why I do the thing I do and start saying no and that’s not supposed to go there and why aren’t you writing a test first? Because I don’t want to. I will arrive at your conclusion don’t worry just please give me time. I will move this code around don’t worry about it. There you see now it’s all fine. But they don’t understand that I just want something to work fast and then I’ll refactor and make it nice.

And then they start demanding and that’s when my flow breaks and then I do shit work because they force me to accept their views.

Ofc granted those are shitty workplaces but chances are I’m not gonna work with the person interviewing me and also when you are looking for a job without a job then you kinda don’t care and want it make it work somehow.

Bah.

Thanks for the clarification! I see. I guess, I have to decide where and to what extend I would like to level my skills. And that where do I start.

Not going to lie, getting Oobabooga to work on my laptop was no small feet. Conda didn't work, python modules didn't work. cmd_macos failed with an error... But eventually I got there by hand. :D So the "Getting Started" was quite steep.

Hi, I'm a noob in this plane, but I'm learning slowly. Could you please tell me how this is different from running something like [Msty](https://msty.app/)?

That also supports running local models that you already installed using ollama and things like, adding pdf and preserving chat history. So what am I looking at here with ooba?

Thanks!

I like this. Very low tech. I distractions. No fluff. Just text. Amazing. I use obsidian but I find myself always spending time trying to tweak it. And it gets nowhere other then me spending unnecessary time on it.

Exactly. Sometime I’m surrounded completely without changes of going around something with two three level 5-6 creatures in the beginning at level 3.

Hello folks.

Thought I'd share a small script I wrote for this annoying Github registry issue to delete container type packages.

The problem I found if your package has `/` in it it wouldn't find it using the REST api. So we need to escape that using %2F. That works and the script does this for you.

Anyways, I linked the gist that contains the script. It could be further enhanced to work for Organizations or other types like docker npm, etc.

Ah okay, thanks.

Hello dear HN readers.

I would like to show you, with pictures, how I tracks todos and tasks using Obsidian and it's daily notes section.

It's far from perfect, but it works for my mind. :) I thought maybe someone could appreciate the write up and cohesion that it might bring or further ideas. :)

Cheers!

Well. Now the license change and Mitchell leaving makes sense.

Hello Peeps!

A couple of months ago I wrote about a tool I created to parse CRDs and generate a Sample YAML out of it including values with the right type.

I created a website for that which was the classing client-server website which I hosted.

Now it's a static WASM based SPA site using go-app that is hosted on GitHub pages. It's pretty neat. No more JavaScript. :D I jest. But it IS pretty neat.

The link above is a post detailing the process and linking to the site.

Thank you! :)

Hello people! :)

I wrote a GitHub Action that can automate specific processes involving dealing with issues for ProjectV2.

Amongst others, the following:

- moving issues in "columns" aka setting status - this isn't just moving issues ( for that there are other commands ). There is a feature that can be executed via a cron job called "scan". It will move through all Open Pull requests and in a given time moves all associated issues into a desired state, like `In Review`. - then, upon update of the pull request, it will move associated issues back into `In Progress`. - the scenario here is that you don't have to remember to set the state for your tickets, it can be done automatically so everyone can see what's going on at a glance of the board. I think that's neat. :) - slash commands - I'm working on adding more, for now `/assign` exists which assign the pull request to you if you comment it on the pull request AND ALSO it will assign every connected issue to the user. - automatically assigning issues to projects upon creation - creating issues from pull requests ( Pending feature )

And I'm adding more stuff here to help manage projects and issues. Caretaker comments back on pull requests and leaves reactions to comments it processed. Those can be disabled.

All of this can happen on Action Triggers so Caretaker doesn't need to be hosted anywhere and doesn't require any webhooks.

I hope someone finds this useful. :) Please feel free to leave any feature requests! I'm happy to extend this thing and make it a useful tool for developers who, like me, just plain forget to move that dang issue into review. ;D

Oh, interesting. Thanks, I'll look at that. Unfortunately, caddy was not an option I could have used here. :/ But it's still interesting and I didn't know it could do that. :)

Well, aside from the fact that it's calling out to the internet, the service is in early access. Using cert-manager at and a self-signed cert, you can set this up in a fenced environment as well.

Indeed. Kubernetes setup is not for the faint of heart. But then again, it could be a LOT more confusing than this, I think. Here, you just download something, put it somewhere, and you're basically done. Once this is in place you can basically forget about it. Which was my aim. :)

I believe dealing with certificates that include the ability to end-to-end test them is never simple.

Hello folks!

Thought I'd share some insights in how I set up test frameworks using cert-manager for generating self-signed certificates. And then making them locally trusted using mkcert for easy install.

I think it's neat, but it does have its caveats as everything I suppose. :)

Thanks!

I meant to write `sig-autoscaler`. :)

Just a small remark that dog-autoscaler which is `cluster-autoscaler` supports scaling to and from 0 pretty nicely. I even implemented that functionality in cluster api provider aws using cluster autoscaler. Which probably why the native support isn’t going along so nicely.

Thanks very much!!!

I _WISH_ this would be 2005. Did you ever work at a bank? They are still using 1.6-1.9 maybe.

I still know modern Java codebases where long descriptive class names are a must. So sadly, while I understand your sarcasm, it is not the case.