HN2new | past | comments | ask | show | jobs | submit | smusamashah's commentslogin

Can we have icons like these again please.

I started from Windows 98 and always loved the icons. They actually represented the application and purpose. These days they are more focused on looking modern. Lots of times they are not even distinguishable between each other.


C:\Windows\System32\moricons.dll is still available, you can set them on shortcuts in Properties with the "Change icon" button.

I have two twitter accounts. On one I like indie games, ai stuff, Gaussian Splats etc and some other things and that is what my feed is filled. I discover so many good games here. I keep it that way by not watching and liking randomly recommended things. My second twitter account is full of crappy videos (fights, accidents and sometimes very horrific).

I have never seen things of my interest on Facebook ever. It is full male focused staged thirst crap. I log in to that account may be once or twice a year.

I think Facebook should be steerable like Twitter. I haven't tried because it's super clingy.


So there is a spreadsheet of websites. That is very interesting. There was an article here sometime ago about a media group who have so many super SEOd websites. They all have common footer text. I searched and added as many as I could find in uBlacklist. I have a gist listing them and how I searched for them. You might find that useful.

Edit: https://gist.github.com/SMUsamaShah/6573b27441d99a0a0c792431...


With this speed, you can keep looping and generating code until it passes all tests. If you have tests.

Generate lots of solutions and mix and match. This allows a new way to look at LLMs.


Not just looping, you could do a parallel graph search of the solution-space until you hit one that works.

Infinite Monkey Theory just reached its peak

You could also parse prompts into an AST, run inference, run evals, then optimise the prompts with something like a genetic algorithm.

Agreed, this is exciting, and has me thinking about completely different orchestrator patterns. You could begin to approach the solution space much more like a traditional optimization strategy such as CMA-ES. Rather than expect the first answer to be correct, you diverge wildly before converging.

And then it's slow again to finally find a correct answer...

It won't find the correct answer. Garbage in, garbage out.

How about if you run this loop (one year from now) on this kind of hardware but with something like Claude/Kimi K2. How about that? Because that's where it'll go.

This is what people already do with “ralph” loops using the top coding models. It’s slow relative to this, but still very fast compared to hand-coding.

This doesn't work. The model outputs the most probable tokens. Running it again and asking for less probable tokens just results in the same but with more errors.

Do you not have experience with agents solving problems? They already successfully do this. They try different things until they get a solution.


don't forget JavE! http://www.jave.de/

As an aside but still relevant question, why is CSS preferred over JS when these days it can do lots of things like JavaScript and probably uses similar resources.

CSS and JavaScript specialize in two very different areas and are commonly used in tandem

You can do layouts with javascript?

Couldn't imagine ever wanting skip grid and flexbox for whatever has been concocted up for JS.


multi-window interfaces in the browser... simulating a desktop, or other user navigable environment such as in a game or simulation, where a user my want to customize their environment beyond a grid snap.

The issue then becomes how do you make it accessible to screen readers? It's not impossible, just very tedious and requires cross browser and cross-device testing

While I understand the concern... Not everything ever made needs to be fully accessible to everyone.

That said, there are already aria labels for UI contexts such as modal windows. Desktop OSes are already multi-modal.


1. This isn’t CSS. It’s a declarative JS drawing framework with CSS flavor to the syntax

2. Without actual CSS JavaScript wouldn’t be of much use for drawing much of anything unless you were just going to use canvasses and forego the DOM entirely


I agree. This seems like it would make more sense as a canvas library unless there's a use case I'm not understanding.

If I understand this, it’s all vector space. So it avoids one of the most irritating issues with working with canvas. I love not having to think about scale or resolution or aliasing.

I think there is value in making it a declarative model

I have seen Gemini one shotting Vampire survivor clone on Twitter few months ago. This is clearly made with an LLM, why is this on front page?

I did use an LLM as a coding assistant, but this wasn’t one-shot generated.

Found this video https://youtu.be/6YcK05z--n8 wish there were more and as interesting as the pictures in this article

This is another video https://www.youtube.com/watch?v=bqPfW62aqzk

Mockus was the best mayor Bogotá ever had. Cultura ciudadana, or "Citizen Culture," was the most transformative policy we ever implemented.


The photos in the article seemingly are of Brazilian traffic mimes.

You meant this perhaps? https://svylabs.github.io/smacthat/


Yeah, smacthat!

I get a 404 immediately


I got a cool name for the project: smacthat, so renamed the repo. But I couldn't delete this post or edit it. Sorry, but thanks for trying out.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: