Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: What are you working on? (May 2025)

news.ycombinator.com

Post reply on HN
Print view View original

191–200 of 1001 posts

Re: Ask HN: What are you working on? (May 2025)

#191

Post by BSTRhino »

https://easel.games

A game engine that lets you code multiplayer games without coding the multiplayer! My idea was to put multiplayer into the fabric of the programming language itself. This allows the engine to automatically turn your game into a multiplayer game, without you needing to learn anything about networking or synchronization. I imagine there are lots of people who have the talent and creativity to create a multiplayer game but don't have the interest or patience in learning how to code multiplayer, and so that's who this is for!

I've been working on this for 3 years and there were lots of tricky parts rolling back and deterministically executing a whole programming language, but it's working now! My next phase is to increase the breadth of features so better games can be made with it!

Re: Ask HN: What are you working on? (May 2025)

#193

Post by static_void »

A card game in Ruby on Rails, with an emphasis on deep reaction trees, and where the resolution order of action trees depends on whether actions resolve before or after their triggers.

Its real purpose is twofold: I enjoy data modeling, and doing just enough Rails work to regain fluency after a gap.

https://github.com/alexnyeoverride/causality-rails/

Re: Ask HN: What are you working on? (May 2025)

#195

Post by recsv-heredoc »

jsemrau wrote: post #160

I still haven't found what an actual working product in AI Agents could be and write about my journey into capabilities, frameworks, and restrictions here: https://jdsemrau.substack.com/ Initially I thought there is a use-case in finance, but the barriers of entry are incredibly small and the value add is not that large. Currently, there seems to be a lot of traction in code generation (Cursor, Lovable, et al), but I…

From our observations on why - you need to have an extremely tight validation loop on everything you do for AI agents to be useful. They also need a ton of highly specific instructions and context. This requires a deep understanding of the platforms and tooling or a highly standard way of working (coding).

This is why tools like cursor work so great, they’re able to work in a super tight feedback loop with the compiler, linter and tests. They operate in a super well-known, documented environment.

If we can replicate the same thing on business systems… that’s when the magic happens - just very hard to do without deep knowledge of those platforms and agentic AI because everyone does stuff differently in each org. The overlap of people with skills in both AI and specific business ops areas is absolutely tiny.

An example of where we’re using this is in a fully AI native CRM (part of SynthGrid - see https://mindfront.ai). We don’t even have any way to interface with it outside of AI, but we’d also never want to do so again anyway because the efficiency gains are so huge for us.

The Pareto frontier will continue to inexorably advance forward, dragging even the complex or non-standardized domains in with it. For those tightly integrated business systems, we’ll probably see huge gain in utility, if not function, from the improved underlying models combined with the excellent tools. Be sure to try out Claude 4 Opus hooked into some systems if you haven’t already!

Re: Ask HN: What are you working on? (May 2025)

#197

Post by Alex-Programs »

kegs_ wrote: post #189

2 hours in and this thread is already stacked, but I'll bite since I am stuck on this problem and need help. I am working on a language learning solution that involves llms. The way I am branding it is "Anki meets Ai" because it combines a flashcard-esque method of generating complete exercises such as multiple choice, cloze, etc. with the tried-and-true SRS methodology. I think it works great! The problem is, I thin…

Hey, I happen to have run into a similar issue with my project!

I've documented a lot of my research into LLM translation at https://nuenki.app/blog, and I made an open source hybrid translator that beats any individual LLM at https://nuenki.app/translator

It uses the fact that

- LLMs are better at critiquing translations than producing them (even when thinking, which doesn't actually help!)

- When they make mistakes, the mistakes tend to be different to each other.

So it translates with the top 4-5 models based on my research, then has another model critique, compare, and combine.

It's more expensive than any one model, but it isn't super expensive. The main issue is that it's quite slow. Anyway, hopefully it's useful, and hopefully the data is useful too. Feel free to email/reply if you have any questions/ideas for tests etc.

Re: Ask HN: What are you working on? (May 2025)

#199

Post by samirsd »

working on an app that lets you take multitrack “voice memos” by plugging your phone into an interface. then the audio is automatically synced to the cloud, akin to a primitive dropbox for audio. there’s a simple mixer to adjust levels for local playback. for now i use it to get hi fi recordings of band practice and shows.

https://carnyx.ai

Re: Ask HN: What are you working on? (May 2025)

#200

Post by aag »

I got tired of using Markdown and Org mode for writing web pages last year. They're so limited, and so full of odd gotchas and limitations. Instead, I started writing raw HTML, but with a post-processing step to add titles, headers, footers, and CSS, and to do macro-like things, e.g. insert pull quotes and YouTube viewers. But raw HTML is not great, either. I'm now working on an editor that lets me use Emacs-style commands and key bindings (e.g. character, paragraph, sentence, and word motion, deletion, and transposition; Emacs-style undo/redo; incremental search; and case conversions) to edit HTML in a WYSIWYG view. The new editor does it all in a Webkit-based HTML view built with Tauri. Editing this way is so much more pleasant and more powerful. I plan to publish it under an MIT license once it's good enough.
Post reply on HN