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

Ask HN: What are you working on? (September 2026)

news.ycombinator.com

Post reply on HN
Print view View original

991–1000 of 1001 posts

Re: Ask HN: What are you working on? (September 2026)

#991

Post by adrianwaj »

dr_dshiv wrote: post #232

I’m working on translating the Renaissance — it was mostly written in Latin and 90% of it has never been translated to English. Even big names, like Ficino, Durer, Kircher, Fludd and Drebbel. And that’s Latin— things much more sparse in other classical languages. https://SourceLibrary.org is a hub that now provides over 20,000 translations of classical texts in dozens of languages (Latin and Greek but also Chinese, S…

It's a great endeavor. I put it at the bottom of my new search hub https://search.future-secured.com

I would just make sure all the titles can be clearly seen when superimposed over the pictures. Looks like a pressing matter to fix.

I actually liked the original https://archive.sacred-texts.com/ site because it's so fast and lightweight and classified by religions. I would also look at classification by languages. Is that being done at present?

I started a conversation about interfaces here if interested. https://share.gemini.google/ciOqbMFyCpBv

Re: Ask HN: What are you working on? (September 2026)

#992

Post by bredren »

I have been building Contextify.

Contextify (https://contextify.sh)

It is an application for Linux, macOS and Windows that files your AI chats into a database your AI can read using a skill or MCP.

It backs up the valuable content of your Claude Code and Codex conversations (JSONL transcripts) and turns them into a complete history you don’t have to manage.

Use Contextify during normal flow of CLI AI conversations:

   1. Resume unfinished work → "where did we leave off on that?"
   2. Recover the intent/scope → "what was the actual goal of this whole effort?"
   3. Verify it got done → "did we ever finish that, and which session proves it?"
   4. Recall a fix → "how did we fix this the last time it broke?"
Just add "use total recall" or invoke the skill directly via /total-recall or $total-recall.

You can self-host and keep all data local or use the source available Contextify Cloud sync server so your history on all machines in your cluster or desktop and laptop are always backed up and immediately available for use.

I have not been a Windows user since the days of the Hackintosh. But urging here and from users to cover the platform pushed me into it and I ended up really enjoying the challenge.

Right now I’m working on an expanded version of the windows app and support for integrating a 3rd party local LLM.

Re: Ask HN: What are you working on? (September 2026)

#996

Post by patlasswell »

I'm working on a programming language for agents to write and humans to read. Think Lean4 but built for software engineering. It's dependently typed based on Observational Type Theory. A few key differences from Lean4:

1. Constructivist rather than propositional logic, so propositions can be true or false or unknown.

2. Compiler which targets the Linux ABI (in progress). An interpreter exists and acts as an oracle for tests.

3. Built-in row polymorphic effect model.

4. Fine-grained capability and resource-based security.

I aspire for it to have a comprehensive first-party software catalog.

https://github.com/swe-toolkit/ken

Re: Ask HN: What are you working on? (September 2026)

#997

Post by mfog »

triwats wrote: post #527

Still plugging away on https://flopper.io Added models and pricing there recently too. Aiming to be a huge place for all GPU data when it comes to mathematical performance. Things like FP8 for AI inference etc. Tempted to make the infrastructure a bit more sound. Aside from that I'm working on https://solarable.org - which is aiming to provider understanding of solar panels before finding an installer and plotholes.a…

Would be cool if you could add coverage, for each GPU, of both rental price as well as purchase price (i.e. buy price := one of the spectrum of completed transactions on ebay, liquidator price, second hand price, new price, MSRP).

Always found the relationship between rental price and purchase price a bit peculiar.

Re: Ask HN: What are you working on? (September 2026)

#998

Post by __alexander »

I have been exploring reverse engineering using LLMs and Ghidra. A particular interest is in trying to understand how complexity can affect the process. I’m still learning but it’s been a fun way to explore LLMs. Here are my results using opencode https://alexander-hanel.github.io/StressingLLMs/opencode.htm...

Re: Ask HN: What are you working on? (September 2026)

#999

Post by adl »

gschier wrote: post #801

Yaak [1] is a local-first desktop API client (Postman alternative) and I've been building MCP workflows around growth lately. I already have one for product that takes in a bunch of sources (feedback, PRs, my own tasks, etc) and makes it trivial for an agent to guide me through the day's tasks. Now I'm doing the same thing for growth by taking in web/social triggers, web analytics, and other signals to spin up market…

I love Yaak! my favorite HTTP client!

Re: Ask HN: What are you working on? (September 2026)

#1000

Post by mfog »

torsi0n wrote: post #659

https://github.com/t0rsion/leone Working on my local LLM runtime Leone. It has a CPU path that verifies the CUDA kernels.

why not just compare against a Pytorch reference instead of comparing the kernels to a separate scalar Rust implementation?

(is it for perf reasons because bottleneck becomes verification? Though in that case perhaps it's better to just run a Pytorch reference on the GPUs, which should be good enough. Though, perhaps if you could get this current design to the point where `overlap(CPU, GPU cuda) < GPU cuda + GPU reference` that would be pretty awesome.

Post reply on HN