Getting a LLM writing a python module I'm considering
Ask HN: What are you working on? (May 2025)
671–680 of 1001 posts
Re: Ask HN: What are you working on? (May 2025)
#672Reimplementing from scratch every major idea in modern deep learning, to help people transition into deep learning research. https://github.com/tanishqkumar/beyond-nanogpt
Username... does not check out! Which resources are you using?
I just carefully read the papers inventing the techniques I want to implement, and verify correctness of my implementation by checking outputs from my code match the papers and pass sanity checks (ie. diffusion produces nice images, RL improves rewards, etc).
Re: Ask HN: What are you working on? (May 2025)
#673A "Postgres lock diagnostics" tool for migrations. We've been bitten a few times, by seemingly innocent migrations that ended up acquiring locks we did not expect (yes, RTFM). So I thought: what if, when you opened the PR, you had a tool that actually executed your migrations, saw what locks they acquired (in runtime) and then reported it as a comment to your PR? This would provide you with yet another data point you…
A thing I always go back to, regardless of what language I'm working in, is the checks performed by strong migrations
https://github.com/ankane/strong_migrations?tab=readme-ov-fi...
It's pretty popular, and has even been ported to other languages
Re: Ask HN: What are you working on? (May 2025)
#674Working on a PDF driver for my PostScript interpreter. The drawing model is very similar between the two so it should be fairly straightforward. The only exception is the graphics save stack which persists between pages in PostScript but doesn't in PDF.
Re: Ask HN: What are you working on? (May 2025)
#675I recently quit my salary job after 16 years and am consulting in nuclear engineering now. I have a few passion projects that I'm working on (between the somewhat substantial consulting work that came out of the woodwork): - Nuclear Reactor Starter Kit --- an open source set of procedures, processes, templates, and maybe even some IT advice that should help newcomers start companies with nuclear quality assurance pro…
Do you think there's a path for ..say.. a competent engineer from a small country to obtain such a starter kit and jumpstart a weapons program? I'd guess all the information would already be available on the internet, but is there competitive advantage here?
No I don't think it'd help that much with a weapons program. The starter kit would be mostly focused on QA, compliance, configuration management, cost modeling, etc. and would not include any non-public information about like how to optimize a core for high-quality plutonium or U-233 production or how to do the chemistry needed to separate plutonium from other actinides and fission products. And of course it would say nothing at all about weapon design. That said, vast quantities of normal core design knowledge are widely available.
Re: Ask HN: What are you working on? (May 2025)
#676I'm working on Logchef: https://github.com/mr-karan/logchef
It's an open-source (AGPLv3), purpose-built log analytics UI specifically for ClickHouse. I've been using ClickHouse for logs and love its power, but found the existing UIs either too heavy, too focused on ingestion (which we already have covered with tools like Vector/Promtail/Fluentbit/Logstash etc), or not ClickHouse-native enough.
Logchef aims to be a lightweight, powerful log explorer that sits on top of your existing ClickHouse setup.
Key things:
- Schema-agnostic: Works with your existing ClickHouse log tables. No need to change how you store things.
- Dual Query Modes: Simple search syntax (e.g., level=error service=api) for quick looks, and full SQL for complex analysis when you need to dig deep.
- Lightweight & Focused: Single binary, easy to deploy. It doesn't do ingestion, letting you use best-of-breed tools for that.
- Team-centric: Built with multi-tenancy for teams and access controls for different log sources from the get-go.
I've got a demo running at https://demo.logchef.app if you'd like to try it out. It's still evolving, so feedback (especially from other ClickHouse users) would be super valuable!
Wrote an announcement post with a bit more detail: https://mrkaran.dev/posts/announcing-logchef/
Re: Ask HN: What are you working on? (May 2025)
#677A daily briefing that bundles content from every source you care about into a simple newsletter format.
Re: Ask HN: What are you working on? (May 2025)
#678AI Augment - Enabling SMEs to keep quality jobs local, without giving up data.
Novelty: Massively Agentic multi-modal AI Model with task-based agent synthesis and no measurable hallucinations. New consensus algorithm able to integrate empathic reasoning and emergent motives.
Re: Ask HN: What are you working on? (May 2025)
#679CodeVideo - https://github.com/codevideo
Basically, event sourcing for the IDE - build entire software lessons and course videos without actually recording anything.
You can get up and running pretty quickly at our studio https://studio.codevideo.io
Or locally with the CLI tool: https://github.com/codevideo/codevideo-cli
Re: Ask HN: What are you working on? (May 2025)
#680I'm working on https://nuenki.app . It's a browser extension that translates sentences at your knowledge level into the language you're learning, so you learn through immersion as you go about your day. I've been doing a lot of experiments evaluating LLM translation performance, and I used what I learnt (that LLMs make mistakes, but different LLMs make different mistakes, and they're better at critiquing translations…
Nice work and blog! You might consider crossposting some blog entries to LessWrong for more visibility of your research. People are very interested in new ideas on benchmarking.
That's a good idea, thanks for the suggestion. I'm thinking of writing one big blog article about everything I've found, loosely based on the talk, and I'll crosspost it there.