sprobertson

Much better than your current demo image! The default vignette is too obviously digital

or easier, open ~/.claude/projects/[project]/[session].jsonl (excluding the system prompt)

The official one doesn't work for me on mobile but this does: https://artemistracker.com/

it likes mentioning itself in commit messages, though you can just tell it not to.

Pretty much everyone saying worktrees but I lean heavily on hot-reloading of both backend and frontend (to actually see what I'm doing) so it's too annoying to deal with dependencies and ports when things are isolated. Instead I keep everything on the main branch and just make sure to keep tasks pretty separate in scope, e.g. add this API route vs. fix this layout issue, so they don't step on each other. When I consider something done I have the agent commit what it worked on and start a new one.

yes:

skip all the gratuitous affirmation and tell me when I'm wrong

Yeah the UX itself is clearly slop

You're certainly not alone, though it's not that much worse than usual over the last 5 years https://telliott.me/posts/is-github-getting-less-reliable/

Sure, but use the tool for the job IMO. GPTs are much more complex so should demonstrate a much more complex task.

This kind of thing is pretty easy to do with a much leaner model https://docs.pytorch.org/tutorials/intermediate/char_rnn_gen...

No - it costs $0 already

It seems like everyone commenting here is already part of the hivemind. So maybe someone can answer an important question that I'm not getting at all from the docs: what does this actually do?

I printed the casing just to see - it actually does fit in my levi pockets (front and back), but looks and feels a bit ridiculous to actually walk around with

software dev is a job for me that needs to generate income

sir, this is a hackernews

Who's the designer? I assume the same as agents.cloudflare.com, finally something that looks creative and not based on purple gradients

not for the standard hosted APIs using structured output or function calling, best you can get is an array

And #1 on the jobs chart

Here's the screen recording version of a long exposure (thanks for the nerd snipe) - https://gist.github.com/spro/7599415b0e47de65311557b3454771a...

my first instinct was "the slow one is struggling so it must be doing some heavy work"

Yeah good call, I missed that. I don't think there's a correct answer here, but it could be another step of the read or write. Either it would do another lookup of "my daughter" -> Name on read, or do a lookup on write if you already have a "my daughter is Name" memory. Whatever's less expensive in the long run. The graph memory someone else mentioned also seems like a good option there.

It's been asked a few times, here are some links to get you started:

* https://news.ycombinator.com/item?id=36370684

* https://news.ycombinator.com/item?id=30752540

* https://news.ycombinator.com/item?id=9896369 (Python specific)

You might want to put your actual username in those links. Would also be good to see an example PDF or at least screenshot in there.

As a simplified example:

(memory) [2023-07-01] My daughter is 10

(memory) [2024-05-30] My daughter turned 11 today!

System prompt: Today is 2025-08-21

User prompt: How old is my daughter?

The vector DB does the work of fetching the daughter-age-related memories, your system decides (perhaps with another LLM) if the question needs time-based sorting or something else.

FYI using pagination erases your filters

Would be good to have the "play" button above the fold on the detail page

Previous discussion: https://news.ycombinator.com/item?id=1912530

for anonymization it would probably be a good idea to blank out links (linkedin/github/etc)

simply tell it what command it should call to compile, or better yet add that info to Claude.md

I have used OpenSCAD not CadQuery, to me (as a vim person) the appeal is not having to do a bunch of mousework. But it doesn't hold up for long, at some level of complexity it is certainly easier to just go learn SolidWorks/OnShape/Fusion.

This in combination with making the tool call / json response itself also "batched" is a good pattern. Instead of returning a single `{english, foul}` object per record, pass in an array of records and have it return an array of `[{english, foul}]`. Adjust the inner batch size depending on your record size and spread the rest over the batched API.

Feedback:

* Font is very hard to read, needs some aliasing adjustment maybe?

* The dropdowns seem to be simple prompt injection, so sometimes just selecting "NES" as the color scheme will add an NES to the image. Similarly, the default "anime" style makes it looks like anime and not pixel art