HN user
https://github.com/y1zhou
Some tasks could take a while, especially with /goal mode. At work we often have multiple agents running on different tasks, and each could spawn subagents that you may or may not want to track. With the herdr skill you can have subagents be managed via tabs or panes, which gives them more visibility and could also explode your terminal :)
GitHub iOS logged me out. Tried logging in twice before realizing the problem is on their side...
On the flipside, this "fingerprint" approach is also what makes Shazam work poorly if you just sing into it. You're likely to generate different hashes than the original song, even if you are a very good singer! This is why newer, machine-learning-based systems are built to handle humming and singing, by matching on melody rather than exact frequencies.
So this is why singing/whistling a song to my phone never worked! I've always imagined the tech as some sort of wave pattern matching but the DFT is obviously more efficient for many scenarios. Cool article!
I'm in the same boat and https://github.com/neurosnap/zmx has been working very well for me.
Been trying out [SiYuan](https://github.com/siyuan-note/siyuan) as a local alternative and love it so far. The files are single-line JSONs so not as ideal as Obsidian .md files, but it seems to be trivial to export to various human-friendly formats.
I use the two together with no issues. Micromamba works great for conda dependencies that’s not on PyPi (for many reasons), and everything else I go for uv.
FWIW since cherry’s switch patents expired there has been a ton of great keyboards from China. I got a PMO Wave 75 and loved it.
Mine is same as the green one in this post: https://www.reddit.com/r/MechanicalKeyboards/comments/1h8011...
Has there been recent developments in the style detection and deanonymization tools you mentioned? I would assume many would not work well given the high usage of LLMs nowadays.
A few years back a journal editor maticulously reviewed all dashes in our manuscript and pointed out places where em dashes should have been used. Since then I started noticing different dashes everywhere around the internet.
Adding dependencies to the script directly was a game-changer. I was able to write a script for a friend with no coding background at all and everything ran smoothly on his machine. No more rabbit holes of bundling Python packages and setting up environments!
Support was added for this a few versions back and made life much easier!
Does uv handle cases where dependencies go beyond Python? For example, many bioinformatics-related conda packages would come with external binaries written in C++ or other languages. Would uv be able to pull from conda?
Cool work! Curious if this can be applied back in LLMs as a discrete diffusion model with partial masking.
This is so neat! MOO has been a integral part of my work yet it has never occurred to be that Pareto optimization could be applied in kart picking.
For prediction intervals that are guaranteed coverage, check out conformal prediction [1]. Works great especially for time series data.
On a related note, if you come from the R community and is already familiar with the `tidyverse` suite, `tidypolars` would be a great addition.
Closest paper I could find is https://doi.org/10.1126/science.abn7850, but it made no mentions to amino acids in the main text or the supplement.
MathJax supports server-side rendering.
https://docs.mathjax.org/en/latest/upgrading/whats-new-3.0.h...
Could this be used as a replacement for MathJax and KaTeX? If so, would the performance be comparable?
I spent more time working on the Hugo theme (https://github.com/y1zhou/hugo-northeast) than actually putting content out. Gonna spend more time writing!
I've recently setup AdGuard Home on my Raspberry Pi, and noted that requests made from an Android phone weren't going through the DNS server at all. Turns out it uses a default IPv6 DNS first (which I found no method to modify) and then falls back to the manually entered IPv4 DNS server.