dad creativity in: genomics, bioinformatics and solving problems languages: rust, nim, python, go, c, bash, awk, jshttps://bsky.app/profile/brent-p.bsky.socialhttps://bsky.app/profile/brent-p.bsky.social/post/3muiiii45tk2ga new release of somalier with a user fix for long reads: https://github.com/brentp/somalier/releases/tag/v0.3.4 somalier continues to get better and if you're not using it or another means to detect sample swaps, you're operating blind reach out to me for consulting if you want to do some in *depth* QC on large cohorts01 Sep 2026 22:40 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3muiiii45tk2ghttps://bsky.app/profile/brent-p.bsky.social/post/3mt32xitkic2m"load-bearing" is coming up a lot lately. Anyone seen a sort of LLM etymology of how these phrases come to be so commonly-used? Doesn't seem obvious how this happens with RL as I understand it. And to obviate the obvious response: "you're right to call that out!"14 Aug 2026 21:09 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mt32xitkic2mhttps://bsky.app/profile/brent-p.bsky.social/post/3mlu4ahigvc2vhttps://thehackernews.com/2026/05/mini-shai-hulud-worm-compromises.html It's hard to find all these in one place: ~/.npmrc min-release-age=7 ~/.config/uv/uv.toml exclude-newer = "7 days" ~/.bunfig.toml [install] minimumReleaseAge = 604800 ~/.config/pnpm/config.yaml minimumReleaseAge: 1008015 May 2026 00:46 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mlu4ahigvc2vhttps://bsky.app/profile/brent-p.bsky.social/post/3ml7m5thy6c2zLLMs open up the possibility of providing more functionality and customization for command-line tools and anything that takes a config without drowning users by exposing a scripting interface. That's basically it, but a bit more here: https://brentp.github.io/latest/blog/2026/04/28/ai-scripting/06 May 2026 21:05 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3ml7m5thy6c2zhttps://bsky.app/profile/brent-p.bsky.social/post/3mkq7oerke22vthanks to @ankane.bsky.social echtvar is now available in brew: brew install echtvar dat is superleuk! [contains quote post or other embedded content]30 Apr 2026 18:12 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mkq7oerke22vhttps://bsky.app/profile/brent-p.bsky.social/post/3mkpuxp6dws2pWith help from contributors, echtvar is still improving: github.com/brentp/echtvar https://github.com/brentp/echtvar30 Apr 2026 15:00 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mkpuxp6dws2phttps://bsky.app/profile/brent-p.bsky.social/post/3mk65ixgbwc2thttps://github.com/brentp/blammo-qc Does QC and EDA for long reads. It was written entirely by LLM.23 Apr 2026 13:45 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mk65ixgbwc2thttps://bsky.app/profile/brent-p.bsky.social/post/3mjpdlpktd22gNew version of somalier is up: https://github.com/brentp/somalier/releases/tag/v0.3.2 Has a CHARR contamination estimator, a CONPAIR pair-wise inter-sample contaminatoin estimator, an AB/depth plotter and other improvements to make it of greater utility in cancer samples. [contains quote post or other embedded content]17 Apr 2026 16:24 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mjpdlpktd22ghttps://bsky.app/profile/brent-p.bsky.social/post/3mjklq6c4v22fas part of my work with isabl genomics, I've been updating somalier with more info for cancer samples (contamination and concordance). somalier extracts small "sketches" of ~20K sites from a BAM/CRAM/VCF and then does rapid all-vs-all kinship checks. (1/n)15 Apr 2026 19:07 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mjklq6c4v22fhttps://bsky.app/profile/brent-p.bsky.social/post/3mgvliom4q224the codex desktop app was only available as a Mac DMG. A while ago, I downloaded it and told codex cli to make it work on linux. It did. Now I use that nearly exclusively. https://gist.github.com/brentp/b6bd8d009f8936bc0b0c12f29473e0ef12 Mar 2026 23:46 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mgvliom4q224https://bsky.app/profile/brent-p.bsky.social/post/3mgvamw7py22qAgain, I think making a custom GPT (or similar) as I did for bedder will be the default for most software shortly. If it can't answer common questions, improve the docs and refresh and the maintenance burden of answering user questions goes down. [contains quote post or other embedded content]12 Mar 2026 20:32 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mgvamw7py22qhttps://bsky.app/profile/brent-p.bsky.social/post/3mgdxttx4fc2hI would guess that soon, every tool will have one of these. It is more interactive than just documentation and better than waiting for a grumpy maintainer to answer the same question in a github issue or personal email. [contains quote post or other embedded content]05 Mar 2026 23:39 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mgdxttx4fc2hhttps://bsky.app/profile/brent-p.bsky.social/post/3mgdrlmapzc2jFeeding my markdown docs to this, then using it in anger shows where the docs need to be bolstered and provides a nice reference for users: https://help.openai.com/en/articles/8554397-creating-a-gpt here's one for bedder: https://chatgpt.com/g/g-69a6027ff4e4819199e7d27f3c6a83e5-bedder05 Mar 2026 21:47 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mgdrlmapzc2jhttps://bsky.app/profile/brent-p.bsky.social/post/3mfhwzbzu6s2xI had a similar, less principled way of doing this, but this does seem to work quite well after trying a couple of large changes with it: https://boristane.com/blog/how-i-use-claude-code/ (I use codex, but same idea)22 Feb 2026 20:10 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mfhwzbzu6s2xhttps://bsky.app/profile/brent-p.bsky.social/post/3mdmfhbumpc2jAGENTS.md > skills https://vercel.com/blog/agents-md-outperforms-skills-in-our-agent-evals30 Jan 2026 03:48 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mdmfhbumpc2jhttps://bsky.app/profile/brent-p.bsky.social/post/3md4cnfgpts2yscratch that. better to use gemini-cli, turn on preview features and then get gemini 3.0 pro and flash for free for a decent amount each day. [contains quote post or other embedded content]23 Jan 2026 18:16 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3md4cnfgpts2yhttps://bsky.app/profile/brent-p.bsky.social/post/3mcdsffmxjk2acrush coding agent with a free plan from cerebras using glm4.7 model is an easy way to start with coding cli tools. https://inference-docs.cerebras.ai/support/rate-limits https://github.com/charmbracelet/crush crush is static binary that detects your CEREBRAS_API_KEY. GLM 4.7 isn't as good as Opus, but it is very solid.14 Jan 2026 00:21 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mcdsffmxjk2ahttps://bsky.app/profile/brent-p.bsky.social/post/3mcalq65z5s2wanother reminder about htsvcf which allows either 1. rust library to support javascript expressions (filtering/modification) on variants from VCF/BCF 2. javascript (node/bun) library for reading/writing/modifying variants from vcf/bcf. both very performant. github.com/brentp/htsvcf https://github.com/brentp/htsvcf12 Jan 2026 17:44 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mcalq65z5s2whttps://bsky.app/profile/brent-p.bsky.social/post/3mb7s7uumtk2da post about how, with LLMs, I wrote a new set of libraries. Wrapping htslib is my hello-world but doing it in rust to allow javascript expressions handled by V8 was more substantial project. The tools and LLMs are now getting quite good: https://brentp.github.io/latest/blog/2025/12/30/ai-coding/ let me know what you think30 Dec 2025 16:42 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mb7s7uumtk2dhttps://bsky.app/profile/brent-p.bsky.social/post/3mb36vzflx22bmy LLM-coded version of the single-page year calendar that's been going around. This one can be scripted with javascript to highlight cells, change the borders and texts. It also allows saving and sharing via url: https://brentp.github.io/single-page-js-calendar/calendar.html with code here: https://github.com/brentp/single-page-js-calendar?tab=readme-ov-file#single-page-calendar28 Dec 2025 20:46 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mb36vzflx22bhttps://bsky.app/profile/brent-p.bsky.social/post/3mazn2j3sps2uso Opus 4.5 is excellent for coding. I've been using these LLMs for coding for a while and it felt like slow improvement with each new model and sometimes more work to guide the model. Opus 4.5 just works. It solves hard problems and has good taste. Anyone else notice this?28 Dec 2025 05:54 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3mazn2j3sps2uhttps://bsky.app/profile/brent-p.bsky.social/post/3malkpslkyc2lI'll write more about this later, but as an experiment in LLM coding (GPT-5.2 and Opus 4.5), I wrote a JS wrapper for rust-htslib to read/write modify VCFs: www.npmjs.com/package/htsvcf works on linux (x64/aarch64) and OSX. https://brentp.github.io/htsvcf/latest/htsvcf_napi/index.html22 Dec 2025 15:35 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3malkpslkyc2lhttps://bsky.app/profile/brent-p.bsky.social/post/3m5wbftgxxc2jI vaguely recall seeing #bioinformatics tool that could compare variants (or set of variants) using haplotypes rather than simply normalize and check for equality. That murky description is about as specific as I can get. Any pointers? IIRC, it's not just a truth-set evaluation tool. Has API18 Nov 2025 17:32 +0000at://did:plc:b4x3jqelv3boot7op2yo7kkw/app.bsky.feed.post/3m5wbftgxxc2j