Hacker News: postalcoder

New comment by postalcoder in "A Claude Code skill that makes Claude talk like a caveman, cutting token use"

postalcoder — Sun, 05 Apr 2026 14:58:26 +0000

I disagree with this method and would discourage others from using it too, especially if accuracy, faster responses, and saving money are your priorities.

This only makes sense if you assume that you are the consumer of the response. When compacting, harnesses typically save a copy of the text exchange but strip out the tool calls in between. Because the agent relies on this text history to understand its own past actions, a log full of caveman-style responses leaves it with zero context about the changes it made, and the decisions behind them.

To recover that lost context, the agent will have to execute unnecessary research loops just to resume its task.

New comment by postalcoder in "Caveman: Why use many token when few token do trick"

postalcoder — Sun, 05 Apr 2026 14:31:03 +0000

Pruning an assistant's response like that would break prompt caching.

Prompt caching is probably the single most important thing that people building harnesses think about and yet it's mind share in end users is virtually zero. If you had to think of all the weirdest, most seemingly baffling design decisions in an AI product, the answer to "why" is probably "to not break prompt caching".

New comment by postalcoder in "Samsung Magician disk utility takes 18 steps and two reboots to uninstall"

postalcoder — Fri, 03 Apr 2026 16:13:19 +0000

> I worked on a disk utility in the 90s called PartitionMagic

PartitionMagic was a S-Tier windows utility. Thank you.

New comment by postalcoder in "Show HN: I built a frontpage for personal blogs"

postalcoder — Fri, 03 Apr 2026 13:14:02 +0000

If anyone looking for something even more minimalist, give the HN x Small Web RSS feed a try

https://hcker.news/feeds/atom?period=day&limit=50&smallweb=t...

New comment by postalcoder in "Axios compromised on NPM – Malicious versions drop remote access trojan"

postalcoder — Tue, 31 Mar 2026 04:05:59 +0000

PSA: npm/bun/pnpm/uv now all support setting a minimum release age for packages.

I also have `ignore-scripts=true` in my ~/.npmrc. Based on the analysis, that alone would have mitigated the vulnerability. bun and pnpm do not execute lifecycle scripts by default.

Here's how to set global configs to set min release age to 7 days:

  ~/.config/uv/uv.toml
  exclude-newer = "7 days"

  ~/.npmrc
  min-release-age=7 # days
  ignore-scripts=true
  
  ~/Library/Preferences/pnpm/rc
  minimum-release-age=10080 # minutes
  
  ~/.bunfig.toml
  [install]
  minimumReleaseAge = 604800 # seconds

(Side note, it's wild that npm, bun, and pnpm have all decided to use different time units for this configuration.)

If you're developing with LLM agents, you should also update your AGENTS.md/CLAUDE.md file with some guidance on how to handle failures stemming from this config as they will cause the agent to unproductively spin its wheels.

New comment by postalcoder in "Tell HN: Litellm 1.82.7 and 1.82.8 on PyPI are compromised"

postalcoder — Wed, 25 Mar 2026 09:43:57 +0000

Yes there is. You can use those configs as flags in the CLI to override the global config.

eg:

  npm install  --min-release-age 0
  
  pnpm add  --minimum-release-age 0
  
  uv add  --exclude-newer "0 days"
  
  bun add  --minimum-release-age 0

New comment by postalcoder in "Tell HN: Litellm 1.82.7 and 1.82.8 on PyPI are compromised"

postalcoder — Wed, 25 Mar 2026 07:24:26 +0000

Works for me?

  $ pnpm add -D typescript@6.0.2
   ERR_PNPM_NO_MATURE_MATCHING_VERSION  No matching version found for typescript@6.0.2 published by Wed Mar 18 2026..

You could also set the config this way:

  pnpm config set minimumReleaseAge 10080 --global

You may be thinking about the project-specific config, which uses YAML.

https://pnpm.io/cli/config

New comment by postalcoder in "Tell HN: Litellm 1.82.7 and 1.82.8 on PyPI are compromised"

postalcoder — Wed, 25 Mar 2026 06:18:30 +0000

FYI, npm/bun/pnpm/uv now all support setting a minimum release age for packages.

I updated my global configs to set min release age to 7 days:

  ~/.config/uv/uv.toml
  exclude-newer = "7 days"
  
  ~/.npmrc
  min-release-age=7 # days
  
  ~/Library/Preferences/pnpm/rc
  minimum-release-age=10080 # minutes
  
  ~/.bunfig.toml
  [install]
  minimumReleaseAge = 604800 # seconds

New comment by postalcoder in "Tell HN: Litellm 1.82.7 and 1.82.8 on PyPI are compromised"

postalcoder — Tue, 24 Mar 2026 13:29:27 +0000

This is a brutal one. A ton of people use litellm as their gateway.

New comment by postalcoder in "Ripgrep is faster than grep, ag, git grep, ucg, pt, sift (2016)"

postalcoder — Tue, 24 Mar 2026 11:22:55 +0000

idk if this was your issue but I’m posting this because it’s not obvious (especially the default behavior):

  rg      : Searches git tracked files
  rg -u   : Includes .gitignored files
  rg -uu  : Includes .gitignored + hidden files
  rg -uuu : Includes .gitignored + hidden + binary files

New comment by postalcoder in "I'm OK being left behind, thanks"

postalcoder — Fri, 20 Mar 2026 14:32:22 +0000

I agree with you, which is why I think it's a straw man. How many real devs are actually banging the "you're getting left behind!" drums?

New comment by postalcoder in "I'm OK being left behind, thanks"

postalcoder — Fri, 20 Mar 2026 14:28:25 +0000

The thing is, this post is hitting a straw man. ngmi culture was deeply toxic and pervasive in crypto. I think the people who are really into LLMs are having a blast.

New comment by postalcoder in "Afroman found not liable in defamation case"

postalcoder — Thu, 19 Mar 2026 13:06:53 +0000

It gives me immeasurable delight seeing afroman at the top of HN.

Love me some freedom, sweet soulful music, and pie in the face of bad cops.

Dang/Tom, please don't downrank this. America needs this win.

New comment by postalcoder in "$96 3D-printed rocket that recalculates its mid-air trajectory using a $5 sensor"

postalcoder — Wed, 18 Mar 2026 14:16:45 +0000

For the non-Americans, the modern technology you're referring to is the HOA.

New comment by postalcoder in "Kagi Small Web"

postalcoder — Tue, 17 Mar 2026 12:45:33 +0000

Kagi used to use Google's search API. They discuss the current state of their index here: https://blog.kagi.com/waiting-dawn-search

New comment by postalcoder in "Kagi Small Web"

postalcoder — Tue, 17 Mar 2026 10:52:02 +0000

And Small HN:

https://hcker.news/?smallweb=true

https://news.ycombinator.com/item?id=46618714 (Ask HN: Share your personal website, 2414 comments)

New comment by postalcoder in "The “small web” is bigger than you might think"

postalcoder — Tue, 17 Mar 2026 03:48:47 +0000

Multiple layers of curation works really well. Specifically, using HN as a curation layer for kagi's small web list. I implemented this on https://hcker.news. People who have small web blogs should post them on HN, a lot of people follow that list!

New comment by postalcoder in "I don't use LLMs for programming"

postalcoder — Thu, 12 Mar 2026 11:07:29 +0000

Fair.

New comment by postalcoder in "I don't use LLMs for programming"

postalcoder — Thu, 12 Mar 2026 10:44:37 +0000

Edited my comment before your response. But yeah, lighten up, it’s a joke! I’m not that lazy.

The only thing I do that I’d consider remotely lazy is put my API keys in my AGENTS.md so I don’t have to keep pasting it in my chat.

New comment by postalcoder in "I don't use LLMs for programming"

postalcoder — Thu, 12 Mar 2026 10:28:17 +0000

Me:

  codex
  “run my dev server”

My laziness knows no bounds.