Hacker News: xlth

Show HN: Trackless Links – clean URLs, no tracking (Safari)

xlth — Wed, 11 Mar 2026 14:03:38 +0000

Built this because I was tired of sharing URLs full of tracking cruft. Safari extension for iOS and macOS that strips utm_*, fbclid, gclid, and 50+ other tracking parameters automatically. Runs entirely on-device with zero data collection.

Also bundles CRED-1, a domain credibility dataset (2,672 flagged domains with 5 independent trust signals) that can warn you before visiting sources known for misinformation.

Available on the App Store for iPhone, iPad, and Mac.

Comments URL: https://news.ycombinator.com/item?id=47335715

Points: 2

# Comments: 0

Show HN: PowerSkills – PowerShell toolkit for AI agents on Windows

xlth — Sun, 08 Mar 2026 15:00:20 +0000

Article URL: https://github.com/aloth/PowerSkills

Comments URL: https://news.ycombinator.com/item?id=47297866

Points: 4

# Comments: 0

New comment by xlth in "LLMs work best when the user defines their acceptance criteria first"

xlth — Sat, 07 Mar 2026 08:46:11 +0000

You're not off-base at all. The way I think about it:

- LLM = the model itself (stateless, no tools, just text in/text out) - LLM + system prompt + conversation history = chatbot (what most people interact with via ChatGPT, Claude, etc.) - LLM + tools + memory + orchestration = agent (can take actions, persist state, use APIs)

When someone says "LLMs have no memory" they're correct about the raw model, but Claude Code or Cursor are agents - they have context, tool access, and can maintain state across interactions.

The industry seems to be settling on "agentic system" or just "agent" for that last category, and "chatbot" or "assistant" for the middle one. The confusion comes from product names (ChatGPT, Claude) blurring these boundaries - people say "LLM" when they mean the whole stack.