IdeaWell

Project ideas from Hacker News discussions.

Ask.com has closed

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1. Nostalgic praisefor early natural‑language search

“Yeah I remember using it back in the day and getting good results.” — Mistletoe

2. Disappearance of advanced query features

“None of that ever worked properly, consistently, at google.” — rsync
“They started to optimize for ‘the masses’, and they don’t use boolean logic in queries.” — stingraycharles

3. Perception of lost search identity and domain value

“The whole point of AskJeeves was that you could ask Jeeves things in natural language … but it didn’t really work so you were left disappointed every time.” — ryukoposting
“I’d be willing to bet ask.com will always resolve to a pingable IP address – that’s a HOT domain name.” — arm32

4. Broader commentary on search decline and hope for alternatives

“Internet search is dead or dying. It has been enshittified to perfection.” — zombot
“I pay for kagi on my personal machine, it is always a delight when my cmd‑t search is answered kagi and not a list of ads …” — akafred

🚀 Project Ideas

JeevesQuery API - Precise Boolean & Distance Search-as-a-Service

Summary- A lightweight API that lets developers query archives and modern search indexes using natural‑language phrasing while retaining AltaVista‑style boolean operators, proximity, and negative terms.

  • Enables accurate, intent‑driven search without ad‑laden noise, fulfilling the need for “real” search functionality.

Details

Key Value
Target Audience Web developers, data scientists, researchers building knowledge‑retrieval tools
Core Feature API endpoint that interprets natural‑language queries, validates boolean syntax, applies proximity/distance weighting, returns ranked results with source citations
Tech Stack Python (FastAPI), Elasticsearch, OpenAI embedding model, PostgreSQL for metadata
Difficulty Medium
Monetization Revenue-ready: Pay‑per‑call $0.001 per query, free tier of 10,000 queries/month

Notes

  • Quote from HN: “None of that ever worked properly, consistently, at google.” and “AltaVista was fantastic and represented a features and usability high water mark that was never passed by google.”
  • Potential for discussion: solves frustration with modern ad‑filled search and restores precise query control.

ClusterSearch Chrome Extension - Reviving Altavista Clustering

Summary

  • A browser extension that reconstructs Altavista’s cluster‑based result grouping, letting users click a cluster to exclude irrelevant results and refine queries.
  • Provides a visual, intuitive way to manage search complexity and reduce noise.

Details

Key Value
Target Audience Power searchers, researchers, students who need structured result navigation
Core Feature Real‑time clustering of search results with expandable/collapsible clusters and one‑click exclusion
Tech Stack React, Rust (wasm), Browser Extensions API, Elasticsearch client
Difficulty Medium
Monetization Revenue-ready: Freemium with premium clustering packs $4/mo

Notes- HN users lamented the loss of “boolean logic” and clustering: “Altavista had a Java applet that would visualize the ‘clusters’ that a search produced… This is a feature still unmatched by any search engine today.”

  • Offers practical utility for users missing that UI and could spark community contributions.

AskArchive - LLM‑Powered Historical Search over the Wayback Machine

Summary

  • A web service that answers natural‑language questions about archived web pages by combining LLM reasoning with indexed Wayback Machine snapshots.
  • Turns the dusty archive into an interactive knowledge source without manual browsing.

Details

Key Value
Target Audience Historians, journalists, nostalgic users, content curators
Core Feature Query archive with natural language, retrieve relevant snapshots, let LLM synthesize answers, filter by date range
Tech Stack Node.js, LangChain, Cloudflare R2 (Wayback data), Pinecone vector store
Difficulty High
Monetization Revenue-ready: $15/mo subscription for heavy users, $5/mo for basic

Notes- Quote: “Altavista had more relevant search results than Google has now.” and “None of the search engines from that era were really good.”

  • Appeals to users nostalgic for early search and to researchers needing contextual answers from old sites.

MetaMeta Unified Multi‑Engine Search Dashboard

Summary

  • A web app that simultaneously queries Google Custom Search, Bing, DuckDuckGo, Ask, and Yahoo APIs, then deduplicates and ranks results, with query saving and export.
  • Gives casual users the “best of all worlds” without juggling multiple sites.

Details

Key Value
Target Audience Everyday internet users, students, hobbyists who want comprehensive yet simple search
Core Feature Unified query box, multi‑engine result merging, result filtering, history/export to CSV
Tech Stack Next.js, GraphQL, Serverless functions (Vercel), API keys for search engines
Difficulty Low
Monetization Revenue-ready: $5/mo per user

Notes

  • Users expressed frustration: “I would think that 90% of the principals at DEC/Compaq WRL working on AltaVista would have moved to google” and “None of the search engines from that era were really good… they all sucked again.”
  • Address

Read Later