HNHacker News
TopNewBestAskShowJobs

abhishake85

3 karma · joined September 23, 2016

submissionscomments
abhishake85··on Show HN: Requirements Engineering with Formal Verification
very cool!
abhishake85··on ChatGPT Now Uses Google Index – Proved Using Orphan Page Experiment
Proof that ChatGPT Plus is secretly Google-powered - “orphan page” experiment

1. I coined a nonsense word. 2. I then put it on a page not linked anywhere. 3. Forced Google to index it via Search Console; left all other engines unaware. 4. Asked ChatGPT Plus to define the term → it quoted my hidden page verbatim. 5. Ran the same query in Bing, DDG, Yandex → no results

abhishake85··on I analyzed ChatGPT with Chrome devtools to uncover its web search query patterns
I analyzed 50 real ChatGPT conversations by intercepting network traffic to uncover the patterns behind when and how ChatGPT searches the web.

Takeaways:

- ChatGPT has an internal understanding of when to lookup information using web search. It will lookup information from the internet only when unsure, or the user has explicitly asked it to lookup.

- Adding keywords like “lookup X” or “latest Y” or “Z near me” will force a web search lookup.

- Ranking boosters like “best”, “2025“, <current year>, “reviews” appear in ~30 % of ChatGPT’s generated web search queries.

- For very well established verticals (like “wordpress hosting”) ChatGPT provides an answer off the cuff from what it remembers.

- When generating web search queries, ChatGPT likes to:

    . add keywords “top”, “reviews”,

    . the user’s location and the current year

    . and authoritative sources (like “who“, “cdc” for medical queries)
Details here: https://acme.bot/blog/how-chatgpt-decides-when-to-search-the...
abhishake85··on Show HN: A Neural Networks Based Amazon Review Summarizer
Not exactly sure what you mean by "detecting bias" - but we use statistical modelling on the reviewer history to infer whether a product contains too many incentivised/spammy/fishy reviews. If we do detect any fishy behavior, we warn the user about it.
abhishake85··on Show HN: A Neural Networks Based Amazon Review Summarizer
Reading product reviews on websites like Amazon is tedious. We felt that there has to be a better way to grasp large volume of reviews without having to go through them manually.

We've built TheReviewIndex to try and help solve this problem.

You can use our summarizer to create a scorecard for any electronics / gadgets / appliance Amazon product.

We also call out products with fishy / spammy patterns, so that you don't end up reading biased reviews.

Do let us know if you find this useful !