HN Search Is Misleading | Hacker News Reader

HN Search Is Misleading

When I search USMLE, it gives me list of topics, but none is related to USMLE, rather it gives topics whose title have "useless" word in it.

Am I missing something?

View on HN

Works fine for me. As you scroll down/go through the pages you'll get looser matches, but the first topic results I get for "USMLE" all seem exact:

1. "ChatGPT Has Passed the United States Medical Licensing Examination (USMLE)" - https://news.ycombinator.com/item?id=34483030

2. "Show HN: Medical LLM API on par with Google Med-PaLM 2. 92% USMLE accuracy" - https://news.ycombinator.com/item?id=37038021

3. "Med-PaLM 2 reaches a score of 86.5% on MEDQA (USMLE)" - https://news.ycombinator.com/item?id=35971465

And for comments:

1. "[...] akin to passing the bar or USMLE could go a long way to solving this problem [...]" - https://news.ycombinator.com/item?id=43315182

2. "[...] very clear that there will be significant disparities by race in USMLE boards in pretty much exactly the same pattern [...]" - https://news.ycombinator.com/item?id=43234441

3. "[...] Suggests MEDIAN USMLE step 1 scores for White, Asian, Hispanic/Latino, and [...]" - https://news.ycombinator.com/item?id=43226207

Generally find HN search works very well, being fast and comprehensive.

I see USMLE related results: https://imgur.com/a/e2JrsQC But when limiting to the shorter timeframe like 'last month' the titles containing 'useless' are returned more. Seems to be the usual too-helpful search engine behavior trying to return possible typos rather than no results.

How should I set time frame please?

When I queried using the term, it seemed to perform well, listing 13 results on a singular page. [1] All have the term [2]

Stories that are dead do not show up though.

For date ranges, there should be a pull down box up the top of the search page. I prefer to just get all by date and page to older entries as far as I need.

[1] https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

[2]

HN Search Is Misleading ... When I search USMLE,

Show HN: LLM reasoning partner for multiple choice questions ... I made this for my gf who is studying for US medical exams (USMLE Step 2)

Show HN: I made a USMLE prep AI study buddy (VQA-RAG) for my girlfriend.

Show HN: Get Your Medical Exam Results Explained – Free for Launch (Usually $10) ... USMLE by over 20 points.

Show HN: Medical LLM API on par with Google Med-PaLM 2. 92% USMLE accuracy

Med-PaLM 2 reaches a score of 86.5% on MEDQA (USMLE)

Chat GPT Passes US Medical Board Exam(USMLE)

ChatGPT Has Passed the United States Medical Licensing Examination (Usmle)

ChatGPT Has Passed the United States Medical Licensing Examination (USMLE)

Performance of ChatGPT on Usmle: Potential for AI-Assisted Medical Education

Usmle moves to pass/fail system -- (https://www.usmle.org/announcements/)

[Med Tech]: Future family doctor looking to contribute to med tech ... My USMLE (United states medical licensing exam) scores

I get good resultr, but in the configuration section there is an option "typo tolerance", you may try disabling it.

Reply on news.ycombinator.com