Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

aboutideasnow.com

Post reply on HN
Print view View original

21–30 of 121 posts

Re: Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

#21

Post by phgn »

afrnz wrote: post #10

Love it - I knew /about and /now but /ideas is new to me. I am adding this to my site as well. Do you filter results based on last update?

Yep at least the /ideas pages that turn up on the page load are sorted by recency. You can also see the most recent /now pages by clicking the filter icon next to the search bar and selecting /now.

And we re-scrape all indexed pages every few days.

For now the search is purely sorted by relevancy (keyword + embeddings), but it yes would be nice to prioritise recent posts somehow. Since those are much more actionable.

Re: Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

#22

Post by phgn »

city41 wrote: post #20
phgn wrote: post #19

Earlier quoted context omitted.

GPT-3.5: https://github.com/lindylearn/aboutideasnow/blob/main/apps/a... But in this case the date seems very close to the unix timestamp 0 which we use when there's no date found. Maybe it's a local timezone issue... Edit: this should be fixed now!

Oh interesting. I never would have thought AI would be used for this. Does it also find things like the meta "revised" tag or anything like that? Doing some Googling it seems like officially it should be "revision", but seems like it's very common to use "revised"

I tried using https://www.npmjs.com/package/metascraper before which I believe does check this meta tag.

But a few websites set their updated date to the current date which was annoying, maybe to rank better in Google? And some people (including me) only mention the update time in the page text content.

I've used GPT to parse human formatted dates in another project too, it's quite reliable if you validate the output timestamp. And relatively cheap too if you only pass in the first part of the page text.

Re: Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

#25

Post by city41 »

phgn wrote: post #22
city41 wrote: post #20

Earlier quoted context omitted.

Oh interesting. I never would have thought AI would be used for this. Does it also find things like the meta "revised" tag or anything like that? Doing some Googling it seems like officially it should be "revision", but seems like it's very common to use "revised"

I tried using https://www.npmjs.com/package/metascraper before which I believe does check this meta tag. But a few websites set their updated date to the current date which was annoying, maybe to rank better in Google? And some people (including me) only mention the update time in the page text content. I've used GPT to parse human formatted dates in another project too, it's quite reliable if you validate the output…

I can see how it's a tricky problem. I wish html had more structure here (and people followed the structure, a whole other problem...). FWIW, my page has a "last updated" date on its now page but comes up as 1969 in aboutideasnow.

Oh, now aboutideasnow shows no date at all.

Re: Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

#26

Post by phgn »

Retr0id wrote: post #18
nonethewiser wrote: post #13

Cool site. Maybe implement /ideas and /now pages? I kinda figured they'd be there but they are 404.

Perhaps they could redirect to https://github.com/lindylearn/aboutideasnow/issues and https://github.com/lindylearn/aboutideasnow/pulls , respectively

Very nice idea, I just implemented this!

Feel free to create some GitHub issues at https://aboutideasnow.com/ideas to grow the list :)

Re: Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

#27

Post by bilsbie »

Neat idea! But where are all these websites coming from? People have personal websites AND they share random ideas?

Btw I don’t have a website but here are some ideas I’d love to work on:

Anything applying LLM to robots

Device that rides on top of roomba and taps its bumper (to turn it) if it sees something bad.

An employment agent for letting people try out new careers.

Cheap simple pickleball out ball detection

Re: Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

#29

Post by phgn »

city41 wrote: post #25
phgn wrote: post #22

Earlier quoted context omitted.

I tried using https://www.npmjs.com/package/metascraper before which I believe does check this meta tag. But a few websites set their updated date to the current date which was annoying, maybe to rank better in Google? And some people (including me) only mention the update time in the page text content. I've used GPT to parse human formatted dates in another project too, it's quite reliable if you validate the output…

I can see how it's a tricky problem. I wish html had more structure here (and people followed the structure, a whole other problem...). FWIW, my page has a "last updated" date on its now page but comes up as 1969 in aboutideasnow. Oh, now aboutideasnow shows no date at all.

The vast majority of indexed websites don't have a date unfortunately :(

What's your website so I can take a look at the parsing?

Re: Show HN: AboutIdeasNow – search /about, /ideas, /now pages of 7k+ personal sites

#30

Post by phgn »

desireco42 wrote: post #24

Like others said, I love it. This is good way to connect people by connecting their independent sites.

That's the idea! And especially if more people start writing /ideas pages I think the search will become better and the results more actionable.

Also, if you find something to improve, people open a GitHub issue: https://github.com/lindylearn/aboutideasnow/issues

Post reply on HN