Hacker News new | past | comments | ask | show | jobs | submit | rishikeshs's comments login

I was thinking about this recently! Can we make an LLM-in-a-box ?


The ability is there for sure: https://www.prepperdisk.com/products/prepper-disk-premium-ai...


Oh good! Offline slop! /s


Just a small suggestion, please take it constructively! I had to click on the app store link and check the screenshots and description to understand what this was. Some info on the landing page would have been better!


But is it a front end issue?

I can understand the time taking to process, but the whole tab becomes unusable after a while!


Yeah sounds like a lack of ‘virtual scrolling’. Behind these scenes, they are often creating a tremendous number of markdown renderer components, which are expensive and not meant to have hundreds of instances on a page. It’s a tricky problem imo, and involves tradeoffs with the scrollbar ergonomics.


I think it's both. If you look at the response coming back from claude you can get an idea for how the front end library probably works. (When updating an artifact, claude's response is inefficient. Perhaps this is more reliable at scale but... why not use a diff format similar to git?)


Working on my plain text cricket live cricket site[1]. Just realised after starting that getting reliable live cricket or sport data is super hard. Tried a paid provider[2] and it’s frustratingly inaccurate. Need to figure out a better way!

Also, figuring out a way to visualise manhattan chart, score worm, wagon wheels etc using plain text ascii

[1] https://criclite.com [2] https://cricdata.org


https://llmstxt.org/


Never thought a post about Kerala will reach the front page of HN.


I was looking for live score data. Not sure how I can get that from discussions


People will presumably quite quickly start discussing events. Score data can be inferred from that.


Interesting! That’s a great angle to look at.

But don’t know how hard it would be to process the data.


That's why you pass it to a big, clever neural network with some reasoning and let it figure it out for you.

Put multiple different sources through multiple different LLMs and take a consensus of them. If there isn't a clear winner, declare it uncertain.

It'd be a bit of a bodge, but it'd work.

If you can get a live feed of the broadcast, you could also try feeding visual information in to be processed by a multimodal LLM.


I was thinking something like this:

Have the cricket score site open in a VM or something. Then a webjob can take screenshots and process the screenshot to extract the score.


Thanks a lot for this!

Yes the site was inspired by plaintextsports.com and I have credited them as well!

One thing I decided was I will use only ASCII and will stay away from unicode chars. But now when I’m trying to plot a wagon wheel or a heatmap, it’s becoming tricky. i will update you once I have something solid!


Is gnuplot plain text ascii?

do you mind sharing the link to ploot?


  gnuplot> set terminal dumb

  Terminal type is now 'dumb'
  Options are 'feed enhanced size 79, 24 aspect 2, 1 mono noattributes'
  gnuplot> plot sin(x)

    1 +--------------------------------------------------------------------+   
      |                *  *              +  *  **         +       *  *     |   
  0.8 |-+             *   *                 *    *          sin(x* *******-|   
      |              *     *                *    *               *    *    |   
  0.6 |-+            *      *              *     *               *     * +-|   
      |              *      *             *       *             *       *  |   
  0.4 |*+            *      *             *       *             *       *+-|   
      |*            *        *            *        *           *        *  |   
  0.2 |*+           *        *            *        *           *        *+-|   
      | *          *          *          *         *          *          * |   
    0 |-*          *          *          *         *          *          *-|   
      |  *         *          *         *           *         *           *|   
 -0.2 |-+*         *          *         *           *         *          +*|   
      |  *        *            *       *             *       *            *|   
 -0.4 |-+*        *            *       *             *       *           +*|   
      |   *      *              *      *             *      *              |   
 -0.6 |-+ *     *               *     *              *      *            +-|   
      |    *    *               *     *               *     *              |   
 -0.8 |-+   *   *                *   *                 *   *             +-|   
      |     *  *       +         **  *   +             *  *                |   
   -1 +--------------------------------------------------------------------+   
     -10              -5                 0                5                10


wow this is so cool


git://bitreich.org/ploot


Haha great implementation! I also have a similar one on my blog called: Useless Button[1]. This one is implemented with CF Workers as the backend.

This button has been clicked 2475 times for no reason.

[1]https://rishikeshs.com/button/


This is really cool, now I'm tempted to add some kind of interactive element like this to my website. reminds me of the 2000s internet online user/views counters


Check out this mosaic as well: https://rishikeshs.com/mosaic/

I'm mesmerised by how many people take time to click all these! :D

It has been updated by people from: AE, IT, FR, US, IN, BE, GB, AU, PL, MX, DK, BR, JP, DE, CH, IL, SK, ES, SE, CA, NL, AR, NZ, PF, IS, CO, CL, GE, PT, RO, FI, PH, CZ


Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: