Hacker Newsnew | past | comments | ask | show | jobs | submit | hmdai's commentslogin

I absolutely love this, bonus: I can now read Cuneiform numbers, if I ever need that.

Suggestion: You can potentially show the Cuneiform time in the url.

sent at: 𒌋:𒎙𒐛:𒐏𒐗


Since the introduction of Model Context Protocol, I've been wondering why this protocol is so complicated to work with and after many wasted hours and a few MCP spec updates, I've decided to write down what I think MCP should have been and I call it Naive Context Protocol ¯\_(ツ)_/¯.

Maybe this approach is in fact naive (please tell why!), the "spec" is very minimal at this point (I will expand based on feedback here) and it probably ignores some use-case (some of them on purpose), but I would like to hear:

1. What everyone here thinks the MCP/NCP should've been/should be?

2. What use-cases would you like a context protocol to support?


I'm building a client-side encrypted personal management tool for myself, with support for file encryption:


Does anyone know if the search will be available through their api? seems like a unique offering that not even google has (to my knowledge at least)


Inspired by Paul Graham tweet:

"You could probably make a lot of money simply by investing in companies that a significant percentage of the latest YC batch use. They're the quintessential early adopters."

I wanted to see what services are used on batch W24 company websites. Of course this not a complete representation since it doesn't include internal, server side or behind authentication services.

These numbers are out of 197 total companies in batch W24:

- Octolane is the most popular company in W24 so far, used by 10 within the batch.

- 88 use Google analytics vs 2 that use Plausible

- Many use webflow (50) or framer (42) to build the website

- Youtube (19), loom (5) and Mux (4) for video


Do you have a link to the source of this info?



Try this one: https://uneven-macaw-bef2.hiku.app/app/

It loads the LLM in the browser, using webgpu, so it works offline after the first load, it's also PWA you can install. It should work on chrome > 113 on desktop and chrome > 121 on mobile.


Read more about why webgpu is required here: https://webllm.mlc.ai/ (that's the project that is used here)


You're good, sorry dropped the /s


Can you at least provide guidance on how to avoid wrongly getting flagged like this?


The underlying model is created for webgpu, not to mention the feasibility of running LLM with webgpu, read more here: https://webllm.mlc.ai/


I wanted a privacy preserving AI chat bot since I didn't feel comfortable about putting personal issues into ChatGPT. It uses webllm models to run in the browser. Make your own chat by saving a message.

Disclaimer: I'm the creator of https://hony.ai that was used to build this demo.


Ok, but how is Marcus Aurelius involved? Does it use search the Meditations with RAG, or does it use a model fine-tuned on extracts from the Meditations? Or...?


Ah, a most excellent inquiry, my dear! adjusts philosopher's cloak As Marcus Aurelius, I must say that the connection between Marcus Aurelius and this chatbot is purely nominal.


You can make any chat you want by saving the last message, e.g. if you just need a friendly chat bot, you can create that as well.


Cool! Can webLLM handle inference of models with any meaningful size?

Can I ask what model is used?


Thanks! It's using Llama 2 7B, It supports bigger models but those take longer to download and also infer (if run at all depending on the device)


[flagged]


Rubber ducking is worth a lot, and have the Meditations vaguely quoted back at you seems like it would help out on occasion.


LLM's are surprisingly good at giving advice about minor personal problems and social situations, probably on par with a newspaper advice columnist. It's not therapy but that doesn't mean it's not useful. The main problem is it's relentlessly positive and it's not going to give you 'tough love' advice, so if you're really a mess you're not going to hear what you need to hear.


I think I can't trust a therapist enough to be completely honest, while I can just write anything into the chat, knowing no one, (except google) will ever know :D


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

Search: