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

If you want to build a prod ready RAH architecture with decent benchmark scores I can recommend this blog post based on our experiences what techniques actually work: https://bytevagabond.com/post/how-to-build-enterprise-ai-rag...

thank you for sharing your experience

This article is interesting cause of its scale, but does not touch on how to properly use RAG best practices. We wrote up this blog post on how to actually build a smart enterprise AI RAG based on the latest research if it's interesting to anyone: https://bytevagabond.com/post/how-to-build-enterprise-ai-rag...

It's based on different chunking strategies that scale cheaply and advanced retrieval


Interesting to see still solutions being developed for RAG. We developed a solution similar to yours: Automatic indexing from GDrive, SharePoint etc. and then advanced hierarchical chunking, context header based markdown conversion etc... All the tricks that were published last year while RAG was still the "new" kid in town. We finally open sourced everything as the competition from the big players (Notion AI, Google etc.) was daunting. If anyone is interested, this blog post about all the techniques we tried and what actually works is still relevant and up2date: https://bytevagabond.com/post/how-to-build-enterprise-ai-rag...


Thank you so much for this, started reading it a few min ago and already learnt quite a lot!

I like how clean and compressed the info is


After building enterprise RAG from scratch, sharing what I learned the hard way. Some techniques I expected to work didn't, others I dismissed turned out crucial. Covers late chunking, hierarchical search, why reranking disappointed me, and the gap between academic papers and messy production data. Still figuring things out, but these patterns seemed to matter most.


Believe it or not, but this is actually a usecase for crypto (evm)


Downvoted comment but it's not completely wrong.

While NFT sales and token releases on ethereum and other chains with smart contracts are absolutely botted and scalped/sniped very aggressively, there has been a lot experimentation on how to avoid this. Whitelists are one (bad, IMO) solution, various type of auctions are a fairer solution that aren't that popular in crypto because people are mostly interested in flipping the asset rather than obtaining it at a fair market price.

For things like tickets to a concert or things that people actually want to keep/consume, auctions make a lot of sense.


>While NFT sales and token releases on ethereum and other chains with smart contracts are absolutely botted and scalped/sniped very aggressively,

I didn't quite understand what you're referring to with this. Would you mind elaborating a bit? Honestly curious.


Bots will "snipe" hyped up releases of NFTs or tokens, using various techniques depending on the chain. On ethereum and most others the miners participate in this in exchange for a tip (usually an open market where people can bid to be first in line). Since miners produce the blocks they can order the transactions however they want and put the highest tipper first.

The goal is just the same as ticket scalping. The botters/snipers will just list the the asset on the market immediately at anywhere from 2x-10x the original sale price.

At its peak during NFT mania there were tips of millions of dollars being paid in tips to miners to buy up 75% of the supply of an extremely anticipated NFT collection for another couple of million dollars ( the sale price ) and still coming out in profit.


Another nice stoic-inspired visualisation of a life clock:

https://lifeslay.vercel.app/


Netlify Scheduled Functions for basic TS Code.

It's free.

https://docs.netlify.com/functions/scheduled-functions/


You only need to look for pimples or blemishes on the skin. Thats how you know what the real face is.


Yep. Everyone is focused on the background, but you don’t need that—all the fake faces look too perfect.


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

Search: