Ask HN: How is your organization using LLMs?

https://news.ycombinator.com/item?id=41000150

by cloudking • 2 years ago

19 8 2 years ago

news.ycombinator.com

I'm curious to hear how different organizations are integrating large language models (LLMs) into their workflows. Are you using them for customer support, content generation, data analysis, or something else? How have LLMs impacted your operations, and what challenges or benefits have you encountered?

brianjkim21• 2 years ago

PMs and Execs are pushing devs to do docs classification and extract data from text ("see I can do it on chatGPT!") so they're using Taylor (trytaylor.ai) to build production grade text pipelines ;)

In all seriousness, customer support was the first but the least impactful area for LLMs. Currently, LLMs are primarily used for developer efficiency and for info retrieval.

iAkashPaul3• 2 years ago

Heavily,

Call summarisation at scale, for further insights of key topics/sentiments for diarized & fused channel call data

Lead mining quality improvement by asking questions instead of merely classification

Internal simple RAG chatbot across ~1000 PDFs, not yet down the GraphRAG route

Better chatbot experience over RASA

Contract mining for speeding up mundane governance tasks

fnetisma• 2 years ago

I’m curious, which software do you use for the lead mining quality improvement use-case you mentioned?

iAkashPaul3• 2 years ago

Mostly Mistral/Mixtral models in addition to Whisper.

gkorland• 2 years ago

Would love to hear your feedback once you do try GraphRAG.

1oooqooq• 2 years ago

how do you know the caller didn't mention anything confidential if you're only seeing it after you send it to openai?

iAkashPaul3• 2 years ago

Where did you find OpenAI in my reply?

nicbou• 2 years ago

I run a website alone. I often use it to find better ways to explain things in plain English. It's good to see how ChatGPT explains a thing, and use it to improve my own explanation.

I'm now testing automated translations of my Markdown-based content, but the unpredictable nature of LLMs mean a lot of potentially costly errors could slip through. However, offering a website in multiple languages without additional labour is very appealing. I could reach a far greater audience, and help a lot more people.

I also use the far better modern text to speech APIs to tell my readers how to pronounce complex German words.

Ask HN: How is your organization using LLMs?

Related Stories