Recent submissions

Comment · 1 year ago
My bad. I shouldn’t have mentioned LangChain here because it’s a little besides my point. What I mean is, MCP seems designed for a world where users talk to an LLM, and the LLM calls software tools.

For the foreseeable future, especially in a business context, isn’t it more likely that users will still interact with structured software applications, and the applications will call the LLM? In that case, where does MCP fit into that flow?

Comment · 1 year ago
The question I’m wrestling with is will anybody care about MCP? I’m working on my own MCP proxy to manage security, auditing, and server management and the more I think deeply about the actual use cases the more I wonder if I’m wasting my time. Can anyone think of a world where MCP is relevant if generic chatbots (ChatGPT, Claude Desktop) don’t become the primary human-AI interface? If LLMs are still wrapped in application wrappers, isn’t ̶a̶n̶ ̶a̶p̶p̶r̶o̶a̶c̶h̶ ̶l̶i̶k̶e̶ ̶L̶a̶n̶g̶C̶h̶a̶i̶n̶ a more traditional agentic approach going to make more sense?
Comment · 9 years ago
The author raises good points which are hard to disagree with. I'm troubled by the decision to choose the odd and seemingly arbitrary "95-20 ratio" though, instead of a more typical 80-20.

Does anybody know what these graphs look like at 80-20?

Comment · 11 years ago
I've got a side project on Google Play called Airboss that does exactly this using WebRTC.

https://play.google.com/store/apps/details?id=com.airboss.ai...

It's far from perfect and has some obvious shortcomings but it works reasonably well with low latency.

Comment · 11 years ago
Comment · 12 years ago
Comment · 12 years ago
Comment · 13 years ago
Am I the only one who thought London?
Comment · 13 years ago
Ha, wish I had seen this link first. My Nexus 4 arrived last Thursday so I was well within the 15 day window. I just called their customer service (https://support.google.com/googleplay/answer/3053039) and they processed a refund for the difference. They're a bit swamped right now so you may have to hang on for 10-15 minutes before you can talk to someone.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton