2 Stanford founders showed me how on-device AI will disrupt user interface | Nexa AI, Octopus

Описание к видео 2 Stanford founders showed me how on-device AI will disrupt user interface | Nexa AI, Octopus

Reach out to Alex & Zack   / wei-chen-stanford   |   / zack-z-li   | I interviewed Alex & Zack, who are the co-founders at Nexa AI, an on-device LLM AI agents startup funded by Fusion Fund here in Silicon Valley. Alex and Zack were both Stanford graduates who are passionate about making fundamental breakthroughs using AI. NEXA AI envisions a world where software interactions occur through natural human language. They aim to transform human-computer interaction from GUI interactions to natural language prompts using accurate, fast, and affordable AI agents. It was absolutely a wonderful experience hearing them share their stories on how they met, build and iterate. Let's dive in!

Nexa AI Links:
https://www.nexa4ai.com/about-us
https://huggingface.co/NexaAIDev

Follow me for more:
LinkedIn/Twitter/Insta/Discord:   / shen-sean-chen   |   / shenseanchen   |   / sean_tech_stories   |   / discord  

Hope you'll enjoy this video~
00:00 A few interesting shots...
00:15 Starts!
00:37 Founders' intro
01:11 Use case of Nexa AI
01:33 Original product demo
02:32 Voice controlled demo
04:09 Fundraising stage
04:24 Initial Ideas that got them the investment
04:53 More enjoyable to build for fundamental breakthroughs
05:43 How they met and came up with the ideas
06:34 What they studies to prepare for AI startups
07:56 'Aha' moments for deciding to build together
10:12 Android VS iOS for building product demos
10:37 Can it be a webApp
11:04 Potential competition from Siri or Gemini
12:14 Cloud-based VS on-device tech solutions
12:29 Nexa is a general framework solution enabling on-device models
14:12 Outperformance over RAGs (Retrieval Augmented Generation)
17:02 Why is RAG slower than Nexa AI
18:48 Stories of Pivoting
21:47 Quick Questions Quick Answers
22:56 Experience at Stanford University
24:29 Role Model in Tech/Entrepreneurship
25:29 Closing Comments

Комментарии

Информация по комментариям в разработке