x NeonPulse | Future Blueprint
Posts
🤖 ZoomAI, ChatGPT reflects, and Elon fires shots at Gates

🤖 ZoomAI, ChatGPT reflects, and Elon fires shots at Gates

3/29/23

Neon Pulse Team
March 29, 2023

Good morning and welcome to today’s issue of neonpulse!

Here’s what we have for you today:

🤖 Zoom Rolls out AI Features

🔫 Elon takes a shot at Gates

🌊 GPT-4 Gets 30% Performance Boost

Lets get into it!

Zoom AI

Zoom’s new AI whiteboarding tool

Zoom is the latest arrival to the AI party, releasing it’s new “Zoom IQ Smart Companion” developed in partnership with OpenAI.

Highlights of the new AI features:

A smart transcription service that generates real-time captions during meetings
An intelligent note-taking service that automatically captures and organizes meeting notes for users
A virtual meeting assistant that provides meeting insights

And best of all, Zoom IQ will help catch you up on everything you missed when you arrive late to meetings, making it slightly less likely that your boss will replace with you with a robot.

You can learn more on the company’s blog here.

Musk vs Gates

Apparently Elon Musk is not impressed with Bill Gates A.I. aptitude.

Responding to a tweet about Gates’s essay “The age of A.I. has begun,” in which Gates spoke about his interactions with the team at OpenAI, Musk commented “I remember the early meetings with Gates. His understanding of A.I. was limited. Still is.”

Originally a board member of OpenAI, Musk left in 2018 citing conflicts of interest due to his work at Tesla and has since become a vocal critical of OpenAI’s shift to a for-profit business.

“I have no control & only very limited insight into OpenAI,” Musk said, adding that his confidence in its safety was “not high.”

Musk has since paused OpenAI’s access to the Twitter database, claiming that he would need to know more about the company’s “governance structure & revenue plans going forward” to allow them access.

What would an AI newsletter be without a little tech billionaire drama?

GPT-4 Boosted 30%

According to a new research paper, the performance of language models like ChatGPT can be improved significantly by training them to reflect on their own errors.

By using "meta-cognitive training," in which ChatGPT is asked to reflect on it’s errors and analyze it’s thought processes, the performance of ChatGPT can be improved by up to 30%.

Imagine how powerful I could be if I took the time to learn from my mistakes.

I’m not actually going to do it, but just imagine…

You can check out the full paper here.