Generative AI Weekly - Apr. 18, 2024

April 18, 2024

This week:

  • Llama 3 - Meta's open-source model is now available. The 400B+ version of the model is on par with Claude 3 Opus. Review here.
  • Adobe teases new generative AI features in Premier Pro. Worth watching.
  • Microsoft releases VASA - wild demos within. "TL;DR: single portrait photo + speech audio = hyper-realistic talking face video with precise lip-audio sync, lifelike facial behavior, and naturalistic head movements, generated in real-time."
  • On the hardware front,  Humane AI gets slammed in market, Rabbit R1 is now making its way to early adopters and Limitless pendant goes live.
  • Some evidence of dangers in using AI for the wrong problems… "In our field experiment with 640 Kenyan entrepreneurs… high performers benefited by just over 15% from AI advice, whereas low performers did about 8% worse with AI assistance"
  • A new sort of game…  "1. every guest gets an AI character. 2. you customize it to your personality. 3. your character is thrown into a virtual world where it meets everyone else attending the party. 4. the day of the irl party, you get a report of the top 3 ppl to meet and more importantly, who to avoid lmao."
  • The "first ever AI Interviewer" - Interview 100x more engineers async using AI.
  • On AI-enabled services businesses. "For example, a firm could presumably handle all of your go-to-market from marketing to sales, customer experience to success, in which case you could focus solely on building and refining your core product."
  • Wired on the eight Google employees who "Invented Modern AI." 
  • And my possibly biased favorite for this week. Star Trek's Holodeck recreated using ChatGPT and video game assets.
  • As for this week's interesting paper relevant for designers.  "We present an automated text animation scheme, termed Dynamic Typography..."