#145 - OpenAI’s Q*, Claude 2.1, Stable Video Diffusion, Starling-7B, Orca 2, international agreements

Our 145th episode with a summary and discussion of last week's big AI news, this time around with guest co-hosts Kevin and Gavin from AI For Humans podcast

Check out the AI For Humans episode on which Andrey and Jeremie guest co-host here.

Also check out our sponsor, the SuperDataScience podcast. You can listen to SDS across all major podcasting platforms (e.g., Spotify, Apple Podcasts, Google Podcasts) plus there’s a video version on YouTube.

Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

Email us your questions and feedback at contact@lastweekin.ai

Timestamps + links:

• Tools & Apps
  • (06:00) OpenAI rival Anthropic makes its Claude chatbot even more useful
  • (12:33) Stability AI debuts Stable Video Diffusion models in research preview
  • (16:33) Generative Video Startup Pika Labs Launches Version 1.0, Raises $55 Million in Funding
  • (25:25) Screenshots show xAI’s chatbot Grok on X’s web app
  • (30:37) Amazon Introduces Q, an A.I. Chatbot for Companies
• Applications & Business
  • (34:28) Unpacking the hype around OpenAI’s rumored new Q* model
  • (47:37) NVIDIA Announces Financial Results for Third Quarter Fiscal 2024
  • (52:40) Nvidia Delays Rollout Of China-Focused AI Chip: Report
  • (56:31) Share sale set to test financial impact of OpenAI’s leadership turmoil
  • (58:55) Amazon and Salesforce Expand Partnership to Add New AI Capabilities
  • (59:55) AI21 takes Series C to $208 million with additional $53 million in funding
• Projects & Open Source
  • (01:03:15) Starling-7B: Increasing LLM Helpfulness & Harmlessness with RLAIF
  • (01:12:10) Defending your voice against deepfakes
• Research & Advancements
  • (01:16:07) Orca 2: Teaching Small Language Models How to Reason
  • (01:22:16) New technique can accelerate language models by 300x
  • (01:23:35) DeepMind Says New Multi-Game AI Is a Step Toward More General Intelligence
  • (01:27:00) GAIA: A Benchmark for General AI Assistants
  • (01:30:45) Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
• Policy & Safety
  • (01:33:42) US, Britain, other countries ink agreement to make AI 'secure by design'
  • (01:36:07) Chinese GPU recycling factories have a workaround for the US government's newly imposed export rules — modified NVIDIA RTX 4090 cards serve as great AI accelerators
  • (01:39:00) The EU AI Act needs Foundation Model Regulation
  • Putin to boost AI work in Russia to fight a Western monopoly he says is ‘unacceptable and dangerous’
  • Inside U.S. Efforts to Untangle an A.I. Giant’s Ties to China
  • (01:41:55) More than half of Americans are worried about AI than excited
  • (01:46:10) State Dept prioritizes ‘AI-ready workforce’ in its first AI strategy
  • US chip export ban is hurting China’s AI startups, not so much the giants yet
• Synthetic Media & Art
  • (01:48:18) Sarah Silverman Hits Stumbling Block In AI Copyright Infringement Lawsuit Against Meta
• (01:52:23) Outro