#122 - AI for Word and Excel, leaked Google memo, ImageBind, LLAva, robot soccer, Midjourney 5.1
Summary
Hosts Andrei Kurenkov and Jeremy Howard discuss recent AI developments, including Microsoft’s expanded Copilot features, Waymo’s autonomous driving growth, and IBM’s strategic pivot with Watson. They analyze a leaked Google memo questioning proprietary model advantages amid the rise of open-source alternatives like Llama. The episode also covers Palantir’s military AI tools, Chegg’s stock volatility due to ChatGPT, and concludes with career advice for aspiring AI professionals.
Topics discussed
Sponsor message: Box and AI agents
Intro, host banter, and listener comments
Microsoft Copilot and enterprise AI tools
Informatica's generative AI for data management
LinkedIn AI messages for hiring managers
Waymo expands autonomous taxi service
Leaked Google memo on open source AI threat
IBM Watson relaunch and hiring pause
Military AI and autonomous drone ethics
Chegg stock reaction to ChatGPT fears
Microsoft and AMD chip partnership
Former Google researchers launch new AI startup
Meta's open source multimodal models
Open source code models and multimodal vision
OpenAI research on interpreting AI neurons
AI mind reading and autonomous science
Robotics: quadrupeds and soccer robots
US export controls on AI chips to China
Constitutional AI for alignment
AI scraping tools and voice cloning scams
AI in politics, SAG-AFTRA strike, and music
AI music copyright and Midjourney v5.1
Listener Q&A on college and AI careers
Listen ad-free on Castria