#165 - Sora challenger, Astribot's S1, Med-Gemini, Refusal in LLMs

Our 165th episode with a summary and discussion of last week's big AI news!

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

Email us your questions and feedback at contact@lastweekin.aiand/or hello@gladstone.ai

Timestamps + links:

• Tools & Apps
• () GitHub releases an AI-powered tool aiming for a 'radically new way of building software' (zdnet.com/article/github-releases-an-ai-powere…)

• () China unveils Sora challenger able to produce videos from text similar to OpenAI tool, though much shorter (scmp.com/tech/tech-trends/article/3260790/chin…)

• () ChatGPT’s AI ‘memory’ can remember the preferences of paying customers (theverge.com/2024/4/29/24144680/chatgpt-plus-m…)

• () Rabbit R1 review: Avoid this AI gadget (tomsguide.com/ai/rabbit-r1-review)

• () Amazon Q, a generative AI-powered assistant for businesses and developers, is now generally available (aboutamazon.com/news/aws/amazon-q-generative-a…)

• () Yelp’s Assistant AI bot will do all the talking to help users find service providers (theverge.com/2024/4/30/24144812/yelp-assistant…)

• Applications & Business
• () Video of super-fast, super-smooth humanoid robot will drop your jaw (newatlas.com/robotics/astribot-s1-fast-humanoi…)

• () Tesla’s 2 million car Autopilot recall is now under federal scrutiny (arstechnica.com/cars/2024/04/feds-concerned-so…)

• () Tesla shares soar as Elon Musk returns from China with FSD 'Game Changer' (thestreet.com/investing/stocks/tesla-shares-ju…)

• () OpenAI inks strategic tie-up with UK’s Financial Times, including content use (techcrunch.com/2024/04/29/openai-inks-strategi…)

• () OpenAI Startup Fund quietly raises $15M (techcrunch.com/2024/04/26/openai-startup-fund-…)

• () Huawei backs HBM memory manufacturing in China to sidestep crippling US sanctions that restrict AI development (tomshardware.com/pc-components/dram/huawei-bac…)

• Research & Advancements
• () Capabilities of Gemini Models in Medicine (arxiv.org/abs/2404.18416)

• () Let's Think Dot by Dot: Hidden Computation in Transformer Language Models (arxiv.org/abs/2404.15758)

• () NExT: Teaching Large Language Models to Reason about Code Execution (arxiv.org/abs/2404.14662)

• () SenseNova 5.0: China’s latest AI model surpasses OpenAI’s GPT-4 (interestingengineering.com/innovation/china-se…)

• () Octopus v4: Graph of language models (arxiv.org/abs/2404.19296)

• () Better & Faster Large Language Models via Multi-token Prediction (arxiv.org/pdf/2404.19296)

• Policy & Safety
• () Refusal in LLMs is mediated by a single direction (lesswrong.com/posts/jGuXSZgv6qfdhMCuJ/refusal-…)

• () Rishi Sunak promised to make AI safe. Big Tech’s not playing ball. (politico.eu/article/rishi-sunak-ai-testing-tec…)

• () DOE Announces New Actions to Enhance America’s Global Leadership in Artificial Intelligence (energy.gov/articles/doe-announces-new-actions-…)

• () The Chips Act is rebuilding US semiconductor manufacturing, so far resulting in $327 billion in announced projects (techspot.com/news/102770-how-chips-act-remade-…)

• () Analysis-Second global AI safety summit faces tough questions, lower turnout (finance.yahoo.com/news/analysis-second-global-…)

• () Sam Altman, Jensen Huang, and more join the federal AI safety board (fortune.com/2024/04/26/federal-ai-safety-board…)

• Synthetic Media & Art
• () Air Head creators say OpenAI's Sora finicky to work with, needs hundreds of prompts, serious VFX work for under 2 minutes of cohesive story ↺ (notebookcheck.net/OpenAI-s-Sora-finicky-to-wor…)

• () Eight newspaper publishers sue OpenAI over copyright infringement (cnbc.com/2024/04/30/eight-newspaper-publishers…)

See Privacy Policy at art19.com/privacy and California Privacy Notice at art19.com/privacy#do-not-sell-my-info.

#165 - Sora challenger, Astribot's S1, Med-Gemini, Refusal in LLMs

Kapitelklicken springt direkt hin

Shownotes