• Bye bye, human TikTok dancers – Animate Anything is a state-of-the-art image-to-video framework that you can use to make anyone (or any drawing/anime character/etc) dance the exact dance you want, using only a single reference image. - Animate Anything
• Talking heads – GAIA (Generative AI for Avatar) is a framework that can make someone talk, or animate facial movements, from a single reference image, it “enables different applications like controllable talking avatar generation and text-instructed avatar generation”. - GAIA
• In 4 minutes, learn how to create your own AI photo model like Aitana using free and openly available tools (also a great video to see how evolved the image generation ecosystem already is). - YouTube
• Using advanced voice analysis methods, Forbes has doxxed Beff Jezos, the creator of the influential e/acc movement – turns out it’s Guillaume Verdon, founder of deep tech startup Extropic (“building the ultimate substrate for Generative AI in the physical world” and “fundamentally new computing”), with experience working on “special projects at the intersection of physics and AI at Alphabet, X and Google,” and a theoretical physicist before that. - X (Twitter), Forbes
• Lawmakers in Porto Alegre, Brazil, unknowingly passed a legislation written by AI – a councilman used ChatGPT to draft a proposal concerning water consumption meters, it was passed unanimously, and he disclosed his use of AI only after the fact. - AP News
• Understand LLMs layer by layer, with detailed visualizations and explanations, down to every tiny step the model makes – explore the LLM Visualization created by software engineer Brendan Bycroft. - LLM Visualization
• The consequences of OpenAI drama are still unfolding – according to one source, “things still feel shaky” at OpenAI, and they still haven’t been told why Sam Altman was originally fired, only that it wasn’t over AI safety. - Semafor
