It keeps getting more impressive. This audio2video can take a still portrait of a person, an unrelated audio source, and make it come alive. It’s still an academic research project called Emote Portrait Alive (EMO), but it shows yet another leap in AI progress. There’s more here: https://lnkd.in/gEXvFJga https://lnkd.in/g6r2SHCj
This is next level! OpenAI dropped a surprise announcement of Sona, a new AI model to create text-to-video. It uses awareness of real-world physics to generate hyper-realistic videos. This goes way beyond anything we’ve seen to date. Take a minute and at least skim through the videos.“𝘈𝘭𝘭 𝘷𝘪𝘥𝘦𝘰𝘴 𝘰𝘯 𝘵𝘩𝘪𝘴 Read more…
This is scary. $25M swindled by a deep fake video conference call, completely fooling the staff member into thinking that his team was instructing him to make 15 money transfers totaling HK$200 million. https://interestingengineering.com/culture/25-million-swindled-as-deep-fake-cfo-tricks-finance-worker
Create a custom GPT that creates events on your Google Calendar! I created a video walkthrough to show how to build this and how easy it is. This is my first time in a few years creating a YouTube video, so I’m just getting back into the swing of things, Read more…
Microsoft released exciting research on a method that can improve even existing LLMs, which they have called Layer-Selective Rank Reduction (LASER). They are seeing 20-30% model improvements with reduced memory requirements. It appears relatively easy (for the engineers) to implement, so I believe we’ll hear more about LASER as it’s Read more…