We of course covered everything Apple Intelligence, from the tech stack perspective (3B param model on device with GQA, LORA swapping, spec. decoding) and the handoff to private cloud
x.com/altryne/status/1800540764612841911
Also covered the AI OS use-cases, like a smarter Siri, Genmoji & Image Studio + built in LLM text features into the OS, everywhere you write, you'll have AI assistant helping you rewrite, summarize, expand or contract and much more like notification summarization & priority
x.com/altryne/status/1800207009033474142
While we focused on Apple, we also had other things to talk about, @swyx came by to talk about the upcoming @aiDotEngineer event, spilled the beans on my surprise talk and shared a coupon code for ThursdAI listeners (you'll have to listen to the episode to get them)
x.com/aiDotEngineer/status/1798028596910653591
Of course, friend of the pod @daken_ sent us a preview of @LumaLabsAI the second it was public, so we had to cover the LDM (Luma Dream Machine) and it's comparison to SORA and other video models (thanks to @amebagpt for the comparison)
x.com/amebagpt/status/1801234328225018021
It really was a great conversation, if you found it, the newsletter or this thread helpful, please follow @altryne and @thursdai_pod to never miss an episode or when we go live (every ThursdAI 8:30am pacific!) and give us a boost to the first tweet, because X deboosts š
x.com/altryne/status/1801650253403197744