In the newsletter this week, next to the notions of news on AI, robotics and beyond, I reflect on some thoughts triggered by the newly announced MM1 through a paper in combination with some thoughts triggered by two artists: Marina Abramovic and James Bridle. I end the newsletter with some possible exciting events for this week.
Triggered thought
The new battlefield is multimodal AI, I did mentioned it before, among others with the new Rabbit-device (edition 275). A logic successor in the process from chat to agents. Apple published a research paper last week to claim some of the space here. It would be wise for them of course, to put their foot between the door specifically in that domain; here lies the potential unique proposition for Apple with the huge installed base… Ben Thompson wrote about it in his update this Monday that MM1 is a family of models, the largest of which is 32 billion parameters; that’s quite small — GPT-4 reportedly has ~1.7 trillion parameters — but performance appears to be pretty good given the model’s size. Unfortunately, the models were not released, but the paper is chock full of details about the process Apple went about creating the models and appears to be a very useful contribution to the space
Check out the published paper, and some articles on MM1, like this and this.
Another way of thinking about AI connected to physical world interventions is described in a lovely new experiment by James Bridle: AI Chair 1.0. How do you prompt AI for instructions on how to design a chair based on a pile of scrap wood? They are also building the chair based on the instructions.
Browsing the article, I have to think about the recipes that Marina Abramovic is creating for her performances. Last weekend, I was at the exhibition in Amsterdam, and the well-known performance artworks are intriguing and make you think and feel through the absorption of the performances. An extra asset is the recipes that are hanging next to the pictures or videos of the work. Clean instructions are a key part of the experience. The AI Chair is also made through this kind of instruction, so you wonder if the interaction of generative AI with humans, as soon as the real world is part of the equation, is an act of performance art? What does that act of interacting mean? Who is the creator of the performance?
Read the notions of the news, paper for this week and events to track, via the newsletter. You can also subscribe for weekly updates on Tuesday 7 am CEST.
target is new 