(Not so) new AI breakthrough: learning robots
Sometimes I post things related to AI that believe will affect our work as designers. As Jorge Arango wrote in his newsletter (https://jarango.com/2024/01/21/five-future-roles-for-designers), design jobs will change shape and probably completely new design roles will arise. Personally, I believe these new roles will be a lot closer to the job profile of Susan Calvin, the fictional robot trainer and psychologist in Asimov’s novels.
Some times ago I posted about Rabbit AI and its learning model (LAM) related to making actions in the digital space. Now, I am flabbergasted, because these approaches are moving fast from digital to physical. I can now ask rabbit to “book a flight to Munich, second class, window, less that $100” and it will do it, as my digital concierge (agent). But we are really close to have a personal robot (nothing fancy as Asimov’s, just imagine a big arm on wheels, with robotic “hands” and accessories) and picture yourself, from your couch or desk, while you are working, shouting “Hi Dave, prepare an omelette”. Dave (the robotic arm on wheels) will take out ingredients from your fridge, kitchen cabines, prepare your recipe and cook it on the stove. Dave will also serve you at your desk and clean everything out, on your command.
How close are we to this scenario? Well, things are rolling out in the business world: these robots will dominate the supply chain… Or, to be precise: robots have been dominating our supply chain for decades, now we are going to link them to learning AIs orchestrating them.
I am not writing about something that is a scientific or industrial possibility yet to be researched: such research trains have gained steam about 10 years ago! Now, in 2024, we are on the verge of real applications that will roll out faster and faster (remember: AI is an exponential transformation technology), in particular because they will generate a huge lot of savings in the supply chain (and, in the future, in other contexts).
[…] Three of OpenAI’s early research scientists say the startup they spun off in 2017, called Covariant, […] unveiled a system that combines the reasoning skills of large language models with the physical dexterity of an advanced robot.
The new model, called RFM-1, was trained on years of data collected from Covariant’s small fleet of item-picking robots that customers like Crate & Barrel and Bonprix use in warehouses around the world, as well as words and videos from the internet. In the coming months, the model will be released to Covariant customers. The company hopes the system will become more capable and efficient as it’s deployed in the real world.
Full paper here:
