AI robotics startup Generalist AI has introduced its latest model, GEN-1.5, which the company characterizes as a "one-shot learner."
In a blog post published Wednesday, the robotics firm said the new robot foundation model can learn a new task in seconds without gradient updates or fine-tuning.
"The model learns new tasks in seconds when prompted with 3 to 12 seconds of a single demonstration, no training required," the company said.
The team also said in a thread of posts on X that the model can generalize prompts to new situations, learns from mistakes, and improvises new strategies.
"For example, after fine-tuning to use a brush to sweep a block into a bowl, it could use other tools like a dustpan to accomplish the same task with a very different strategy," the team added.
According to the company, such capabilities emerged from pretraining on large amounts of physical interaction data.
"Although the tasks are simple and success rates are modest, it's the first model we know of that exhibits the general ability to learn a wide range of dexterous closed-loop physical tasks from just one or few demonstrations," the team said.
The launch of the new model comes after the AI robotics firm was reportedly in talks in July to raise a new funding round at a $3 billion valuation. Venture capital firm 8VC is expected to lead the round, according to a Business Insider report.
Earlier in June, Generalist AI announced $400 million in new funding, which put its valuation at $2 billion and brought its total capital raised to more than half a billion dollars.