Isnt this a general reinforcement learning agent with a transformer as the policy discriminator? Very cool, but not necessarily a giant leap forward, more like a novel combination of existing tools and architectures. Either way impressive.
I haven't read the paper yet but it looks like the breakthrough is that it uses the "same weights" for tasks in completely different domains.
Which implies that it can draw from any of the domains it has been trained on for other domains. Speculating here but for example training it on identifying pictures of dogs and then automagically drawing on those updated weights when completing text prompts about dog properties.
If my interpretation is correct then this is a pretty big deal (if it works well enough) and brings us a lot closer to AGI.
