Relearning to walk — what AI is missing | Khurram Javed, Oak Lab
AI systems lack brain-like plasticity to update deeply ingrained weight-level knowledge when it becomes outdated
“Something that has been true for 20 years when it stops being true it can go and update that and get rid of that. And that is the capability I think that's extremely useful we would want in our systems.”
Khurram Javed of Oak Lab argues that current AI systems rely on a split between frozen fundamental skills in weights and personalization via context, but this architecture is insufficient. Using the analogy of humans who lose proprioception and must relearn to walk over years using visual feedback, he contends that AI needs the ability to update deeply ingrained weight-level knowledge when it stops being true. This frames continual learning and neural plasticity as a critical unsolved gap in current AI systems.