There's a bunch of conditions that can be treated using different kinds of implants into the brain - deep brain implants for epilepsy, cochlear implants for severe hearing loss, we've been putting shunts into babies' brains to treat hydrocephalus for probably more than 50 years at this point. Musk is just pushing the envelope, as usual.
The closest analogies to Musk's implant are probably the cochlear implant and some of the experimental artificial vision stuff, but they're putting signals into the brain, there's nothing that extracts useful information.
What comes close is some of the research using fMRI or electroencephelograms to pick up gross brain concepts from outside the skull, but Musk needs to significantly level up from that for Neuralink to be seen as a success.
The reports seem to be that he's looking at people with quadriplegia as a result of ALS* or spinal injury as his test subjects. IMO that suggests he's significantly backed off from what he expects to achieve. The only way they're getting someone walking with the current setup is on the level of sending gross instructions to a powered exoskeleton: stand-up, walk forward, sit down, and without a complete sensory feedback loop , that's already been done without the need for a brain implant.
* I think there's an ethical issue in recruiting people with ALS, it's such a fast progressing condition I'm not sure people are really in a healthy position to give informed consent, and how do you allow for the progress of the condition when measuring the success of the implant.