/sci/ - Science & Math

>>11991904
Anyone with a decent grasp of the literature knows that Neuralink promises the world but has not delivered on anything that actually advances the field. I'd be happy if I'm shown wrong, but so far their output is distinctly underwhelming.

For example, their device:
- contains a single channel per thread (1 is not enough for stable single-neuron resolution since you can't triangulate activity; probes that are widely used today already contain ~300 channels).
- performs pre-amplification not at the probe level (again, the standard) but at the chip level where multiple probes are combined (meaning any measurements will have very low SNR)
- contains thin threads which is good to minimize bleeding but also precludes sub-cortical implantation, which is at odds with the stated goal of developing treatment plans in humans (literally all DBS treatments target sub-cortical areas and as such require rigid probes that can reach the appropriate depth).

>>11991926
>There will be an official update on the 28th of August.
I'm not sure if you've read the responses to this announcement from neuroscientists, but this will be a joke. They're planning to show "neurons firing", which we were able to do literally 80 years ago (pic related).