#assistive technology

An IEEE Spectrum opinion piece argues that exoskeletons and brain-computer interfaces should now be judged less by controlled demos and more by whether they deliver durable value in the messy conditions of real-world use

Researchers implanted rhesus macaques with roughly 300 electrodes and used AI to decode brain signals into movement in virtual environments, pointing toward more intuitive interfaces for paralysis and mobility aids.

A long-term user’s experience shows how real-world feedback is influencing the design direction of modern exoskeletons, including self-balancing systems intended to make robotic walking more practical.

Patrick Darling lost his ability to speak and sing to ALS. Using AI trained on noisy pub recordings and kitchen videos, he created a voice clone so accurate that audiences could not tell the difference.