Interesting approach! However the article ignores the long diffusion time of insulin released under the skin. Unless the insulin produced by their beta cells is

Interesting approach! However the article ignores the long diffusion time of insulin released under the skin. Unless the insulin produced by their beta cells is different from what I shoot, it takes hours for subcutaneous insulin to get fully absorbed into the bloodstream where it is needed. There are two main problems with that: Regulation: With lead times measured in fractions of hours, the response is very delayed compared to a pancreas that releases directly into the bloodstream where insulin has immediate effect. When a device under the skin measures higher blood sugar, it is already too late and you need a muted response to prevent a swing in the other direction later. Locality: Insulin leads to tissue growth and must not be released in the same area for a long time. The body will grow a lump of fat tissue around a device that releases insulin. So I wonder how they want to address these issues.