Its a neat headline, but its unclear what problem this is trying to solve. I hope the first viable versions it don't colonize the poor grad student.

I think these analogies miss the forest for the trees. This is a wholly different idea and approach than prior art related to insulin, synlogic's trials, or fermentation (the production of products by microbes in a reactor)

> 1) Medical insulin is already produced by engineered microbes, and has been for 40 years. So we know microbes can do this safely and effectively.

This is in a sterilized tank with defined media and careful controls, not on a human body.

> 2) Insulin is a single gene, so it's easy to engineer. We can move it around to different organisms and add different control systems to explore where and how it is expressed.

Syn bio control systems are so coarse and unreliable I would never trust my life to one. Would you?

> 3) Diabetes is a well studied disease. We know a lot about how much insulin patients need, at what times, in response to what signals. So it is a well defined engineering problem.

We do, but microbes don't

The synlogic trials are essentially for augmented digestion, something we've always relied on microbes for. That's a lot safer than a feedback-control mechanism for an essential hormone.

This is the creation of a (hopeful) symbiosis with the hope that the steady state for one organism is also the optimal steady state for the other.

This idea is also the release into the wild of an organism designed to live on or in humans with the hope it does not become a pathogen, and the hope that its host will never enter a metabolic state where it assumes the role of a pathogen.

Isn't this just gain-of-function research with another name? Feedback and control in syn-bio is coarse to say the least. Is it responsible to be developing this tech with a microbe that can live on mammals?

I must add, I am _profoundly_ pro-syn-bio. I think this field needs to take itself more seriously before something bad happens.