""It's essentially a grow box with some sensors for collecting data," Cerqueira, a dietitian who worked as a project manager at the Open Agriculture Initiative for two years [...] The boxes were not air-tight, so staff couldn't control variables like the levels of carbon dioxide and even basic environmental factors like temperature and humidity, Cerqueira and the other person said. "
Looking at https://www.media.mit.edu/groups/open-agriculture-openag/pro... it looks like the food computer project was just one among several, though it might've been the more prioritized 'audacious goal' among them considering it's prominence on the main open ag site: https://www.media.mit.edu/groups/open-agriculture-openag/
The last paper I saw (2018-10-20) ( https://link.springer.com/chapter/10.1007%2F978-3-030-02683-... ) on food computer version 2 is more about the building process of it, not so much on results yet. https://wiki.openag.media.mit.edu/faq discusses the version differences.
Finally, Caleb comments on the BI article here: https://forum.openag.media.mit.edu/t/can-we-get-a-comment/51...
(sidebar: the food computer thing weirdly reminds me of reddit.com/r/spacebuckets )
On top of that, the headline claim that the "personal food computer" "simply doesn't work" seems to be outright untrue. I had a look and the designs seem to be pretty widely replicated and tested - everything from an MIT-supplied version of the latest design in a New York museum operated by unaffiliated volunteers to a homebrew version of an old foam-based design built from scratch by a Baltimore school using material from Lowes. Some versions do seem to be a little temperamental on the electronics side, to the point members of the team have advised people not to built v2 on the forum, but they do work and grow things. I wouldn't be surprised if some of the photoshoots and static displays were faked, though; turnaround time for a real grow is at least a month.