Before the headline: what did the AI food study actually test?
A practical reporting brief on the Tac and Kuhl preprint, with claim boundaries, interview questions and a headline test.
All publications
Find the reporting, analysis and practical guides that matter to your work.
News · Guides & tools · How our formats work
Browse publications
Publications in date order, newest first.
GLP-1 food: start with what people actually eat
Your formula can hit its protein target while the product misses the eating occasion. Here is the question I would put at the top of the brief.
Food biotech runway needs a payment calendar
A production deposit and a customer's payment date can reshape the cash plan. Put them beside the next commitment before signing it.
What should change after an investor says no?
Seven prompts help founders review the questions raised during diligence, separate direct feedback from hypotheses and choose the next useful check.
What a food biotech cost model must show investors
One bar is measured. The others are modelled. Make the difference visible before an investor asks.
Can your plant-based team turn texture data into a better product?
Instrument readings, consumer preference and manufacturing performance answer different questions. The organisational test is whether the team can connect them and act on the result.
Why food biotech platforms should start with an ingredient
Shared biology can reduce development work. Customer qualification, regulatory scope and manufacturing costs still have to be worked out product by product.
When does a food biotech pilot become a purchase?
A pilot can produce useful evidence and binding commitments. Map what still separates the trial from the next buying decision.
Five risks to check in a food biotech business
Check customer demand, unit economics, operations, funding and regulatory planning before the next major commitment.
How ready is your food biotech business for its next stage?
A funding round does not establish commercial readiness. Examine customer commitments, economics, cash, growth and exit assumptions.
What belongs in a sensory data room?
A practical guide to taste, texture, consumer language and repeat-purchase evidence. See which claims each kind of evidence can support.
How much protein reaches the end of the recovery train?
A hypothetical five-stage mass balance connects recovery losses to the ingredient a customer can accept and the cost model behind it.