Look, there is still a big jump from the usual founder sharing partial vague optimistic statments (e.g., "our algorithm is great"), to the not so common direct lie (e.g., "we tested it with statistical significance, it works, here you can see these results on these real world inputs", while completely faking the results)