As for animal experiments, they are not nearly as difficult to get going as people think, SOPs for mice take a month on average to get approved in my experience. The vast majority of my dislike towards animal testing lies in the realm of cosmetics, not animals, and I'm sure most people would agree. Even so, the trend in research now is to move away from traditional animal models to organ-on-chips, simulations, and other human substitute models.
Finally, Biology is SLOW. You haven't seen a "unicorn" because an MVP for a biotech company in therapeutics is 10 years away from when it starts needing funding, in a lot of cases. There are very few options for series A-B funds to fill the gap between seed and revenue stages for biotech startups, which is why licensing and partnerships between researchers and industry are the predominant options seen today, alongside the traditional NIH/DARPA/NSF funds.
Maybe I'm biased because of my location (huge Biotech hub) and affiliations, but anytime I read complaints about biotech and medicine from software devs it reads like ignorance from people who are used to "moving fast and breaking things".