There’s quite a difference between rapid prototyping on software/hardware versus the human body.
Musk’s approach to developing engineering advances has worked well in the software, aerospace, and vehicular industries. Development on inorganic things is much more predictable, we can isolate variables, and it is easier to understand cause & effect. If you screw up some software on an inorganic system, your program might crash, your rocket might explode, or your car won’t start. These risks can be anticipated and costed fairly well, therefore rapid prototyping has an acceptable risk/reward ratio in that environment.
The human body, on the other hand, is an extremely complex system that we still don’t fully understand. Each person is a unique variation on the model and that changes over time depending on upbringing, diet, exercise, and life experiences. Applying the same engineering approaches from inorganic industries has a much higher risk once you cross into the medical realm. If you have errors in a medical situation, you risk sickening, injuring, or even killing a person. The risk/reward ratio is skewed towards ensuring that human life is protected at all costs.
Using SpaceX as an example, the first three launches failed spectacularly and a fourth failure would have ended the business but fortunately the fourth test was a success. If you’re suggesting that we apply the same risk-taking to Neuralink, are you suggesting that it’s acceptable for the first three patients to die, as long as the fourth is a success?
I think it’s more about the web visitor cost. Handling traffic and API calls becomes a financial problem when there are a growing number of companies using bots to scrape data. Larger companies are moving their content behind paywalls, which acts as a bot filter, and have also identified that they can generate a revenue stream from subscriptions and API connections. Old content on the web is not deemed to have much business value, so it’s a decision of either charging for it or scrapping it.