Hacker News Reader

Comment by selfmodruntime | Hacker News Reader

No. Science has all along been testing if something works and then finding out why. Often the „why“ is orders of magnitudes harder to find out than the „if“.

For example, we can observe how the brain functions in some instances, but why it works like does will elude us perhaps forever. Highly complex systems like the inner biomechanical workings of our brains are essentially a black box.

View on HN

"Why" questions in general are pretty difficult question to answer in a satisfactory way.

[1] https://www.youtube.com/watch?v=36GT2zI8lVA

I think “how” is more appropriate in the GP’s comment.

We can understand a thousand instances "how", but it will always be "why" that ties them together.

Like software incompatibility: every program is the story of "how", written explicitly, and translated into instructions. The only way to fit them together is to implement "why". If we could write both stories together, we could factor out incompatibility.

I knew it was the Feynman video!

Feynman doesn't disappoint. Although, I would have liked for him to make clear how quickly we (including physicists like him) get to "we have no idea". Sort of the overall arc of his answer made it seem that, well we know why but it would be impossible to explain it to you because it is incomparable to anything else you might know. But the reality is, we quickly arrive at the four fundamental forces (the we know of) for which we have no explanation, no further reducibility (that we know of). We know they exist because we can observe them or work them out mathematically. But we have to just accept them on their face as they are in order for the rest of the universe to make sense.

I'm not sure that invalidates GP's point. Your comment is basically just repeating his premise. If we don't understand the why, we don't understand the how. We can spend years testing that "it works," but we're only testing for a specific outcome and checking for specific risks. Without understanding the how or why, we have little way of knowing what we don't know.

We start with what then much later to how which finally leads us to why.

Ground rhino horn and tiger penis don't have measurable results, sure it's not ideal we might not understand how aspirin does what it does, but at least we know that it does something. That's not even in the same ballpark as Chinese medicine so comparing them just for that reason is disingenuous.

We don't test medicine for marketing, we test them for results. If homeopathy and Chinese medicine would teach us one thing it's that testing only for marketing would be the greatest waste of time and money ever.

Indeed, chinese medicine and homeopathy show us that spending a dime on testing that you could be spending on marketing instead is always a waste

Not only that, but it's based on a very certain, but false, idea of why it "works".

For an example in CS, we currently use machine-learning algorithms that work very well in practice. However, why they work so well is still unclear in many cases (although we are making progress).

I find it fascinating that describing an automated learning process is so much easier than manually teaching. Once trained, looking inside a model to gather information about how it comes up with an answer to a question is somewhat similar to doing exactly that with a real brain.

Not always in pharmacology. There are many cases in drug development where people decide to target a particular biological process and tailor a molecule to do that.

Reply on news.ycombinator.com