kj98uo

Your argument is that the majority opinion is never psychopathic? Technically, if you were referring only to insanity, I might agree that an operational definition of insanity is anything that a majority of the population doesn't believe. But that doesn't mean the view is not psychopathic.

This definition of insanity might also not be correct. After all, was Galileo insane because most people (assuming they followed the church) didn't agree with his views?

That's like saying humans have been enslaving (among other things. this is just one example) other humans since the dawn of civilization and on a large scale, that didn't stop the growth of art, culture, science etc.

But our ethics progress and evolve. Similarly, treating other species as commodities for our consumption has been done for millennia but it doesn't mean we should keep doing it.

Well, how many other humans would you murder to save your grandma's life? Or your wife, your children, someone you love?

What about terminally ill humans who have < 1 year to live (assuming the particular testing results won't get skewed by their particular ailments)? How about prisoners sentenced for heinous crimes?

What about humans who are "not productive"? People who have been sentenced to more than 10 years?

You might draw the line at other species. Someone else might draw it differently to not do experiments on intelligent species like whales, dolphins and primates (who are used in lab experiments too). Or someone might include the categories I mentioned above in the list of test subjects.

The big second question is the effectiveness of transferring the results of tests from animals to humans. Coldly speaking, while there are experiments that yield useful results on animals, there are also extremely pointless experiments that are done for the sake of doing.

I won't go on and on but to your statement about " I am willing to sacrifice an insane amount of animal lives in order to save anyone human", would you do it to save Hitler (who is a proxy for someone who, maybe, you should think twice about saving)?

I agree with you about replacing the meat industry but we can do multiple things as a species at the same time. We can work on replacing the meat industry, working on alternate ways of testing drugs, on understanding animal communications and language (and we keep finding new indicators of intelligence) and on reducing our footprint as we get technologically and hopefully culturally more sophisticated.

I am still learning about Bayesian inference so this might be off-base but isn't the point to compute the full posterior distribution (or an approximation thereof) of the underlying parameters. Whether this is done in the context of a linear model or a deep neural network is a question of tractability.

The other distinction is between discriminative and generative models. In a discriminative model, the output/label is being predicted based on the input features: p(y|x, theta). For example, the probability of an image containing a dog, y based on pixels, x. Theta here refers to the parameters one needs to discover.

In a generative model, one instead models the distribution p(x|y, beta) i.e. given the label, say dog, predicting the joint distribution of all the images.

Neural networks with backproagation can be used for both discriminative and generative models. Bayesian methods can be applied to both discriminative and generative models to compute the full posterior distribution of the parameters, theta and beta.

Edit for clarity: The claim is that the choice of the model vs the choice of inferential methodology (Bayesian vs max likelihood for example) are orthogonal choices.

A neural network doing (discriminative) binary classification based on cross-entropy is maximizing likelihood instead of maximizing the posterior. Most Bayesian examples seem to specify a generative model (a Hidden Markov Model for example) and then infer the posterior. But there's nothing preventing one from using Bayesian methods with discriminative models (generalized linear models) or max likelihood with generative models.

for you. might still apply to others

I thought you work remotely based on your past comments??