3 ms·
I agree that cross examination helps. I agree lawlessness is bad. However, what you you are proposing is pseudo-random solutions instead of lawlessness, which d
by sqeaky 10y ago
I agree that cross examination helps. I agree lawlessness is bad. However, what you you are proposing is pseudo-random solutions instead of lawlessness, which doesn't seem much better. Picking any subset of trials and saying it is okay for them to be wrong could mean setting wrong precedent for years to come. It can also mean bad outcomes and suffering for many and disproportionately poor people.
If the jurors don't understand the fundamentals of what they are deciding how will they interpret cross-examinations? Certainly not on the merit of the argument, because they simply can't.
If it just comes down to who can sound more confident while making their side of the argument then justice goes to whoever can afford the best lawyers. Which largely seems to match the system we have.
It is not a very good but it is better than lawlessness, but only just so.
Though we get more nuanced each time, I fear our debate has started running in circles. Thank you for making reconsidering some of my opinions.
- Retric 10y agoI mean if a group of experts might chose A 55% of the time and B 45% of the time then a 50/50 random outcome is not a real issue. A few places have elections where in case a tie they do a coin flip. The important thing is to remove bias not get extreme accuracy. As to our court cases being biased for the wealthy that's a little more systemic than just highly technical cases. Even in clear cut cases having a competent legal team can make a massive difference, IMO the overall system was designed that way. The problem is how to change things. If doctors made up the jurors in malpractice cases you would see much stronger bias. I would much rather have random people than wall street bankers judging wall street bankers or even worse FBI agents judging other FBI agents. That said, I am open to the idea of civil suits allowing for more restrictive juries if both sides agree. If nothing else it may allow for shorter trials. Which is something I only considered though this conversation. So thanks.
- sqeaky 10y agoClearly we aren't running in circles! I agree that those coinflip cases exist, but if cases are so close a coinflips (or elections) can decide them then why do sums like $500 million depend on them as in the article? A system which promotes charisma instead of factual accuracy doesn/t just result in muddy grey areas where things were already close. Things like climate change, which pretty much all experts agree on, are grey areas to many people. Charisma is how we get trump instead of any other politician. Every year that passes I am less in favor of the idea of democracy than I was the year before. The aggregate of a great many untrained people is generally stupid. Its just better than any system reliant on experts we have devised so far. This seems to be caused by the corruption of our method of choosing experts. Once we a get a reliable system that empowers experts I will be behind wholeheartedly.
- seanp2k2 10y agoThe problem in your last sentence is the trick, one which humanity has yet to solve in any sort of sustainable way. Maybe strong AI can finally save us from ourselves.
- sqeaky 10y agoEven that that pushes the trust problem one step further. Presuming making a Benevolent and Obedient AI leader is possible, how do we trust the people making it? If western industry creates the super AI and somehow Google, MS, Facebook, Amazon, Apple and western society in general trust it why should China and Baidu?