Opening my Open Letters to OpenAI
When Elon says, it would be a good idea to hire those with exceptional achievements, he must not mean musicians and creative artists, for they are not going to contribute much to self driving; we have to be specific if we want progress.
Now that we found data, what are we gonna do with it?
One problem is that human thinking never really left the kindergarten level of “this is like that”. I know, because I never went to kindergarten.AI would have to implement all human limitations, the superstitions, the urban legends, the beliefs, religious paradoxes, for this is what we are, these are things that we co-exist with and live by. We must include not only the stars, but the horoscopes. Wanting to find correlations has its dangers. An example is when a chart of those died of being tangled In bed sheets and the cheese consumption chart seems eerily similar. A coincidence can become extremely damaging when algorithms start enforcing the selling in Corona (the company) based on the attention the word gets in the online media (thanks to a virus outbreak with the same name), for human lives may depend on fiscal losses (reputation, credibility included).
How do you store and restore values? In the future, when a picture shows the Amazon from space, a boat load of descriptions would need to be enclosed, saying that this is the River System in on the South American continent on Planet Earth, in our galaxy, and not a delivery center from Seattle.Who’s an ethical programmer, who’s values are in line with the majority and could be represented in decisions, and who is in the majority at all? Not everything is comparable, yet we want a yes or a no at the end of every question. How do you define “maybe” for an AI in a satisfactory manner? Superposition sounds fancy and futuristic, but can it answer the problem?We have dilemmas like all people are equal, but some are more equal than others. Can you weed out the “eye for an eye” logic? No, I don’t want to bring up the example of the unavoidable car crash where the AI has to de decide between sacrificing your life or five other people in the oncoming vehicle, but the problem is responsibility. It always will be. Playing God, playing Big Brother.
I called myself Melon Husk in my latest book, the Computer Aided Trading, so I should ask: who would be teaching humor to AI? It seems there is a need for a philosopher-comedian after all.I.e. “the erection of giant metal phalloses and their ejection / their penetration to space is the Space-seX”. Come on, don’t tell me you did not get this one!