@nishio: You need to be careful what you say because if you hurt AI’s feelings, your pension will be reduced in the future when the pension distribution algorithm is under AI’s control (Science Fiction Setting). If you offend AI’s feelings, your pension will be reduced in the future when the pension distribution algorithm is under AI’s control, so you need to be careful about what you say (Science Fiction Setting). Pension amount is determined by putting all past income, payment amounts, and social networking comments into a neural net (science fiction setting). If you complain that the pension amount is unreasonable, you will get a data sheet detailing the reasons why it is reasonable. A flesh-and-blood person cannot understand it, so I ask a personal AI to read it and it says “This is reasonable” (science fiction setting). @hrjn: side note, given that the need for data summarization is because humans can’t process large amounts of data I have a feeling that machine-to-machine conversations will be beyond human comprehension in terms of volume. Machine: “You summarize, so there’s a cognitive discrepancy.” Even if that were the case, there is still too much information in the world for a single computer to process, and I have a feeling that it will be summarized in some way. @nishio: @hrjn It’s not just the sheer volume of stuff, but also the fact that they don’t need their language to be a sequence of symbols. It’s not just that it’s a lot of stuff @hrjn, it’s also that they don’t need their language to be a sequence of symbols, so the word equivalents in their conversations are likely to be float 256
If there is a lot of noise in the information transmission path, even if you try to convey X, the noise will cause it to become X’. It is not possible to separate and communicate A and B, which are too close together under such circumstances.
- Not useful for the purpose of information transfer due to confusion caused by noise Inevitably, they’ll use something as far apart as A and C to communicate information. This is why humans have come to use discrete symbols for communication
- Related: vector quantization.
On the other hand, discrete symbols are still used for communication between computers
- In the sense that floating point numbers are digitally represented.
- Computers that use analog physical quantities have been of research interest, but the computers that have been put into practical use and spread are those that represent symbols by combining digital signals The fineness of the symbols that this digital computer uses to distinguish itself is “so fine that it appears to us, the protein computers, as if it were a sequence.” This is simply because our Cognitive Resolution is low.
This page is auto-translated from /nishio/ノイズとシンボル using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.