story
Also the blind person wouldn't confidently answer. A simple "the objects blur together" would be a good answer. I had ChatGPT telling me 5 different answers back to back above.
If the legally blind person never had had good vision or corrective instruments, had never been told that their vision is compromised and had no other avenue (like touch) to disambiguate and learn, then they would tell you the same thing ChatGPT told you. "The objects blur together" implies that there is already an understanding of the objects being separate present.
You can even see this in yourself. If you did not get an education in physics and were asked to describe of how many things a steel cube is made up, you wouldn't answer that you can't tell. You would just say one, because you don't even know that atoms are a thing.
ChatGPT can't count, the problem is the tokenizer.
I do find it funny we're trying to chat with an AI that is "equivalent to a legally blind person with no correction"
> You would just say one, because you don't even know that atoms are a thing.
My point also. I wouldnt start guessing "10" and then "11" and then "12" when asked to double check only to capitulate when told the correct answer.
First of all, it obviously changes everything. A shortsighted person requires prescription glasses, someone that is fundamentally unable to count is incurable from our perspective. LLMs could do all of these things if we either solve tokenization or simply adapt the tokenizer to relevant tasks. This is already being done for program code, it's just that aside from gotcha arguments, nobody really cares about letter counting that much.
Secondly, the analogy was meant to convey that the intelligence of a system is not at all related to the problems at its interface. No one would say that legally blind people are less insightful or intelligent, they just require you to transform input into representations accounting for their interface problems.
Thirdly, as I thought was obvious, the tokenizer is not a uniform blur. For example, a word like "count" could be tokenized as "c|ount" or " coun|t" (note the space) or ". count" depending on the surrounding context. Each of these versions will have tokens of different lengths, and associated different letter counts. If you've been told that the cube had 10, 11 or 12 trillion constituent parts by various people depending on the random circumstances you've talked to them in, then you would absolutely start guessing through the common answers you've been given.
