binary fun

[Impenitent]

https://www.zmescience.com/science/news ... aking-out/

AI has apparently solved several advanced math problems, but somehow explains the findings incoherently...

I would expect no less from a glorified abacus...

how long until the claim that computers have supposedly surpassed human thought takes hold?

when the metal gods assume control of society, will you blindly obey?

-Imp

[MikeNovack]

https://www.zmescience.com/science/news ... aking-out/

AI has apparently solved several advanced math problems, but somehow explains the findings incoherently...

Part of a broader matter.

For example, the typical "disassembler" can produce from the machine language of a program the source code in an assembler language, but only in the sense "an assembler could assemble it". If doing this for a PRACTICAL (human readable) source code << that can be edited/maintained) much hard work remains* >>

When these neural net AI's trained not included in the training "and present in coherent form". Early days yet.

* I have done this in my working days dealing with a "lost source code" problem. It is a PITA, slow going. Sort of like old fashioned cryptanalysis, I used strips pf paper I could slide against each other back and forth to discover common addresses (to which a human writer of an assembler program would have assigned a label and then made addresses local to that label relative to that label). Lots of tedious work, and the program I was recovering source for only a few hundred bytes long.

[Impenitent]

I never wrote programs in direct assembly code... well I did create graphic shape tables in basic on an Apple II in the early 1980's but I never got past basic, pascal and cobol - had a look at fortran but never did anything with it...

broader... the computer processes 1s and 0s... ascii codes can create certain characters in certain alphabets

programming/interpreting simple "speech" in mathematical language seems difficult enough, let alone Arabic, Russian or Greek...

-Imp