←back to thread

146 points MaysonL | 2 comments | | HN request time: 0.448s | source
Show context
geeal ◴[] No.43960034[source]
Advances in computing sciences were not accomplished by investment strategies and nondisclosure agreements. It was accomplished by dedicated professional and academics with the highest levels of integrity and transparency. So your post comes at a critical time where parties are jumping to the bandwagon of AI.
replies(2): >>43960139 #>>43960229 #
watwut ◴[] No.43960229[source]
Where do you have integrity and transparency from? They were of all kinds.
replies(1): >>43960310 #
ktallett ◴[] No.43960310[source]
Transparency is the case, up to a point in research. Many engineering papers I find, have the told you the building blocks of what is required, but they try to avoid stating the secret sauce that binds them together.
replies(1): >>43960756 #
1. mschuster91 ◴[] No.43960756[source]
Yeah because otherwise you'll have some cloner in China reproducing it the very next day.

Transparency is only working if everyone plays by the same rules, particularly when it comes to patents, and there are a few players who have been getting away with openly sharting on the rules for decades now.

replies(1): >>43962892 #
2. ktallett ◴[] No.43962892[source]
I'm not talking about consumer products, I'm talking about academic research. Most research has no patent attached nor any contents that is patent worthy. Most methods in most papers are based on the very basic building blocks coming from ones fine motor skills.

The fact these methods are omitted has numerous problems. One key issue this provides is reproducibility. All science has to be reproducible and a building block for future science. Omitting key details makes this unfit for purpose.

Secondly, it also means research can not be checked for accuracy and truth, which I do sometimes wonder if thats on purpose. Perhaps they are only presenting the most successful attempt, but not the average situation.

Lastly, this fully goes against the spirit of academic research. I want anyone out there to develop either a brand new usage for my work or adapt and improve my work, either building on it, or finding a new way of doing it that's easier, or more repeatable/reliable. It saves me having to do it, and as I am still working in that field, I will benefit from it.