Re:They Write the Right Stuff (2021)
by cratermoon on Friday April 09, 2021 @10:55AM
(#26751312)
AI - really Machine Learning - is "just" massively parallel software for using linear algebra on high-dimensional matrices representing huge data sets.
Any programmer today can feed anything into the machine, but as the old saying goes, GIGO.
That every programmer today can build and train an ML model is one of the biggest advancements of ML engineering in the past 10 years.
But as you say it's GIGO, the difficulty today is to know what to feed it and to know what that means for the real life performance. There are no great tools for that yet.
> the difficulty today is to know what to feed it and to know what that means for the real life performance.
This has always been the difficulty.
Generalization is the fundamental problem in machine learning. Making easily available tools has led to an exponential growth in applications as more people play with it (many without understanding what they are doing or why), but predictably hasn't lead to an exponential growth in successful applications.
Then it seems like the author completely missed that « they write the right stuff » remains a complete pipe dream throughout software development.
> Looking at the result, it indeed seems like AI is going through what software went through 2-3 decades ago.
That would be because AI is a coat of paint on software, and software has not significantly moved from where it was back then. It anything, it’s gotten worse on everything the essay covered.