Message from @Sophie
Discord ID: 685504400273572058
Ok, so let’s say you start with a program
And you randomly change that program
Even with machine learning, the program needs a goal. It needs instructions
Yes, a goal
That’s what selective pressure does
It gives it a goal
The goal is to exist
But it has no goal
The goal is to reproduce and continue the species, ultimately
Because anything that fails to exist, doesn’t exist
The cells have no algorithm
Yes they do
They respond to stimuli
That’s an algorithm
You can think of a cell as a series of Boolean operators
*operations
I’m going to find something to show you how machine learning works
Right, I also use machine learning in my job
Let’s focus on evolution
Are you sure?
Uh huh
Programs cells
Uh, ok
Which receive random mutations
Which can be positive, negative, or neutral
Ha, ok
Depending on what pressures are selecting on them
Uh huh
Sure
Negative mutations will underperform, positive mutations will overperform
Thus positive mutations will tend to be maintained in a population, and negative mutations will tend to be removed from a population
Uh, ok. So we are assuming they already have basic instructions
Then, through many generations and separation between populations, speciation can occur
What “basic instructions” are you talking about
Cells that survive have the “instructions” to survive
Well first you have selective pressure. What selective pressure?
Then you’d have the basic genes in the cells
With random instructions
It depends on many factors. There’s always selective pressure to successfully breed, and successfully eat
I should say multiply, not breed