|
|
|
|
|
|
by streetcat1
740 days ago
|
|
|
The article fails to describe the core agent algorithm. I.e. how does the agent decide what to code ? If the agent did coded something, how does the agent decide that the code is correct ? (Ecepsically if the core model is based on auto complete and lacks any logic). Like if the agent decides to search for a file X on task Y why file X and not file Z ? |
|
|