

Definitely. It might require significantly more input to gain the same certainty, but all it’s doing is reweighting the possible next tokens before choosing, and code output is still just token output. The rigid syntaxes probably means that the next token probabilities are much more sharply divided (maybe a random sentence the top 1 choice is just 40%, top 3 are 80%, but for a line of code, the top 1 choice might be 90% probability and top 3 hit 99%)










If someone said “do you want to interview my agents first”, I think the interview would be over. Well maybe the first time I’d be tempted to see if I could tokenmaxx someone else’s agent. After that it would be too much work for the enjiyment. Maybe I should automate that workflow… Oh no they got me!