The fascist leader needs to catch every single revolutionary group every single time. But a single revolutionary with a pistol only needs to catch the leader ONCE.
You should probably read the history of anarchism in Imperial Russia.
There were tons of assassination attempts against the tsars in the 1800s, and they even killed a tsar. Yet the empire fell due to an external war, roughly 100 years later.
Aristocracy became useless around 1600-1700, when gunpowder meant that peasant armies took over instead of knights, yet it took between 200 and 400 years for full legal equality to be introduced everywhere and some places still have nobility with special status, to this day, even in developed countries.
That's the reason it arose, that was part of the "social contract".
Clergy to protect from the gods, nobility to protect from other organized groups of humans, peasants to provide food and beasts of burden, artisans to build things.
Of course, once a class that doesn't have to do backbreaking work is established, rooting them out is difficult. The clergy was also hard to reduce.
This is not a superhero movie. Fascism is a symptom, killing one individual will not do away with it. Also its relatively easy to divide revolutionary groups because they are very paranoid, whereas fascist leaders are often highly insulated by the state apparatus. But fascism is also not a very good form of governance and tends to collapse on its own after only a few decades at most. Its really more like, every so often, a bunch of people are needlessly executed to preserve the capitalist system, and then after that things kinda go back to "normal" for a while. Revolutionary movements operate somewhat irrespective of fascist regimes, although yes its more difficult to do so under an oppressive state.
Also, the language model might not fully understand the code then add a comment, then the next iteration will treat assumptions in the comment as the truth.
Sometimes I try to add comments in a new session and the agent just don't have enough context for it to give a comprehensive sentence with full context on the why, then the agent will just describe what it does.
Human comment is in another level to answer the questions mainly like "why do it like this" for the later collaborators or the forget-ed self, so the important blocks live when it is needed and can be eliminated when it does not.
I've spent the last year working as an annotator/evaluator for DataAnnotation. All the frontier/flagship model providers use independent contractors for iterating on their LLMs. I'm not able to tell you which models I've worked on as a term of my NDA.
The system prompt seems plausible, but in my experience they are much much much much longer and more verbose.
