Elevator Pitch

• Frontier LLMs have automated enough of open online CTF solving that the public scoreboard no longer measures human security skill or growth, collapsing the format’s “ladder” and incentives.

Key Takeaways

• Medium (and increasingly hard) CTF challenges have become “agent-solvable,” shifting competition toward AI orchestration and spending rather than human problem-solving.
• This breaks the learning feedback loop: beginners are pushed to use AI early, undermining “active struggle” and making visible progress demotivating.
• Organisers can’t realistically enforce bans or design around LLMs without making challenges worse for humans, so the open online format is deteriorating.

Most Memorable Quotes

• “the old game is not there anymore.”
• “That makes open CTFs pay-to-win.”
• “The format is dead.”