August 11, 2026
Flagged by the bots
AI Is Solving CTF Challenges in Minutes
Hackers got benched as AI speed-ran the contest and the comments went feral
TLDR: An automated AI system reportedly solved every challenge at a major hacker training contest in minutes, raising big questions about whether these events still test human skill. In the comments, people split between “this is old news,” “how are they even doing this,” and “stop pretending this won’t upend jobs.”
The big shocker isn’t just that artificial intelligence helped at a hacker puzzle contest — it’s that, according to the story, it basically took the wheel, solved all 52 challenges in minutes, and left humans eating dust. At BSidesSF 2026, the winning team reportedly used an automated system that watched for new puzzles, launched multiple AI tools at once, and crushed problems faster than any person could. For longtime players, that landed like a jump-scare. One competitor said he went from solo top-five material last year to feeling like he’d sink to 75th without AI. Ouch.
But the comment section? That’s where the real fireworks started. Some readers were unimpressed, with one instantly sneering, “Reads like AI writing again. Can you just post the prompt?” Another basically said, old news, keep up — pointing to the public repo and arguing this stopped being surprising months ago. Then came the existential dread squad: one commenter pushed back hard on the article’s sunny “humans should just focus on what AI can’t do” message, calling that pure cope and saying the real implications for security jobs, software jobs, and knowledge work in general are anything but clear.
There was confusion too: if these systems can tear through security puzzles, why do ordinary people still hit walls trying to get AI to do much simpler code-analysis tasks? That question gave the whole thread a spicy vibe of disbelief, doom, and gallows humor. In other words: AI didn’t just win the game — it started a fight about whether the game, and maybe the career path behind it, still means the same thing at all.
Key Points
- •The article says an autonomous AI system won the BSidesSF 2026 CTF by solving all 52 challenges, with many solved within minutes.
- •The winning team reportedly open-sourced a system that polled the CTF platform, ran parallel AI agents in Docker containers, and used a coordinator model to share discoveries.
- •The article contrasts BSidesSF 2026 with the prior year, when participants mainly used ChatGPT as an assistant instead of full automation.
- •It argues that standard jeopardy-style CTFs may no longer measure uniquely human cybersecurity skill if AI can solve them faster and more reliably.
- •The article says AI still struggles more with real-world security work requiring context, judgment, communication, and coordination, and cites NYU research on the importance of context-setting.