We're Getting the Wrong Message from Mythos
The author argues that Mythos, though not trained for cybersecurity, outperforms experts by chaining vulnerabilities and excels across all knowledge work tasks. Companies will soon replace human workers with cheaper, more productive AI, forcing a major shift in how we work and demanding a rethink of our future roles.
- Mythos can chain low/medium vulnerabilities into critical exploits, a feat fewer than 1% of human pentesters achieve, despite not being purpose-built for cybersecurity.
- Its general knowledge-work abilities (emails, analysis, reports) suggest cybersecurity skill is just a side effect of broader competence.
- Open-source models nearing Mythos's capability at under $1,000 will make AI vastly cheaper than a $84,000/year employee while producing 10-1000x more output.
- This cost gap will trigger widespread white-collar job displacement, demanding urgent retraining, policy, and safety-net planning even as it opens space for more meaningful, non-corporate work.
Mythos isn’t a niche cybersecurity tool. It’s a general-purpose AI that happens to excel at security tasks. The model can link together low and medium vulnerabilities into critical exploits—something fewer than 1% of pentesters ever manage. If it can handle that level of complexity, it can also draft emails, analyze data, write reports and tackle virtually any knowledge job.
In months, open-source models will approach Mythos’s capabilities and cost under $1,000. Companies then face a simple choice: pay Chris from Idaho $84,000 plus benefits for 40 hours of average work, or spend a few hundred dollars on an AI that delivers ten to a thousand times more output, nonstop. That price gap will upend hiring, push roles out of reach for most humans and force businesses to rethink every white-collar function.
This shift will be painful at first. Jobs will vanish before new ones appear. But it also creates a chance to move beyond repetitive corporate grind. People could focus on building community, solving human problems and creating value in ways a machine can’t. We need plans now—retraining programs, policy changes, safety nets—to manage the shake-up and steer toward more meaningful work.