CYBERSECURITY

I think there’s a misconception about how AI will break and change things. The Mythos hype has convinced people that AI is about to be so advanced that it can overcome our great cybersecurity defenses. It’ll creat new zero-days and use them to attack everything. And that it’ll be so smart that it can out-innovate our brilliant companies and genius human workers. So it’ll replace hundreds of thousands of companies and millions of workers by exceeding their quality. I think that’s wrong. The reality is that most companies are very poorly run. And their human workforces are extremely mediocre. The company vision isn’t clear. It’s hard to sell the product. And most of the workforce just tries to do as little as possible to earn their pay. And it’s been this way for decades. The only reason existing companies and workforces have survived this long is because most of their competition is just as bad. With cybersecurity, it’s not that everyone is so secure and there just aren’t any attackers smart enough to bypass their defenses. Most companies are actually easy for an a decent attacker to breach if they applied their focus and attention for a few days or weeks. So, people in general, including most people in AI, are thinking that human companies and workforces and cyber defenses are at 9.5/10. And Mythos is a 9.7 out of 10! In fact our current state of running everything, from companies to cybersecurity, is more like a 3/10. And AI for running a business, doing knowledge work, and attacking companies, is about to come in at a 5. Or a 6. And then a 7. But with the crucial difference that the 5 can be spread and scaled across the entire internet/world very quickly. A 9 attacker in 2019 could only hit very few companies. A level 9 attacker can now/soon build a level 7 AI Harness to smash millions of companies that have level 3 defenses. And it’s the same with human worker replacement. An AI that’s competency level 7 at doing general knowledge work isn’t all that impressive. But it’s revolutionary if it’s going up against level 5 human workers who are costing the company 20x what they’d pay for the AI. My takeaway and recommendation: Don’t think of the current state of things as a 9 that needs a level 10 AI to surpass it. As a species, we’re at a 3. In most things. It’s rather a miracle anything works at all. And here comes AI at a 5. 5 isn’t great. It’s weak and flawed too. But it CRUSHES a 3. And this is the worst version it’ll ever be. Maybe it’s a 6 in 2027. Or an 8 in 2029. Who knows what those numbers would ever mean. But humans took thousands of years to go from 1 to 2, and then centuries to go from 2 to 3. So humans aren’t going suddenly jump to a 6 to counter AIs at a 5. This is why we should take AI a whole lot more seriously, in general. It’s not competing with the best of us humans, or the best us humans can offer. It’s competing with what we have been able to piece together so far: which is a hobbled-together and fragile mess. It’s not 9.5 vs 9.7. It’s 3 vs 5/6/7.

→ View original post on X — @scobleizer, 2026-04-11 20:39 UTC