# Everybody's Lost Their Minds

> A security engineer's frustrated take on how AI hype has broken engineering culture and made us all worse at our jobs.
- Title: Everybody's Lost Their Minds · Netmeister | Shortify
- Summary: A security engineer's frustrated take on how AI hype has broken engineering culture and made us all worse at our jobs. AI is replacing understanding with…
- Keywords: artificial intelligence, ai hype, engineering culture, cybersecurity, vulnerability research, patching, ai ethics, open source security, environmental impact, skill degradation, anthropic, openai
- Source: Netmeister — https://netmeister.org/blog/everybodys-lost-their-minds.html
- Read time: 2 min
- Topics: artificial intelligence, ai hype, engineering culture, cybersecurity, vulnerability research, patching, ai ethics, open source security, environmental impact, skill degradation, anthropic, openai
## The state of tech discourse
Non-engineers pitch world-changing AI solutions from their home labs. Emails read like LinkedIn influencer posts. Online articles have converged into mediocrity.
## AI work has killed joy
Spending 75% of time dealing with AI daily has robbed most enjoyment from work. It feels like everyone else changed while you stayed the same.
## AI companies commit crimes
Major AI firms one-up each other while exploiting intellectual property, concentrating power, and enabling child sexual abuse material generation as a logged-in feature.
## Military targeting and ethics
AI models support military target selection, including elementary schools. Companies brush aside ethics concerns while spending enormous amounts on tokens following a casino model.

> ethics aside...
## Vulnerability research FOMO
Anthropic and OpenAI created time-limited projects like Glasswing and Daybreak. Dozens of highly-paid security engineers diverted all priorities, costing millions per organization.
## Thousands of vulnerabilities found
Despite discovering thousands of new vulnerabilities, security hasn't improved. Finding bugs was never the bottleneck. Getting packages actually updated is.
## The real security work
Actual security needs asset inventories, automated updates, regular reboots, and attack surface enumeration. These basics are what nobody does well.
## Resources are zero-sum
Every engineer busy in their AI silo doesn't help projects needing cross-functional collaboration. Those resources could have overhauled patching infrastructure instead.
## Fake regulation theater
AI companies call for regulation while claiming their products might kill all humans. If they believed that, they could simply stop building. Greed and competition prevent it.
## Environmental destruction now
AI's real harm is happening today. Data centers waste water, pollute air, and burn fossil fuels. Governments lift environmental regulations for these companies.

> the world looks different now
## AI dulls understanding
AI achieves superhuman intelligence by making humans dumber. It replaces understanding with dependency as you actively de-skill yourself.
## The debugging problem
AI finds bugs, generates patches, and reviews code. The more AI in the loop, the less anyone understands the codebase. Debugging opaque systems becomes impossible.

> the human in the loop often is nothing more than a rubber stamp
## Getting off the ride
Complex systems fail in complex ways. Debugging is harder than writing code. Large distributed systems with opaque components will be impossible to debug.
## Key takeaway

AI is replacing understanding with dependency, making complex systems harder to debug and diverting resources from actual security fundamentals like patching.