You're tuned in to Tech Beat, let's get into it.
Confidence in AI security tools is taking a serious hit. A new report from Cobalt finds that only nine percent of cybersecurity professionals trust fully automated AI to find vulnerabilities — down from twenty-nine percent just last year. Nearly four in five said their tools missed critical flaws entirely, and the time to resolve AI-specific issues nearly doubled to thirty-six days.
That skepticism connects to a broader tension playing out in software development. Engineering teams are wrestling with what happens when AI agents generate massive pull requests that human reviewers simply cannot keep pace with. The concern isn't just speed — it's the quiet accumulation of technical debt that gets baked in before anyone catches it, reshaping how software actually gets built.
And zooming out further, a piece in Harper's is asking harder questions about who bears the cost of the data center boom powering all of this. The infrastructure behind artificial intelligence doesn't land evenly — it lands somewhere, in specific communities, with specific trade-offs around land, water, and power that the industry rarely leads with.
Three stories, one thread: the gap between what AI promises and what it actually delivers is becoming harder to paper over. Keep surfing. Tech Beat out.
