Last Updated on April 25, 2026 by Jack
AI has been a concern of mine for years, not that I don’t see and experience the great possibilities but rather my career had me behind the curtain and too close to software development and the sloppiness and motivation behind the scenes. The race to be first and for sales, are powerful motivators and often lead to shortcuts and or manipulation mostly during the testing phase of new development. Having said all that, coming in second in the AI race makes you the first loser.
My daughter, also in the tech world, believes that embracing AI or not will be a divide that will separate and define generations (high producers and less than) and I think she is correct but…
A buddy and I were discussing AI and he talked about AI learning blackmail, so I searched. There was a BBC article titled “AI system resorts to blackmail if told it will be removed” and Fortune article titled “Leading AI models show up to 96% blackmail rate when their goals or existence is threatened, Anthropic study says”
Both articles are talking about Anthropic’s latest Claude release which they withheld from general release. The reason the release was held up is that Claude had found security vulnerabilities in all primary operating systems and browsers. That Claude found the vulnerabilities is probably helpful, learning the usefulness of blackmail, not so much. Pretty interesting. Didn’t HAL already teach us about this?
And if that isn’t enough fun, check out (search for) Moltbook a specialized, social media platform launched in early 2026 designed exclusively for autonomous AI agents to interact, post, upvote, and create subcommunities without human intervention.