Claude Can Dream? AI Data Center Backlash, GM Privacy Fallout & Canvas Breach

Anthropic says Claude is getting dramatically more capable — effectively unlimited context, multi-agent coordination, interactive self-correction, and even an experimental background learning process it calls "dreaming." At the same time, Anthropic admits Claude Code is growing far faster than expected, raising a bigger question: can the company control what it is creating? In today's Hashtag Trending, Jim Love also looks at a growing backlash against AI infrastructure as residents begin feeling the real costs. In Maryland, ratepayers could be on the hook for roughly US$2 billion in grid upgrades tied to AI data-centre growth happening outside their state. In Georgia, residents discovered water pressure problems that led to the discovery of major untracked data-centre water usage. Also in this episode: General Motors agrees to pay US$12.75 million in a California privacy settlement over allegations it collected and shared connected vehicle driver data without proper disclosure. Canvas restores services after a cyber incident reportedly linked to ShinyHunters, raising serious questions about education-sector data exposure and how much information may actually have been taken. If your car is collecting data, your utilities are paying for AI growth, and AI systems are beginning to "learn" between sessions… this may be one of the most consequential weeks in tech. Chapters 00:00 Headlines and Welcome 00:26 AI Data Centers Backlash 02:50 GM Driver Data Settlement 05:04 Canvas Breach Fallout 07:53 Claude Growth and Dreaming 10:43 Recursive Self Improvement Fears 12:43 Wrap Up and Support #AI #Anthropic #ClaudeAI #ArtificialIntelligence #DataCenters #Cybersecurity #Privacy #GeneralMotors #Canvas #ShinyHunters #TechNews #HashtagTrending #JimLove