The Most Expensive Spell Checker Ever Built
Companies are spending $54 million on average to deploy AI tools, yet more than half of employees abandon them and 37 percent don't use them at all. A new WalkMe report and developer survey data reveal a growing trust gap: AI is generating more code than ever, but 96 percent of engineers don't fully trust it, and reviewing AI output now takes more effort than reviewing human code.
Read full story →Running Out of Road: The Political Reckoning With AI's Power Appetite
AI's energy demands are crossing from infrastructure problem to political liability. The US Energy Secretary called growing public opposition a "very real risk," UK MPs opened an inquiry into alternative chip designs, and mainstream analysts are warning that the laissez-faire approach to AI power consumption is no longer politically tenable on either side of the Atlantic.
Read full story →$25 Billion and 5 Gigawatts: Amazon's All-In Bet on Anthropic
Amazon committed $25 billion to Anthropic on April 21, one of the largest AI deals in history. The investment includes immediate capital and access to 5 gigawatts of Trainium chips for training Claude. The deal cements Anthropic as AWS's AI anchor and deepens the cloud provider consolidation reshaping who controls the foundations of enterprise AI.
Read full story →The Adoption Gap Is Not a Communications Problem
More than half of workers are abandoning the AI tools their companies spent $54 million to deploy. The standard explanation is change management failure. Peter Harrison argues the tools genuinely don't work well enough for most workers doing most tasks, that companies deploying them know this but can't stop, and that the moment the tools actually work is when the real reckoning begins.
Read opinion →