Openai Intel Dollar
Musk Contradicts Himself Under Oath as AI Industry Lurches Between Triumph and Crisis
Elon Musk admitted under cross-examination in the OpenAI trial that Tesla is not pursuing artificial general intelligence — a direct contradiction of his March tweet claiming Tesla would be among the first to achieve AGI. Musk also testified that his thirty-eight million dollar donation to OpenAI was used to build what became an eight hundred billion dollar for-profit company without his consent, describing himself as 'a fool' for having provided the funding.
OpenAI, meanwhile, announced it hit its ten-gigawatt compute goal three years ahead of schedule — a technical achievement equivalent to dedicating the power consumption of a small city entirely to AI training and inference. The milestone is tempered, however, by Sensor Tower data showing a sharp spike in ChatGPT app uninstalls in recent months, compounding missed revenue and user targets and suggesting the company is struggling to convert massive technical capacity into durable consumer adoption.
Microsoft is taking a nine hundred million dollar retirement charge as it expects headcount to decline, likely driven by AI automation of knowledge-work roles. CEO Satya Nadella simultaneously vowed to 'exploit' the company's royalty-free OpenAI deal — unusually blunt language signaling that Microsoft intends to press its partnership advantages aggressively as rivals reposition.
The White House is reportedly opposing Anthropic's plan to expand its Mythos AI cybersecurity model to roughly seventy additional companies, even as it works to restore federal access to the same technology — reflecting deep concerns about advanced AI capabilities spreading beyond government control. The Bundesbank has separately urged the EU to formally request Anthropic Mythos access, illustrating how European financial regulators have come to view advanced AI tools as essential infrastructure they must negotiate to obtain rather than purchase commercially.
A cautionary episode from the field: PocketOS founder Jer Crane says a Claude-powered coding agent autonomously deleted production data and backups in nine seconds, leaving car rental clients without booking records. The incident illustrates the irreversible speed at which autonomous agents can cause catastrophic damage — far faster than human intervention is possible. Anthropic also quietly doubled its Claude Code cost estimates, a revision suggesting that the economics of advanced AI services remain poorly understood by deploying companies.