@kazani

The AI Industry Is Lying To You The AI industry's growth is based on a "sense of inevitability" rather than tangible progress, with many announced data center projects being speculative and not under active development. Only 33% of announced US data center capacity is under active development, and the majority of committed power is for utilities only responsible for delivery, not generation, posing a significant problem for power-hungry AI servers. Actual data center absorption (IT load of data centers turned on and in operation) in North America for 2025 is estimated to be around 3GW, a figure described as a "fucking disaster" compared to the projected demand. It takes approximately 6 months to install a quarter's worth of NVIDIA's GPU sales, indicating a significant bottleneck in data center construction compared to GPU production. Globally, only about 5GW of data center capacity is actually under construction, with multi-gigawatt projects expected to take 2 to 4 years or more to complete, and capex growth is projected to slow. The AI industry, including major tech companies, is accused of actively misleading the media and public about the pace of data center construction and AI's capabilities. Hyperscalers are forcing employees to use AI tools, tying performance to "token burn" and encouraging non-technical staff to "vibe-code," leading to potential security breaches and a decline in software quality. The widespread use of LLMs in coding is creating massive, unmanageable codebases, increasing operating costs, and disincentivizing software engineers from learning fundamental skills, potentially rotting the tech industry's software stack from within. The arrest of Supermicro co-founder Wally Liaw for allegedly selling hundreds of millions of dollars of NVIDIA GPUs to China, along with other suspicious dealings involving GPU resellers, raises questions about NVIDIA's oversight and knowledge. The entire AI industry is built on "lies," with inflated data center development figures, GPUs waiting for installation, lack of profitability in AI compute rental, and a decline in genuine software engineering skills, leading to a potential "corporate failure." https://www.wheresyoured.at/the-ai-industry-is-lying-to-you/

0 reply

0 recast

1 reaction