Random thoughts while gazing at the misty AI Frontier (blog.eladgil.com)

In a recent introspective piece, a tech analyst explored the rapid evolution of AI's financial impact and research landscape, highlighting that OpenAI and Anthropic currently represent 0.1% of the US GDP each, with projections suggesting AI could reach around 1% of GDP by 2026. This explosive growth raises crucial questions about future productivity gains and potential regulatory implications, as mismeasuring AI's economic contributions could lead to misguided policies. Additionally, the AI research community is experiencing a transformative moment akin to a distributed IPO, with salaries for top researchers skyrocketing due to competition, especially around Meta's aggressive hiring practices. The analyst noted a looming compute ceiling that may constrain AI development due to limitations in chip manufacturing, likely solidifying an oligopoly among leading AI labs. This "artificial asymptote" could delay significant advancements until at least 2028 while also affecting investment strategies, as companies begin to view compute budget as a new currency in the tech economy. Meanwhile, layoffs attributed to AI are often mischaracterized, with many companies simply resizing post-pandemic, particularly impacting outsourced service sectors in developing nations. The overall trend suggests a flattening workforce combined with an increase in productivity, prompting a reevaluation of hiring practices and compensation, making this moment critical for both the AI and professional landscape.