AI bubble ready to pop?

In a recent podcast, Marc Andreessen and Ben Horowitz, founders of the venture capital firm Andreessen Horowitz, discussed the slowing pace of advancements in generative AI. They noted, "We're increasing GPUs, but we're not getting the intelligence improvements at all." Their firm has invested over $2.25 billion in AI, making them one of the industry's most aggressive investors. While factors like GPU availability, power, and cooling have been cited as reasons for this slowdown, there may be deeper issues at play. Data availability challenges One significant issue is data availability. After existing models have scraped the internet for training data, there is little left to gather. Andreessen highlighted that OpenAI has established a 'Human Data Team' dedicated to sourcing data and refining model training and evaluation. This team comprises human experts who enhance AI responses, which is ironic considering that generative AI was initially thought to threaten jobs by replacing human workers. Understanding Gartner's Hype Cycle Gartner's Hype Cycle for Artificial Intelligence indicates that the technology has surpassed its peak of inflated expectations. Gartner outlines five phases in this cycle: Innovation trigger: A breakthrough in technology. Pe

This article was originally published on December 01, 2024.