Description Foundation Models and the AI Boom: Narrative Momentum vs. Architectural ResilienceCivilization Physics — Economics Series This paper analyzes the structural mismatch driving the current AI boom: narrative momentum (hype, investor FOMO, speculative valuations) vastly outpacing the architectural resilience of foundation model technology. While capital inflows and public enthusiasm mirror the late-1990s dot-com mania, today’s AI lacks the robust underlying standards, open protocols, and stable infrastructure that allowed the internet to crash, recover, and eventually dominate the world. The report highlights four systemic risks: 1. Sustainability & Cost Explosion Foundational models rely on massive, energy-intensive training pipelines whose costs far exceed demonstrated returns. With 95% of enterprise AI pilots failing to show ROI, the industry’s infrastructure spend risks becoming a bubble of stranded assets with no equivalent of Moore’s Law or open protocols to guarantee long-term efficiency. 2. Missing Architectural Foundations Unlike the internet—built on TCP/IP, HTTP, HTML, and universally interoperable standards—the AI ecosystem is fragmented across proprietary models, closed APIs, opaque training pipelines, and brittle architectures. No shared “AI protocol layer” exists, weakening the system’s long-term survivability. 3. Competitive Defensibility Breakdown Open-source models, cheap compute, and low switching costs are rapidly commoditizing foundational AI capabilities. Without strong architectural moats, most AI firms risk becoming interchangeable utilities whose valuations depend solely on narrative, not structure. 4. Reliability & Collapse Risk Hallucinations, static knowledge, brittle reasoning, and information inbreeding (models learning from synthetic data) reveal a lack of intrinsic stability. Without negative-entropy grounding—fresh human data, oversight, and structural corrections—foundation models risk long-term performance decay and eventual model collapse. Using the Civilization Physics framework, the paper contrasts the AI boom with previous technological revolutions and concludes that the next phase of AI must prioritize architectural breakthroughs: open standards, efficiency leaps, reliability frameworks, clean-data ecosystems, and governance structures that shift the industry from speculative growth to durable infrastructure. Keywords: AI Boom · Foundation Models · Narrative Momentum · Architectural Resilience · Model Collapse · Information Inbreeding · Open Standards · AI Infrastructure · Civilization Physics · AI Economics
Guo Xiangyu (Tue,) studied this question.