Osservatorio Evolutivo
Ricerche
Ricerca approfondita su AI, automazione e intelligenza operativa. Ogni tema produce un articolo accessibile e un rapporto accademico completo.
Articoli
Can AI Language Models Actually Improve Your Forecasts? What the Research Says
Early experiments with LLM embeddings for forecasting showed near-zero gains. New research reveals the problem was implementation, not concept. Here's what works.
Why Oracle Says AI Models Are Worthless — and What It Means for Your Business
Larry Ellison claims AI models are becoming commodities. Oracle's new AI Database 26ai bets the future on private data. Here's what businesses need to know.
Rapporti
Can LLM Embeddings Improve Time Series Forecasting? A Practical Feature Engineering Approach
The integration of Large Language Models into Time Series Forecasting has emerged as a significant area of research. However, naive approaches like embedding concatenation yield negligible gains (50.0% to 50.47% accuracy). This report evaluates 2024-2026 advances, identifying three effective alignment paradigms — reprogramming, direct vectorization, and multi-layer fusion — that produce meaningful improvements, particularly in zero-shot, data-scarce, and text-rich forecasting scenarios.
Oracle's "AI Database 26ai" and the Shift to Private Data Moats
Oracle Chairman Larry Ellison contends that large language models are converging into functionally equivalent commodities trained on identical public data. The true strategic asset, he argues, lies in private enterprise data. Oracle's AI Database 26ai enables LLMs to reason directly over private data without data movement, supported by a $523 billion contract backlog including a $300 billion OpenAI infrastructure deal. This report analyses the technical architecture, financial performance, industry implications, and strategic risks of Oracle's positioning as the gatekeeper of enterprise AI.