On March 31, 2026, Oracle eliminated up to 30,000 jobs, representing roughly 18% of its global workforce, via a single automated email without prior conversation with HR or warning from line managers. ^[1] Bloomberg-compiled options data implies Oracle's stock could move 10% when the company reports its earnings on Sept. 9. ^[2] Across several architectures and benchmarks, the study found a substantial gap between individual models and their oracle, revealing that different models capture complementary knowledge. ^[3] The researchers found that model complementarity rapidly saturates as more models are added, leaving a persistent subset of queries unsolved even by a large number of models. ^[4] Access to company systems was revoked the same day the layoffs were announced. ^[5] The company's contracted future revenue is at $553 billion, up 325%, almost entirely driven by AI deals. ^[6] The layoffs occurred immediately after Oracle closed a profitable quarter with a $6 billion bottom line. ^[7] At Oracle's most recent earnings announcement on June 10, the implied move was 12.3% but the actual stock price declined 24.7%. ^[8] In five of Oracle's past eight earnings reports, the stock price moved more than the amount implied by options trading. ^[9] Different link prediction models, or even different training runs of the same model, can produce substantially different predictions for the same query. ^[10] The authors propose to measure model complementarity through the performance of an oracle that, for each query, selects the best prediction among a considered set of models. ^[11]