Fable 5 returns as AI job signals diverge

Anthropic restored Fable 5 on July 1 after Amazon researchers found a safeguard bypass that could coax the model into identifying software vulnerabilities. US export controls landed on Fable 5 and Mythos 5 on June 12, prompting a global access suspension, but the controls were lifted as of June 30. Anthropic says a new safety classifier blocks the reported technique in over 99% of cases, and the company is working with Amazon, Microsoft, Google, and other Glasswing partners on a jailbreak-severity framework.

Workforce signals remained mixed. Palo Alto Networks CEO Nikesh Arora warned that roughly 90% of large-company employees lack needed AI skills, while the company relies heavily on hackathons and attrition to reshape its roughly 21,000-person workforce. A Ramp and Revelio Labs study of nearly 22,000 companies found heavy AI adopters grew headcount 10.2%, with entry-level roles up 12%, while Anthropic’s index found 57% of surveyed users said AI raised their skills’ market value and only 10% saw job loss as likely.

Other developments spanned research, education, and products. A Brown professor flagged 40 of 86 perfect take-home midterm scores as an AI cheating case, Meta reported 61% word accuracy for Brain2Qwerty v2, and Google released Nano Banana 2 Lite, which generates images in four seconds at $0.034 per 1,000 images. Anthropic’s hiring of Stanford economist Chad Jones drew attention over a past paper weighing a 33% extinction risk against a projected 55-fold rise in living standards.