Anthropic's Fable 5, its latest AI model, is falling behind competitors as businesses opt for cheaper, more efficient alternatives. Data from fintech company Ramp indicates that spending on Fable 5 accounts for only 11% of Anthropic's enterprise client revenue, significantly less than the older and more affordable Opus 4.8.

Despite Fable 5's initial reputation for being so powerful it was briefly banned, businesses are weary of the high cost of state-of-the-art models, which are priced at $10 per million input tokens and $50 per million output tokens, making it one of the priciest on the market. They prefer less expensive or free models that can adequately handle most day-to-day operational tasks.

Early users complained about Fable 5's hyper-sensitive safeguards making it nearly unusable and its steep price pointing towards a future where only a small elite can afford advanced AI. The emergence of more affordable alternatives from Chinese labs like Moonshot and Z.ai is driving a more sober and frugal approach to AI adoption in the workplace.

👉 Read the full article on Gizmodo

Chinese state-linked hacker groups have more than doubled their attack volume by integrating inexpensive AI models, primarily DeepSeek, into their routine operations, according to a report by Taiwanese cybersecurity research firm TeamT5. These AI models are accelerating reconnaissance and preparatory tasks such as searching for exposed software, extracting public exploit code, and mapping target domains.

DeepSeek is favored by Chinese hackers due to its relative power and minimal cybersecurity barriers, unlike Western models which have stricter safeguards. While more powerful Chinese models like Moonshot's Kimi K3 exist, their operational cost is prohibitive for attackers aiming for high volume.

However, a detailed analysis by Palo Alto Networks' Unit 42 reveals that in the most thoroughly documented campaign, no system compromises were achieved through autonomous AI execution. All confirmed intrusions were the result of manual human effort, suggesting that while AI streamlines attack preparation, it has not yet achieved fully autonomous breaches.

👉 Read the full article on Gizmodo en Español

Elon Musk has consistently promised a future where Americans won't need to work and everyone will be wealthy, thanks to his humanoid robot Optimus. Paradoxically, he frequently complains about the government providing free services, even suggesting that ancient Rome's distribution of free bread contributed to its downfall.

Musk's contradictory stance clashes with his actions, such as overseeing the departure of approximately 300,000 government workers as head of the Department of Government Efficiency (DOGE), which weakened social safety nets. His vision of a robot-driven utopia where no one works is at odds with his criticism of government assistance.

There is no evidence that even widespread production of Optimus robots would automatically lead to a post-scarcity utopia. Social safety nets, like universal basic income, are achieved through political means, not solely technological advancements. Musk's claims that his robots will eliminate the need for the "free stuff" he criticizes are disconnected from how societal wealth distribution actually works.

👉 Read the full article on Gizmodo