Wire flash
Meta's Muse AI hit by zero-day vulnerability allowing Mac device hijack
Editorial responsibility
- No named human review is recorded for this page.
- Source reporting is collected, normalized, translated or condensed automatically when needed.
- Automatically published source-backed update
This article from East Money News reports multiple setbacks for the AI sector, centered on Meta's newly launched Muse AI assistant. Within weeks of its debut, Muse was hit by a zero-day vulnerability that could allow attackers to hijack Mac devices, as disclosed by macOS security expert Patrick Wardle. The product's daily active users remain below 1 million, and third-party monitor SaaSHub has marked its status as 'Degraded.' Performance issues include resource consumption 10 times higher than internal tests, frequent search failures, and a stress test where only 33 of 120 sub-agents ran successfully. Separately, a negative supply chain report about optical modules circulated in investment circles, and OpenAI reportedly paused its latest AI generation. Bank of America analyst Michael Hartnett warned that the MOVE index, measuring U.S. Treasury bond volatility, surged 33% in two days, signaling financial market stress. Analysts cited in the article suggest these three factors could complicate supply chains, slow U.S. AI infrastructure deployment, and pressure AI-related stocks. The piece also notes that AI users would need to spend approximately $1 trillion annually for hyperscale data center operators to achieve meaningful returns on investment.
Source report
The artificial intelligence sector is confronting a fresh wave of challenges, with three major developments threatening to slow deployment and complicate supply chains.
Muse: From Rising Star to Troubled Product
Meta's newly launched AI assistant, Muse—which had briefly revitalized the AI sector and driven the company's stock higher—is now facing significant setbacks. The consumer-grade agent has yet to reach 1 million daily active users and has already been flagged as "Degraded" by third-party monitoring platform SaaSHub.
Critical Security Vulnerability
A zero-day vulnerability has been discovered in the Muse application, posing severe security risks to Apple macOS devices. Security expert Patrick Wardle disclosed the flaw on social media, warning that it allows local malware or attackers to hijack a user's Mac device without their knowledge.
As of now, Meta has not yet patched the vulnerability, though industry observers expect a fix to be released soon.
Performance and Stability Issues
Muse's operational problems have emerged alongside its underwhelming user adoption:
- Actual resource consumption is approximately 10 times the internal testing benchmark
- SaaSHub has marked its status as "Degraded"
- "Unable to search" is the most frequently reported user issue
- In a stress test, only 33 out of 120 sub-agents ran successfully, with the remaining 87 failing entirely
Industry-Wide Bottlenecks
Analysts point out that Muse's struggles reflect broader challenges facing the personal AI agent sector, including:
- General computing capacity
- Memory and storage adaptation
- Browser isolation and security monitoring
- Concurrent task orchestration
- High operational costs
For Muse to achieve long-term viability, it must overcome three core hurdles: successful run rate, unit operating cost, and commercial user retention. Without progress in these areas, the product risks becoming a costly technical experiment.
Supply Chain and Market Pressures
Over the weekend, negative news regarding the optical module supply chain also circulated widely among investors, adding to sector concerns.
Macroeconomic Headwinds
According to MarketWatch, a Bank of America report released on the 25th by Michael Hartnett's team highlighted a sharp spike in the MOVE index—which measures expected volatility in U.S. Treasury bonds—soaring 33% in just two trading days.
Hartnett noted that the U.S. Treasury market serves as the core benchmark for global lending markets. A rapid rise in its volatility is a negative signal for overall financial markets.
AI Profitability Concerns
While top-tier cloud service providers are seeing strong revenue growth—with annualized revenue expected to exceed pre-AI trends by approximately $700 billion by Q2 2026, and reported backlogs surpassing $1.5 trillion—the path to profitability remains uncertain.
Hartnett emphasized that AI users would need to spend roughly $1 trillion annually on AI applications for hyperscale data center operators to achieve meaningful returns on investment, and for the application layer to generate strong margins on computing expenditure.
Source: Securities China
Source
东方财富网-产业透视Eastern
Part of this Story
Meta's Muse AI hit by zero-day flaw, OpenAI pauses model, sector faces headwinds