Microsoft has raised concerns about an increase in cyber attacks targeting internet-exposed operational technology (OT) devices since late 2023. These attacks can compromise industrial processes by manipulating OT system parameters, leading to malfunctions and outages. Microsoft highlighted the importance of enhancing OT security to prevent exploitation, especially given the devices' vulnerability due to weak passwords and outdated software. Recent geopolitical tensions have further heightened risks, with advisories from Rockwell Automation and the U.S. Cybersecurity and Infrastructure Security Agency (CISA) urging protective measures. Additionally, the emergence of malware like Fuxnet demonstrates the destructive potential of such attacks on critical infrastructure.

Microsoft Warns of Surge in Cyber Attacks Targeting Internet-Exposed OT Devices The Hacker News

Salesforce has experienced a slowdown in growth, with fiscal first-quarter revenue increasing by just 10.7% and a forecasted 7% growth rate for the current period, both falling below Wall Street's expectations. Billings grew by a mere 3%, and the company's current remaining performance obligation rose 9.5%, marking its first single-digit increase. Despite achieving over 32% in adjusted operating margins, the market's focus on top-line growth led to a significant 20% drop in Salesforce’s stock. The broader cloud software industry also faces challenges, as companies like Workday, Snowflake, and others report slower deal closures and increased scrutiny, partly attributed to growing investments in AI.

Salesforce: major deals are taking longer to close, if they close at all. The Wall Street Journal

The article discusses the slowing pace of innovation in artificial intelligence (AI) and the growing challenges associated with its implementation and profitability. Despite significant investments and high expectations, AI's improvement rate is decelerating, and it is becoming less efficient due to the saturation of available training data. Large-scale companies may sustain their AI ventures with vast resources, but AI startups face difficulties competing and surviving. Additionally, high operational costs and limited practical applications put into question AI's potential to generate significant returns, much like the fiber-optic boom of the late 1990s that led to a market crash. Evidence also suggests that AI is not the productivity booster it was projected to be, creating skepticism about its transformative impact on industries and jobs.

The AI Revolution Is Already Losing Steam The Wall Street Journal

Google Research and Google DeepMind have introduced Med-Gemini, a suite of advanced AI models fine-tuned for the medical field. These models build on the capabilities of the original Gemini models, enhancing their performance on a wide range of text, image, video, and electronic health record tasks crucial for medical applications. Med-Gemini is designed for sophisticated multimodal and long-context reasoning, showing state-of-the-art results in benchmarks such as the MedQA and NEJM clinico-pathological conferences. Research papers specifically highlight the application of these models in radiology, pathology, dermatology, ophthalmology, and genomics, demonstrating notable improvements in clinical tasks, like generating radiology reports and predicting health outcomes from genomic data. Despite the promising advancements, extensive real-world testing is essential to ensure the models’ safety and reliability. Future collaborations with healthcare professionals and researchers are encouraged to further refine these AI tools.

Advancing medical AI with Med-Gemini Publication