Technology 11/6: Meta cuts staff, DeepSeek mines coal

Tech companies cut staff

In the context of digital transformation and the strong development of artificial intelligence (AI), large technology corporations have made many rounds of personnel cuts to optimize costs and improve operational efficiency.

Meta announced plans to cut 5% of its workforce, with the goal of improving performance and eliminating underperforming employees. Microsoft also conducted performance reviews and laid off some employees, but without providing severance pay, many workers were left struggling.

Despite strong financial performance, Salesforce is cutting staff.

Salesforce, despite its stable financial situation, decided to cut more than 1,000 employees, while HP Enterprise (HPE) announced cutting 2,500 jobs to restructure its operations. Microchip Technology is also following the trend, cutting 2,000 employees, expected to be completed by the end of the second quarter of this year.

In addition, Stripe – an online payment platform – is eliminating 300 employees, mainly in the product, engineering and operations departments.

The layoffs reflect a shift in strategy for tech companies in the age of AI, with a greater focus on recruiting highly skilled people in data analytics and AI. However, the move also raises questions about the long-term stability of the tech industry’s labor market.

WWDC 2025: Apple Praises Developers Amid AI Disappointments

WWDC 2025 ended with a humorous tribute to developers, but the silence on important issues left the tech community with many questions.

While Apple has expanded its AI integration by adding ChatGPT to Image Playground and coding support in Xcode, many developers remain dissatisfied with the company’s progress in this area. The “more personal” Siri AI, promised at last year’s WWDC, continues to be delayed, citing the need for more time to meet Apple’s quality standards. Other AI features largely lag behind Google’s improvements.

The App Store issue was also not mentioned by Apple during the 90-minute keynote, despite the company having just complied with a ruling that forced it to allow US developers to link to payment methods outside of Apple’s system. Changes to payments and anti-fraud were announced ahead of the event but did not appear on the main stage.

Additionally, Apple introduced standalone gaming apps, but focused on the benefits for consumers rather than developers. WWDC 2025 still left many unanswered questions about Apple's direction with AI and the developer community.

OpenAI Launches o3-pro: The Most Powerful Inference AI Model Ever

OpenAI has just introduced o3-pro, an upgraded version of its o3 inference AI model that promises to improve its ability to solve problems in the fields of mathematics, physics, and coding.

O3-pro is now available on ChatGPT Pro, Team, and the Developer API, replacing o1-pro. According to internal testing, O3-pro scores exceptionally well on popular AI tests, surpassing Google Gemini 2.5 Pro on AIME 2024 (a math skills assessment) and beating Anthropic's Claude 4 Opus on GPQA Diamond (a doctoral-level science knowledge test).

The model has access to a variety of advanced tools, including web scraping, data analysis, Python, and personalized responses. However, OpenAI admits that the O3-pro is slower to respond than its predecessor. Some features are still lacking, such as image generation and support for the AI Canvas workspace.

Despite its limitations, the O3-pro still marks an important step forward in AI inference technology, opening up new prospects in practical applications.

China puts DeepSeek into coal mining

China's coal industry is witnessing a dramatic transformation as AI technology from DeepSeek is deployed at major mines, helping to optimize performance, improve safety and modernize mining processes.

At the Liyazhuang mine of Shanxi Coking Coal Group, an AI system nicknamed “Xiao Ya Assistant” is deeply embedded in production operations. With just one click and input of technical parameters, a report on production safety measures is instantly generated, helping technicians control risks more effectively. The system is also capable of updating production status in real time, with more than 500,000 pieces of data processed on the local server.

DeepSeek not only helps improve productivity, but also acts as a “safety guard” that continuously monitors working conditions in the mine. AI creates a tight safety management loop. In addition, the system supports smart maintenance by analyzing machine vibration data, predicting failures early, and switching from passive to proactive maintenance.

The AI revolution in coal mining will not only improve productivity but also improve sustainability and worker safety. If this trend continues, it could reshape the entire mining industry, ushering in a new era of more advanced intelligent standards.

Minh Hoan

Source: https://vtcnews.vn/cong-nghe-11-6-meta-cat-giam-nhan-su-deepseek-khai-thac-than-ar948188.html