Microsoft kills its Windows blue screen of death
Microsoft is replacing the original Black Screen of Death after 40 years. PHOTO: MICROSOFT
After 40 years, Microsoft is replacing its iconic 'blue screen of death' (BSOD) for a cleaner, sleeker black screen.
The infamous BSOD was plastered across monitors in July 2024 when a faulty CrowdStrike update crashed computer systems worldwide. The incident, which affected about 8.5 million Windows devices globally, triggered system problems that grounded flights and crippled the healthcare, shipping and finance industries.
Microsoft revealed earlier in 2025 that the error message , which appears during unexpected restarts on Windows computers , will be renamed the Black Screen of Death.
In a blog post on June 26 , Microsoft announced it would roll out the new BSOD in an update to Windows 11 24H2 devices 'later this summer'. The new screen will no longer feature the frowning face and QR code. It will instead display the stop code and faulty system driver.
Along with it comes a 'quick machine recovery mechanism for PCs that cannot restart successfully', said Microsoft. These changes are expected to reduce disruption in the event of an unexpected restart. The restart is set to take about two seconds for most users.
The original BSOD has lived in the Windows' software for over 40 years.
Netizens reacted to the news with a mix of dry humour and nonchalance.
'A moment of silence for the blue screen of death. May its replacement be equally chaotic and confusing,' one X user deadpanned.
Calling it 'the end of an era', a Windows user posted on X: 'So many lost hours of my life and so much cursing at my computer. I'll miss it... maybe.'
A Reddit user summed it up for most people with this comment: 'Fine, as long as we don't have to change the acronym.'
Join ST's Telegram channel and get the latest breaking news delivered to you.

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles

Straits Times
4 hours ago
- Straits Times
Microsoft kills its Windows blue screen of death
Microsoft is replacing the original Black Screen of Death after 40 years. PHOTO: MICROSOFT After 40 years, Microsoft is replacing its iconic 'blue screen of death' (BSOD) for a cleaner, sleeker black screen. The infamous BSOD was plastered across monitors in July 2024 when a faulty CrowdStrike update crashed computer systems worldwide. The incident, which affected about 8.5 million Windows devices globally, triggered system problems that grounded flights and crippled the healthcare, shipping and finance industries. Microsoft revealed earlier in 2025 that the error message , which appears during unexpected restarts on Windows computers , will be renamed the Black Screen of Death. In a blog post on June 26 , Microsoft announced it would roll out the new BSOD in an update to Windows 11 24H2 devices 'later this summer'. The new screen will no longer feature the frowning face and QR code. It will instead display the stop code and faulty system driver. Along with it comes a 'quick machine recovery mechanism for PCs that cannot restart successfully', said Microsoft. These changes are expected to reduce disruption in the event of an unexpected restart. The restart is set to take about two seconds for most users. The original BSOD has lived in the Windows' software for over 40 years. Netizens reacted to the news with a mix of dry humour and nonchalance. 'A moment of silence for the blue screen of death. May its replacement be equally chaotic and confusing,' one X user deadpanned. Calling it 'the end of an era', a Windows user posted on X: 'So many lost hours of my life and so much cursing at my computer. I'll miss it... maybe.' A Reddit user summed it up for most people with this comment: 'Fine, as long as we don't have to change the acronym.' Join ST's Telegram channel and get the latest breaking news delivered to you.


International Business Times
21 hours ago
- International Business Times
What If Your Cloud Knew Exactly What You Needed? AI Copilots Are Getting Uncannily Close
With growing technology, cloud platforms are already becoming a big part of how businesses run their services, store data, and manage infrastructure. But now, something new is happening these platforms are getting smarter, and in some cases, almost intuitive. Thanks to AI-powered assistants, often called copilots, users are beginning to experience cloud systems that can suggest, fix, and even guide actions in real time. At Microsoft, a seasoned professional, Sadhana Paladugu, has been part of this shift. As a technical leader on the Azure team, she has worked on tools that aim to make the cloud not just more powerful but easier to work with. One of her most important contributions has been building AI tools that reduce user effort, speed up development, and prevent common cloud-related mistakes. Discussing her projects, she mentioned the VM Replication Copilot, a feature built to help users replicate virtual machines across different cloud regions. She highlighted how doing this manually can be time-consuming and full of hidden problems like picking a server type or location that isn't compatible. This tool uses AI to highlight what won't work and suggest better alternatives. According to internal results, it reduces setup time by more than 60%. Additionally, she developed a preflight validation system that checks for problems before cloud resources are deployed. This step has helped reduce customer support issues by around 40%. She worked on improving the process developers use to push code into production too. By detecting unstable tests and automating how they are handled, she helped increase the number of successful code updates by 80%. Her projects go beyond just saving time. They have improved reliability and made the cloud experience smoother for developers and engineers. In fact, she also helped speed up key pages in the Azure portal by about 20%, making everyday cloud management tasks faster. On top of these technical contributions, she has also been focusing on sharing her insights with the industry through research work like, "Demystifying Google Cloud AutoML Vision: A Comprehensive Guide to Automated Image Classification". This work of hers explores how to build custom image classification models using Google Cloud's AutoML Vision. It walks through the key steps preparing the dataset, training the model, checking its performance, and deploying it. The paper also highlights how the platform works and where it can be useful in real-world applications. As is evident across the industry, AI copilots are becoming more common. From writing code to configuring infrastructure, organisations are now looking to bring smarter tools into every part of the cloud experience. Microsoft, Google, Amazon, and others are all pushing to make their platforms more responsive and easier to use. What makes Paladugu's work stand out is how closely it connects AI to real-world cloud problems. For example, she built a system that can recommend better options when a user's cloud setup isn't available due to regional limits or system constraints. Instead of a failed deployment, users now get helpful suggestions right away no need to look through pages of documentation. But none of this has been easy, as she pointed out. Designing these systems from scratch required close teamwork between engineering, design, and product teams. One big challenge was ensuring that the AI's suggestions were clear and trustworthy. She explained how it's not just about being right, rather the user needs to understand why the Copilot is suggesting something. Looking ahead, industry experts like Paladugu agree that tools like these could change how people interact with the cloud entirely. It is believed that copilots will soon go from just offering help to automatically fixing problems, especially in hybrid setups that include both cloud and edge devices. "The shift toward intent-aware cloud systems is inevitable. AI Copilots are becoming embedded not just in development environments but in deployment, monitoring, and diagnostics," suggested the professional. But building trust will be key. Users need to feel confident that the system is acting in their best interest and they need transparency into how decisions are made. In conclusion, this shift in cloud technology is important. It's no longer just about raw infrastructure, but about smarter, more personal tools that understand what users are trying to do and help them get there faster. As cloud platforms continue to grow, tools like these may soon feel less like assistants and more like quiet partners, working in the background to make sure everything just works.

Straits Times
a day ago
- Straits Times
A ‘white-collar bloodbath' doesn't have to be our fate
The distinction between augmentation and replacement can be subtle. PHOTO: REUTERS There's been a lot of talk in recent weeks about a 'white-collar bloodbath ', a scenario in the near future in which many college-educated workers are replaced by artificial intelligence (AI) programmes that do their jobs faster and better. Dr Dario Amodei, the chief executive of the AI company Anthropic, recently predicted that half of entry-level positions in fields like law, consulting and finance could meet this fate in just a few years. Join ST's Telegram channel and get the latest breaking news delivered to you.