Scientists Use AI to Design Viruses to Target Bacteria

Scientists are using AI to design new biological viruses that can target and kill bacteria, which could lead to new treatments for drug-resistant infections. However, this technology also raises concerns about the potential for malicious actors to create biological weapons. Researchers have used specialized AI models to design complete genomes for bacteriophages, viruses that infect and kill E. coli bacteria.

The US Food and Drug Administration (FDA) is seeking feedback on a competency-based approach to regulating medical devices that utilize generative artificial intelligence (GenAI). This proposed approach would evaluate the performance of GenAI-enabled devices in a real-world setting.

Some AI companies are reaching significant revenue milestones quickly, with some achieving $100 million in revenue in just a few years. This is due in part to their sales-led approach, which involves showcasing products using real-world case studies and working with large enterprise clients early.

Artificial intelligence is becoming a fundamental component of enterprise infrastructure, with sectors such as pharmaceuticals, automotive, and financial services leading the charge. Investment in AI reached over $50 billion in 2025, underscoring its integration into core business operations.

OpenAI has developed a custom inference chip called Jalapeno, which delivers better performance-per-watt than NVIDIA's Blackwell chip in some scenarios. Meanwhile, xAI's Grok 4 Heavy AI system has scored above 40% on Humanity's Last Exam by running up to 16 parallel reasoning agents on a single query.

Key Takeaways

• Scientists use AI to design new biological viruses to target and kill bacteria, raising concerns about potential misuse. • FDA seeks feedback on regulating medical devices that utilize generative AI. • AI companies reach $100 million in revenue quickly with sales-led approaches. • AI investment reached over $50 billion in 2025, integrating into core business operations. • OpenAI develops custom inference chip Jalapeno for better performance-per-watt. • xAI's Grok 4 Heavy AI system scores above 40% on Humanity's Last Exam. • AI-generated viruses could be used to develop personalized treatments for patients. • The term 'meat proxy' describes coworkers who blindly share AI output without verifying accuracy. • AI integration offers opportunities in psychiatric training, including early detection of mental health disorders.

AI Creates Viruses, Raises Concerns

Scientists have used AI to design new biological viruses that can target and kill bacteria. This technology has the potential to create new treatments for drug-resistant bacterial infections, but also raises concerns about the potential for malicious actors to create biological weapons. Researchers used specialized AI models to design complete genomes for bacteriophages, viruses that infect and kill E. coli bacteria. The study highlights the need for careful consideration of the ethics behind AI-powered viruses and bioengineering.

AI-Generated Viruses: Future of Medicine?

Scientists are using AI to design viruses that can attack bacteria, with the hope of creating new treatments for drug-resistant infections. This technology has the potential to revolutionize the field of medicine, but also raises concerns about the potential risks. Researchers believe that AI-generated viruses could be used to develop personalized treatments for patients.

FDA Seeks Feedback on GenAI Device Regulation

The US Food and Drug Administration (FDA) is seeking feedback on a competency-based approach to regulating medical devices that utilize generative artificial intelligence (GenAI). The proposed approach would evaluate the performance of GenAI-enabled devices in a real-world setting. The FDA is also considering a two-axis framework for assessing the risk posed by GenAI software functions.

AI Companies Reach $100M in Revenue Fast

Some AI companies are reaching $100 million in revenue in just a few years, thanks to their sales-led approach. These companies are making the initial decision simple for customers, showcasing their products using real-world case studies, and starting to work with large, high-profile enterprise clients early.

AI Becomes Core to Enterprise Innovation

Artificial intelligence is becoming a fundamental component of enterprise infrastructure, with sectors such as pharmaceuticals, automotive, and financial services leading the charge. Investment in AI reached over $50 billion in 2025, underscoring its integration into core business operations.

xAI's Multi-Agent Stack Powers Grok

xAI's Grok 4 Heavy AI system scored above 40% on Humanity's Last Exam by running up to 16 parallel reasoning agents on a single query. The architecture carries a compute cost, but xAI has developed a data center with 555,000 GPUs to support it.

OpenAI's New Jalapeno Chip

OpenAI has developed a custom inference chip called Jalapeno, which delivers better performance-per-watt than NVIDIA's Blackwell chip in some scenarios. The chip is designed to minimize data movement and uses AI in its own development.

New Term: Meat Proxy

A 'meat proxy' is a term used to describe coworkers who blindly share AI output without verifying its accuracy. This phenomenon highlights the need for critical thinking and evaluation of AI-generated information.

AI in Psychiatric Training

The integration of AI in psychiatric training offers several opportunities, including early detection of mental health disorders and personalized treatment plans. However, it also raises challenges such as bias in AI algorithms and the potential for over-reliance on AI.

Sparse Reading for AI Agents

SparseRead is a new training-free, model-transparent reading layer that controls content admission before unnecessary evidence reaches the model context. It reduces token volume and wall time while preserving task quality.

AI Runs Doom on Custom CPU

An AI computing enthusiast has demonstrated Doom running on a custom CPU designed by GPT-5.6 Sol. The CPU design, called Codex-R32, was created using AI and can run Doom in a pulsing schematic of the CPU.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI-powered viruses Biological weapons Drug-resistant infections Bioengineering Ethics AI-generated viruses Personalized treatments Medicine Future of medicine FDA regulation GenAI device regulation Competency-based approach Risk assessment AI companies Revenue growth Sales-led approach Enterprise innovation Artificial intelligence Core business operations Pharmaceuticals Automotive Financial services xAI's Multi-Agent Stack Grok Parallel reasoning agents Compute cost OpenAI's Jalapeno chip Inference chip Performance-per-watt Custom chip design AI development Meat proxy Critical thinking AI-generated information Psychiatric training Early detection Personalized treatment plans Sparse reading Model-transparent reading layer Content admission Task quality Custom CPU design AI computing Doom on custom CPU CPU design AI ethics AI safety AI regulation

Comments

Loading...