Microsoft Project Silica, AI-cured Dog Cancer, Claude introduces "Dispatch", Google launches Stitch

AI Agents Take Control, Autonomous Systems Scale, and AI Enters the Physical World

Each week, AI isn’t just improving, it’s becoming more embedded across the systems that power how we work, design, and store information. This week’s developments reveal a clear shift: AI agents are becoming more autonomous, software is turning into self-operating systems, and breakthroughs are beginning to extend beyond digital workflows into the physical world.

From agents that can execute tasks across devices without supervision, to AI-generated designs moving directly into production, to entirely new approaches in long-term data storage and personalised medicine, the scope of AI is expanding rapidly. At the same time, these systems are becoming more accessible, enabling individuals to build, automate, and solve problems that previously required entire teams.

Together, these shifts signal a transition from AI as a supporting tool to AI as an independent operator within larger systems, reshaping how work gets done, how products are built, and how knowledge is preserved over time.

Microsoft’s Project Silica: Data Storage That Could Last 10,000 Years

Microsoft is pushing the boundaries of long-term data storage with Project Silica, a system that can preserve information for up to 10,000 years using glass.

The technology uses femtosecond lasers to engrave microscopic 3D structures, known as voxels, into quartz glass. Each pulse operates in a trillionth of a second, allowing data to be embedded deep inside the material rather than stored on the surface. A single glass disc, roughly the size of a DVD, can store up to seven terabytes of data.

What makes this breakthrough significant is its durability. Unlike traditional storage systems such as hard drives or magnetic tapes, glass does not degrade from heat, water, or electromagnetic interference. Microsoft has already tested the material under extreme conditions, including boiling, baking, and flooding, with no data loss.

More importantly, the shift from specialised fused silica to standard borosilicate glass makes the technology far more scalable. Microsoft is now developing robotic storage libraries built around these glass platters, aiming to replace tape-based cold storage in data centres.

This signals a long-term shift in how humanity preserves critical information, from short-term backups to civilisation-scale archival systems.

AI Helps Create First Personalised Cancer Vaccine for a Dog

An Australian tech entrepreneur, Paul Conyngham, used artificial intelligence and genomic analysis to develop a personalised cancer treatment for his dog, Rosie, after conventional therapies failed.

After sequencing Rosie’s tumour DNA through researchers at the University of New South Wales, he used tools like ChatGPT and AlphaFold to analyse mutations and identify potential targets for immunotherapy. The findings were then taken to the UNSW RNA Institute, where scientists developed a custom mRNA vaccine tailored specifically to Rosie’s cancer profile in under two months.

Early results showed significant tumour reduction and improved mobility, although researchers emphasise this is not a definitive cure. Instead, it represents an early example of combining AI, genomics, and mRNA technology to accelerate personalised treatment development.

The case highlights a broader shift: AI is not replacing medical expertise, but increasingly acting as a powerful research assistant, helping individuals and scientists navigate complex biological data faster than ever before. It also raises an important question for the future of medicine: how quickly can personalised therapies move from experimental cases to scalable human treatments?

Claude Introduces “Dispatch”: AI Agents You Can Control From Your Phone

Anthropic has introduced new capabilities for Claude that signal a major shift toward fully autonomous AI agents.

One of the key features, Claude Dispatch, allows users to control their AI directly from their phone by simply scanning a QR code. Once connected, users can send complex, multi-step instructions to their computer remotely, without opening a laptop or monitoring execution. Tasks like generating presentations, cleaning datasets, or managing emails can be initiated via a single message and completed autonomously in the background.

Alongside this, Anthropic also introduced Channels, which allow AI agents to respond to live events across platforms like GitHub, Telegram, or Discord. This means an AI system can detect issues, analyse codebases, and even suggest or deploy fixes without human intervention.

The broader implication is significant. AI is no longer just a tool that requires active use, it is becoming a background operator, executing tasks asynchronously while users focus elsewhere. This marks a shift from “using AI” to managing AI workers, a direction that competitors like Sam Altman and Sundar Pichai are also rapidly moving toward.

Google Launches Stitch: AI-Powered UI Design Without Traditional Tools

Google has introduced Stitch, a new AI-powered design tool that aims to simplify how interfaces are created, potentially reducing reliance on traditional tools like Figma in early-stage workflows.

Stitch allows users to generate complete UI layouts by simply describing what they want. From landing pages to complex dashboards, the system automatically creates components, spacing, and structure based on natural language input. It also supports real-time interaction through voice, enabling users to modify designs directly on the canvas without manual editing.

The tool is powered by Gemini 2.5 Pro, which enables a deeper understanding of spatial design and layout consistency. Changes made to one element can propagate across the entire design system, while built-in preview capabilities allow users to test prototypes instantly on mobile devices. Designs can also be exported to Figma or as MCP-compatible files for development tools like Cursor.

Stitch highlights a broader shift in product development. The gap between idea and execution is rapidly shrinking, as AI systems begin to handle not just code generation, but also design thinking, layout decisions, and real-time iteration.

Why This Week Matters!

This week’s developments highlight three clear shifts shaping the AI ecosystem.

AI moving beyond software into infrastructure: Innovations like Microsoft’s Project Silica show how AI is now influencing long-term data storage and physical systems.

AI accelerating personalised innovation: From custom cancer treatments to AI-assisted research, individuals can now solve complex problems once limited to large institutions.

AI becoming autonomous and always-on: Tools like Claude Dispatch and Google’s Stitch reflect a shift toward AI systems that operate independently in the background.

At the same time, AI is embedding itself into everyday products, quietly reshaping how work gets done and decisions are made.

Taken together, these changes show that AI is evolving from a tool into core infrastructure.

At QuensultingAI, we help organisations turn these shifts into practical outcomes, from intelligent agents to scalable AI systems.

The advantage now lies in how effectively you deploy and integrate AI. 🚀