Bloomberg & The Rise of AI Agents: A Deep Dive
Welcome! You can expect the best in tech and AI news delivered to you regularly. This newsletter is designed to keep you informed and at the forefront of this rapidly evolving field.
Don’t forget to confirm your subscription in your inbox and explore additional groups you might be interested in. Also, connect with us on your favorite social media platforms!
The Agent Revolution is Here
As Matt Burns, Director of Editorial at Insight Media Group, points out, the focus isn’t just on AI headlines, but on their practical implications for individuals and organizations. The core idea is that those who master AI will lead the next wave of innovation. And this newsletter is here to help you do just that.
Speaking of leading, Roadmap, a community-driven resource for developers, has achieved a remarkable milestone – becoming the 6th-most-starred project on GitHub, with its founder, Kamran Ahmed, ranking as the 2nd-most-starred individual on the platform. A testament to the power of collaborative learning!
From Developer Tools to Everyday Assistants
The landscape is shifting. AI agents are evolving beyond simple software features and becoming the primary interface for interaction. This week marked a turning point, as agents began to cater to non-technical users.
Perplexity’s ‘Computer’: A Digital Worker
Perplexity launched ‘Computer,’ a “general-purpose digital worker” that goes beyond chatbots and search engines. It automates tasks by deploying sub-agents that browse, research, create, and connect to apps like Gmail, Slack, Notion, and more. As Perplexity puts it, “Chat answers. Agents do tasks. Computer works.” You can explore examples of its capabilities on their product page.
Expanding Agent Capabilities
Anthropic expanded Claude Cowork with 13 new enterprise plugins, integrating with tools like Google Workspace, DocuSign, and LegalZoom. They also released Remote Control for Claude Code, enabling mobile access, and a scheduling tool. Meanwhile, Notion introduced Custom Agents, allowing workflows to run on a schedule or in response to specific requests, supported by a new GitHub toolkit for building custom tool calls. Eric Goldman of Notion describes it as a “programmable partner for custom agents.”
The Agent Framework Wars & Standardization
The competition between agent frameworks is intensifying, drawing parallels to the earlier “container wars.” Anthropic is pushing for standardization with its Agent Skills repository on GitHub, which has already garnered over 76,000 stars. This initiative, following the success of MCP, aims to define the standards for how agents utilize tools. Microsoft, Cursor, Goose, Amp, OpenCode, and even OpenAI have adopted Agent Skills, and the spec is now managed by the Agentic AI Foundation under the Linux Foundation.
Agentic Engineering: A New Paradigm
Andrej Karpathy has coined the term “agentic engineering” to replace “vibe coding,” reflecting a significant shift in how software is developed. He demonstrated this by tasking an agent with setting up a home video analysis dashboard with a single prompt, and the agent successfully completed the task, researching solutions and debugging along the way. This echoes Matt Shumer’s observation that coding agents are now viable, marking a threshold of coherence and tenacity.
The Changing Distribution Channel for Software
Aakash Gupta highlights that agents are becoming the new distribution channel for software, directly accessing CLIs, MCP servers, and documentation programmatically. Benjamin Lee emphasizes the importance of code quality, architecture intuition, and systems thinking in the age of AI coding agents, as developers will be responsible for the consequences of the code generated by these agents.
Cloudflare & Accessibility for AI Agents
Cloudflare launched Markdown for Agents, converting HTML to clean markdown to reduce token usage for AI agents. This infrastructure-level improvement makes websites behind Cloudflare more accessible to AI agents without requiring any code changes.
Political Implications & National Security
The intersection of AI and national security is becoming increasingly complex. President Trump directed federal agencies to stop using Anthropic’s products, citing concerns about its political leanings. Simultaneously, OpenAI secured a deal with the Pentagon to deploy its models on a classified network, with provisions regarding domestic surveillance and autonomous weapon systems. However, the language used in the OpenAI agreement differs from Anthropic’s requests, raising questions about the level of protection offered.
This situation highlights the need for clear standards and regulations in the AI space, particularly concerning national security applications.
Stay tuned for further updates and insights into the evolving world of AI agents!
Community created roadmaps, articles, resources and journeys for developers to help you choose your path and grow in your career.
