The Great Leap of Agentic AI: Between "God-Speed" Infrastructure and the Open Source Rebellion
Weekly summary of the most important news in the world of Artificial Intelligence.
# The Great Leap of Agentic AI: Between "God-Speed" Infrastructure and the Open Source Rebellion
As we begin this last week of January 2026, the Artificial Intelligence landscape is consolidating on three critical fronts: the direct challenge of native AI infrastructures like **Railway** to the cloud giants; the birth of a developer "rebellion" who prefer local and free agents like **Goose** over Anthropic's expensive subscription models; and a growing concern in economies like the United Kingdom, where AI is beginning to displace jobs on a net basis for the first time.
### Railway and the end of the "slow cloud": $100 million to displace AWS
The news of the day is led by **Railway**, the cloud infrastructure platform that has just closed a $100 million Series B round. Its premise is radical: in a world where AI agents like Claude or GPT-5 generate code in seconds, waiting three minutes for a deployment on AWS is unacceptable. Railway has built its own data centers from scratch to offer deployments in less than a second, enabling what its CEO, Jake Cooper, calls "agentic speed." With 2 million users and organic growth without marketing, Railway demonstrates that traditional infrastructure is failing the developers of the AI era.
### The rebellion against $200 a month: Goose challenges Claude Code
The cost of the "code revolution" has reached a breaking point. While Anthropic charges up to $200 per month for **Claude Code** with frustrating usage limits, **Block** (Jack Dorsey's company) has launched **Goose**. This is an open-source AI agent that runs locally on the user's computer, eliminating subscriptions, cloud latency, and privacy risks. With more than 26,000 stars on GitHub, Goose allows programmers to use models like Llama or Qwen for free, marking the beginning of an era where data sovereignty and cost savings could curb the dominance of closed models.
### The social impact: Between job loss and human creativity
A new report from Morgan Stanley reveals an uncomfortable reality: the United Kingdom is losing more jobs than it creates due to AI, a higher rate than the US or Japan. At the same time, a massive study of 100,000 people confirms that generative AI already surpasses the average human in creativity tests. However, there is a ray of hope for our species: the 10% of the most creative humans remain unattainable for AI in complex tasks such as narrative and poetry, suggesting that the "human spark" is still, for now, our best defense.
* **Goose (from Block):** The open-source agent that promises total autonomy in software development without relying on external servers.