Jack Dorsey has been a proponent of open technology for years, and his newest AI venture, Goose, is starting to build momentum within the developer community. The independent AI assistant, launched in January 2025 by Block’s Open Source Program Office, aims to allow users to implement large language models into apps with real-world applications.
Now, just weeks after its debut, Goose is seeing widespread adoption—and its trajectory suggests it’s likely to be a major contributor to the open AI ecosystem.
Dorsey recently celebrated another benchmark of the open-source AI revolution when he reacted to the March 6, 2025, debut of Mansa, an open LLM created beyond conventional tech behemoths with a one-word thumbs-up – “excellent” in a tweet to his more than six million X/Twitter followers.
The endorsement is a harbinger of Dorsey’s ambitious vision for AI – equipping individuals and enterprises with open, adaptable AI tools.
Open-Source AI Developers Flock To Goose
Goose has experienced more interest from developers than expected, Block’s Head of Open Source Manik Surtani wrote in an email response.
“We’ve been blown away by the engagement from the open-source community,” Surtani added.
“Goose was a top trending project on GitHub for several weeks after launch, and we’ve seen huge growth in our Discord community. We’re also seeing some really interesting experimentation within other businesses, like Databricks,” he noted.
Most notable about Goose’s popularity is the fact that it is capable of performing intricate coding work with little human engagement. Unlike traditional coding assistants that require step-by-step approvals, Goose allows for what Brad Axen, Tech Lead for AI & Data Platform at Block, calls “vibe coding.”
In practice, this implies that programmers can provide high-level objectives, and Goose does lots of work independently.
Goose AI Use Cases Moving Beyond Engineering
Originally developed as a tool for software development, Goose is already seeing use in non-development teams within applications. Axen stated that Block’s internal teams are utilizing Goose for sales analysis, content asset management, and even onboarding new hires.
“We’re seeing sales teams analyze thousands of leads in hours instead of days, content teams automating complex asset management, and project managers cutting administrative time by 75%. The emotional feedback we’re getting — like ‘I could cry it was so helpful’ — really shows how these tools are transforming daily work,” Axen wrote in an email exchange.
Among the key technical advances fueling Goose’s growth is its Anthropic Model Context Protocol support. Goose uses MCP to integrate with a wide range of third-party tools and systems, making it extremely flexible.
Basically, MCP serves as an invisible, open-source “rules to the AI road” guide that establishes guardrails to ensure that AI stays:
- Stays helpful and relevant — remembers what’s needed in the moment
- Follows boundaries — doesn’t pretend to be a doctor or lawyer
- Handles questions efficiently — prioritizes what matters
- Protects privacy — doesn’t remember things long-term
“We’ve seen exponential community adoption of MCP, and we’re working with Anthropic to establish a registry of verified, reliable extensions. This is something that will benefit not just Goose users but the entire MCP ecosystem,” Axen explained.
DeAI, Blockchain-Enabled AI And Bitcoin’s Role?
Block is a long-time proponent of Bitcoin and its blockchain, and some consider open-source AI and decentralized finance to be a natural fit. Pete Rizzo — aka The Bitcoin Historian on X/Twitter — recently quoted Dorsey speculating that Bitcoin might be the currency of AI with the rationale that autonomous agents of AI will gravitate toward zero-friction transactions.
Though Block couldn’t verify if Dorsey used those exact words recently, a spokesperson referred to Dorsey’s previous interviews in which he spoke about the use of open protocols and AI agents to reimagine digital transactions.
The prospect of blockchain-enabled AI agents is in line with Block’s overall strategy. Jackie Brosamer, Block’s Head of AI & Data Platform, highlighted the company’s economic empowerment mission.
“Being able to run agentic AI locally means your privacy is protected, and you don’t have to pay for API access. A free, private AI agent has huge implications for democratizing the AI playing field,” Brosamer stated in an email.
What’s Next for AI Assistant Goose?
Adoption among developers of Goose is continuing to gain traction. The project is closing in on a significant milestone, reports Surtani — 10,000 GitHub stars, which is quite an achievement for a project six weeks old. More importantly, Block is getting increasingly high-quality contributions from the open-source community.
“This really shows that people are not just experimenting — they’re figuring out how to improve the tool and sharing those contributions so it becomes better for everyone,” wrote Surtani.
As for the future, Block is working on local AI processing and combining multi-provider models. Axen said the team is already doing work to make Goose support open-source LLMs entirely on-device with less dependence on cloud AI.
“We’re also planning to push the performance envelope by combining the best aspects of models across multiple providers — something that only a flexible tool like Goose can do,” he concluded.
As Goose quickly gains ground speed, and Block remains committed to open-source AI, the project can soon be a standard bearer of the agentic AI revolution.
With the direction of the AI world moving toward autonomous, decentralized modes, Jack Dorsey’s gamble on open-source AI appears like an increasingly safe bet.
