Take part in our daily and weekly newsletters to get the latest updates and exclusive content for reporting on industry -leading AI. Learn more
The identity of the social network with bird motifs Twitter may have been triggered by the new owner Elon Musk, but this did not stop that one of his co -founders, Jack Dorsey, took over new bird names for a new project.
Dorsey’s other company blockThe parent of the point-of-sale-service Square, the cash app for mobile payments, music streaming service and other technically controlled financial instruments, Gans started todaya free, Open source framework This tries to simplify the process of building an AI agent (or many agents) with almost every conceivable large language model (LLM) as an intelligence on the backend, regardless of this Deepseek Or a proprietary model from Openai, Google and Anthropic.
Dorsey actually teased the start of goose less than 24 hours ago Posting on X: “Open source everything.”
The agents that Goose is supposed to build primarily focus on software development. In a blog post in which the new framework was announced, the block developer Advocate Adeewale Abati wrote that the framework has already helped in the following tasks:
- Perform code migrations such as Ember to react, Ruby to Kotlin, Prefect-1 to Prefect-2 etc. etc.
- Immerse yourself in a new project in an unknown coding language
- Transition a code basis from the field-based injection to the constructor-based injection in a dependency injection frame.
- Perform performance benchmarks for a build command with a build automation tool
- Increasing the code cover over a certain threshold
- Scaffolding of an API for database
- Creation of datadog monitors
- Remove or add feature flags, etc.
- Create unit tests for a function
The announcement was made today by Blocks Open -Source program officeA significant step towards democratizing AI-powered systems for developers and beyond.
As Jackie Brosamer, VP from Data & Ai Platform Engineering at Block, in a video call interview recently, there is the goal of enabling users to create agents that can cut off several software providers, transport user data, summarize and accept action in Name of the user in several apps.
“I think the advantage of goose is its ability to work over different systems. It is not only limited to Google Drive – it can also integrate Google Drive into Slack, for example. One of our members of the business team finds it particularly useful to find out what they did last week because they are discussions in 40 hours and cannot remember. And so it is able to lock yourself up in one of these different systems instead of using an agent for Google, an agent for slack, and get married together. “
Google is now under a revealing available Apache 2.0 licenseAnd permits that it is used for a conceivable project from advertising to research.
The aim is to unlock innovative workflows by activating the interoperability between user interfaces, AI models and the systems with which they interact.
Gans are designed so that it should work with Anthropics Open -Source model context protocol (MCP)– A standardized Suite from APIS that enable connections to tools such as content repositories, business applications and development environments.
Structure of the blocks of an open AI ecosystem
According to Dhanji Prasanna, Chief Technology Officer (CTO) from Block, the open source publication is intended to inspire a wave of invention in the industries. “The production of geese open source creates a framework for new heights of invention and growth,” said Prasanna. He emphasized that the framework has already freed it from repeating tasks and enabled more effective work. “Since more people build with AI, we will unlock new ideas and advantages that together reduce the effort and give people time to be their most creative themselves,” he added.
The software is made by design modular and enables developers to integrate goose into their systems or to create custom interfaces.
With the open architecture of Goose, users can also select the LLM of their selection and offer flexibility for various applications.
Bradley Axen, Ai Tech Lead at Block, emphasized this adaptability and explained: “We built goose to work with every large-scale model, from Openai to open source options. This adaptability enables us to combine strengths of different providers for optimal performance. “
Beyond the engineering system
Goose’s first focus is on software engineering, in which it acts as an agent that can automate time-consuming tasks. In real time within development environments, it can be autonomous:
• Search and navigate code bases.
• Reading, writing and editing files.
• Install dependencies and carry out tests.
• Refine the outputs and do additional tasks with minimal supervision.
This flexibility has the potential to save engineers a lot of time and at the same time improve productivity. However, Block presents broader applications, from the fabrication of creative processes such as music composition to improving personalized e-commerce experiences.
Goose’s skills are improved by working through Block with Anthropic, a leading AI security and research company. Together they developed MCP, which enables goose to bring an interface with a growing ecosystem of tools and systems.
The open source nature of goose promotes the municipal business innovation. As Brosamer said: “The open source development enables an ecosystem in which new integrations can come from outside the company. Users do not have to wait for us to publish functions. You can innovate yourself. “
Axen underlined the philosophy behind Goose’s publication: “We do not monetize goose directly – it is completely licensed. Instead, we want to develop products that work seamlessly with agents and benefit from the open source ecosystem. “
With regard to privacy, security and considerations, the Open Source Nature enables flexibility, goose on rooms or in virtual private clouds.
“We definitely have nothing in the middle of goose use – no calls to our servers,” said Axen.
“You can host it anywhere,” said Brosamer. “And as a financial company that is very concerned about user data, this has always been one of the reasons why we are really interested in OpenSource models: instead of handing over our data or user data via the Internet, which is much connected to risks we bring the models to where the data is already hosted. ”
Further information, including installation and contribution leaders, visit Goose’s Github website.
Source link