Other names | Devin AI |
---|---|
Developer(s) | Cognition Labs |
Available in |
|
License | Proprietary (SaaS), using open source software (Ubuntu) in part |
Website | preview |
Devin AI is an autonomous artificial intelligence assistant tool created by Cognition Labs. Branded as an "AI software developer", [1] the demo tool is notable for its software development abilities, including plan implementation, source code generation, and benchmark unit testing. The tool has received praise, concern, and skepticism over implications surrounding the future of artificial intelligence and software development.
Devin AI was created by Cognition Labs, a startup company consisting of ten members including CEO Scott Wu and chief technology officer Steven Hao, with funding from Peter Thiel's Founders Fund firm. [2] [3] Several of the members had participated in competitive coding contests before forming the company. [3] The members developed the software via a combination of training large language models akin to OpenAI's GPT-4 with aspects from reinforcement learning. [3] According to a Bloomberg article, Cognition Labs claimed that Devin AI represents a "breakthrough in a computer's ability to reason." [3] Devin AI has also been considered part of a trend surrounding the advent of autonomous AI agents that can take direct action to solve problems. [1]
Devin AI has been noted for its ability to perform software engineering tasks autonomously. [4] [5] Compared to the GitHub Copilot tool, [3] [4] the software can code, debug, plan and problem solve via machine learning techniques. [5] Devin AI works through a user prompting the software with a task in natural language, with the software responding by showing its plan while implementing the code. [3] It searches online resources during the process to learn to complete a task. [4] The software also takes prompts from users during the implementation process and adjusts its plans accordingly, such as when a user notices an issue or bug. [3] [6]
One application of Devin AI is website creation. A test conducted by Bloomberg revealed that the tool could create a website within ten minutes and could recreate a Pong website in a similar timeframe. [3] In a demo from Cognition Labs, the tool also created a website based on the Llama 2 language model through plan, source code and benchmark testing generation. [1] Other examples include building a project to display images from a blog post, and compiling a computer vision model from an Upwork project. [6] In a benchmark test for analyzing the performance of large language models on real world projects, Devin was found to fix 13.86 percent of encountered issues with no human assistance, compared to an average of 1.96 percent and 4.8 percent for an unassisted and assisted model, respectively. [5] [6]
Later revisions of Devin got multi-agent operation capability, where one of the AI agents dispatch task to other AI agents: [7]
Devin AI has been met with praise, concern and skepticism from journalists and software engineers. [1] [8] [9] Its announcement on X led to praise from investors and software engineers while spawning various memes. [1] Along with the company, the tool has seen optimism amongst AI enthusiasts and anticipation for its public availability. [3] The tool has also been noted for potentially allowing users of a non-technical background to create projects, and aiding developers in solving more complex tasks. [4] The Indian Express claimed that its capabilities could streamline the software development process while avoiding human error. [5] CEO Aravind Srinivas of Perplexity.ai offered praise to Devin, claiming that it "seemed to be 'the first demo of any agent, leave alone coding, that seems to cross the threshold' of human capability." [8] After the release of Devin AI, Cognition Labs experienced increasing growth and interest. Earlier this year, the startup raised $21 million in a deal valuing it at $350 million. It then turned down offers valuing it at $1 billion. According to the Wall Street Journal , the company has been in talks with investors for a deal that would value it at up to $2 billion. [10]
Concern for the software includes its implications for the future of AI and the software development industry. [3] [8] In the wake of layoffs within the tech industry throughout 2023 and 2024, [8] discourse of the tool involves concerns that it may replace engineers and remove lower-level jobs. [4] On social media, various developers expressed criticism for the software's capabilities and potential to incite job layoffs. [1] [8] [9] Skepticism also emerged that the tool may struggle to complete tasks with more intricate requirements and scenarios that would necessitate human creativity, along with its efficiency. [5] [8] Further skepticism regarding its accuracy has emerged following the tool's promotional videos, such as its performance of Devin AI's execution of the Upwork project; YouTube channels such as Internet of Bugs and Computer Vision Project criticized the tool for failing to deliver on the project request, instead writing, testing, and debugging code irrelevant to the Upwork request. [11] However, the tool has also been regarded to encourage software engineers to perform more creative work. [3] [5] Following Devin's debut, various AI software engineering models have been released, such as free and open source replacements like OpenDevin (now called OpenHands) [12] and Devika, [13] and Genie by San Francisco-based startup Cosine. [14]
Synopsys, Inc. is an American electronic design automation (EDA) company headquartered in Sunnyvale, California, that focuses on silicon design and verification, silicon intellectual property and software security and quality. Synopsys supplies tools and services to the semiconductor design and manufacturing industry. Products include tools for logic synthesis and physical design of integrated circuits, simulators for development, and debugging environments that assist in the design of the logic for chips and computer systems. As of 2023, the company is a component of both the Nasdaq-100 and S&P 500 indices.
Business process automation (BPA), also known as business automation, refers to the technology-enabled automation of business processes.
Recursive self-improvement (RSI) is a process in which an early or weak artificial general intelligence (AGI) system enhances its own capabilities and intelligence without human intervention, leading to a superintelligence or intelligence explosion.
Databricks, Inc. is a global data, analytics, and artificial intelligence company founded by the original creators of Apache Spark.
GitLab Inc. is an open-core company that operates GitLab, a DevOps software package that can develop, secure, and operate software. GitLab includes a distributed version control based on Git, including features such as access control, bug tracking, software feature requests, task management, and wikis for every project, as well as snippets.
OpenAI is an American artificial intelligence (AI) research organization founded in December 2015 and headquartered in San Francisco, California. Its mission is to develop "safe and beneficial" artificial general intelligence (AGI), which it defines as "highly autonomous systems that outperform humans at most economically valuable work". As a leading organization in the ongoing AI boom, OpenAI is known for the GPT family of large language models, the DALL-E series of text-to-image models, and a text-to-video model named Sora. Its release of ChatGPT in November 2022 has been credited with catalyzing widespread interest in generative AI.
UiPath Inc. is a global software company that makes robotic process automation (RPA) software. It was founded in Bucharest, Romania, by Daniel Dines and Marius Tîrcă. Its headquarters are in New York City. The company's software monitors user activity to automate repetitive front and back office tasks, including those performed using other business software such as customer relationship management or enterprise resource planning (ERP) software.
Opentrons Labworks, Inc. is a biotechnology company that manufactures liquid handling robots that use open-source software, which at one point used open-source hardware but no longer does. Their robots can be used by scientists to manipulate small volumes of liquids for the purpose of undertaking biochemical or chemical reactions. Currently, they offer the OT-2 and Flex robots. These robots are used primarily by researchers and scientists interested in DIY biology, but they are increasingly being used by other biologists.
Generative Pre-trained Transformer 3 (GPT-3) is a large language model released by OpenAI in 2020.
GitHub Copilot is a code completion and automatic programming tool developed by GitHub and OpenAI that assists users of Visual Studio Code, Visual Studio, Neovim, and JetBrains integrated development environments (IDEs) by autocompleting code. Currently available by subscription to individual developers and to businesses, the generative artificial intelligence software was first announced by GitHub on 29 June 2021, and works best for users coding in Python, JavaScript, TypeScript, Ruby, and Go. In March 2023 GitHub announced plans for "Copilot X", which will incorporate a chatbot based on GPT-4, as well as support for voice commands, into Copilot.
ChatGPT is a generative artificial intelligence chatbot developed by OpenAI. Launched in 2022 based on the GPT-3.5 large language model (LLM), it was later updated to use the GPT-4 architecture. ChatGPT can generate human-like conversational responses and enables users to refine and steer a conversation towards a desired length, format, style, level of detail, and language. It is credited with accelerating the AI boom, which has led to ongoing rapid investment in and public attention to the field of artificial intelligence (AI). Some observers raised concern about the potential of ChatGPT and similar programs to displace or atrophy human intelligence, enable plagiarism, or fuel misinformation.
Generative Pre-trained Transformer 4 (GPT-4) is a multimodal large language model created by OpenAI, and the fourth in its series of GPT foundation models. It was launched on March 14, 2023, and made publicly available via the paid chatbot product ChatGPT Plus, via OpenAI's API, and via the free chatbot Microsoft Copilot. As a transformer-based model, GPT-4 uses a paradigm where pre-training using both public data and "data licensed from third-party providers" is used to predict the next token. After this step, the model was then fine-tuned with reinforcement learning feedback from humans and AI for human alignment and policy compliance.
Microsoft Copilot is a generative artificial intelligence chatbot developed by Microsoft. Based on the GPT-4 series of large language models, it was launched in 2023 as Microsoft's primary replacement for the discontinued Cortana.
AutoGPT is an open-source "AI agent" that, given a goal in natural language, will attempt to achieve it by breaking it into sub-tasks and using the Internet and other tools in an automatic loop. It uses OpenAI's GPT-4 or GPT-3.5 APIs, and is among the first examples of an application using GPT-4 to perform autonomous tasks.
Tabnine is an artificial intelligence (AI) coding assistant developed by Tabnine, which was founded by Dror Weiss and Professor Eran Yahav in Tel Aviv, Israel, in 2013. Initially established under the name Codota, the company underwent a rebranding in May 2021 following the release of the company’s first large language model based AI coding assistant, adopting the name Tabnine.
Mistral AI is a French company specializing in artificial intelligence (AI) products. Founded in April 2023 by former employees of Meta Platforms and Google DeepMind, the company has quickly risen to prominence in the AI sector.
Claude is a family of large language models developed by Anthropic. The first model was released in March 2023. Claude 3, released in March 2024, can also analyze images.
Cognition AI, Inc., doing business as Cognition, is an artificial intelligence (AI) company headquartered in San Francisco in the US State of California. The company developed Devin AI, an AI software developer.
OpenAI o1 is a generative pre-trained transformer released by OpenAI in September 2024. o1 spends time "thinking" before it answers, making it more efficient in complex reasoning tasks, science and programming.