Source: Xinzhiyuan
[Guide] GPT-4 Enhanced Copilot is here! Just now, GitHub released a new generation of code generation tool GitHub Copilot X, writing code is no longer a dream.
Microsoft is really crazy!
Last week, Microsoft just upgraded the Office family bucket with GPT-4. Before people could react, Microsoft came to copy the code again.
Today, GitHub officially announced: Copilot X, a new generation of code generation tool based on GPT-4, is here!
Less than two years after its release, GitHub Copilot has written 46% more code for more than 1 million developers and increased coding speed by 55%. And this heavy upgrade is even more powerful.
The new generation of Copilot X directly integrates a gorgeous chat interface, which can not only generate codes just by talking, but also talk to AI while writing codes, as well as an intelligent document system tailored for you, etc.
GPT-4 supports Copilot, you can write code with your mouth
Specifically, Copilot X supports the following experiences: Copilot Chat; Copilot for Pull Request; Copilot for Docs; Copilot for CLI.
These new features are all powered by OpenAI's GPT-4. It's worth noting that, due to speed delays, the code completion tool is still based on GitHub's Codex model, which is trained on GPT-3.
Now, Microsoft has integrated GPT-4 into various real productivity tools such as search, office, and code writing. It is true that it has the power to start the fourth technological revolution.
Zhu William, a senior big data architect, said that Everything powered by AI is no longer far away:
If it is said that OpenAI GPT-4 is just a model that has completed from 0 to 1, Microsoft has promoted its commercialization a big step forward.
This time, the release of Copilot X directly hits the previous generation of Copilot with dimensionality reduction.
Currently, Copilot X has opened the queuing entrance. But the more troublesome thing is that you have to apply for a queue for each experience.
A netizen joked about this: "Every day I am on the way to apply for Waitinglist."
——I just joined Bard's alternate yesterday, and I'm starting to line up for Copilot X again today.
Copilot Chat
Application address: https://github.com/github-copilot/chat_waitlist_signup/join
GitHub embeds a GPT-4-based chat window in Copilot, focusing on developer scenarios, and integrates it into VS Code and Visual Studio.
However, it's more than just a chat window.
Now, Copilot can not only identify the code content entered by the developer and display the error message, but also conduct in-depth analysis and explanation of the purpose of the code block, and generate unit tests.
You can even give debugging suggestions.
Plus, in Copilot, you don't even need a keyboard to write code anymore.
Just sit in front of the computer, say "Hey, GitHub!", and start coding with your mouth.
Currently, GitHub is experimenting with a new feature called GitHub Copilot Voice, a voice-based interaction system.
In addition to writing code, through Copilot Voice, you can even complete: code jump, control IDE, code summary.
Copilot for Pull Request
Application address: https://copilot4prs.githubnext.com/login
In addition to the editor, the new GitHub Copilot will also provide support for Pull Requests.
When the developer inserts a tag in the PR description, Copilot will automatically recognize it and perform extended supplementary writing.
The currently supported tags mainly include the following:
copilot:summary generates a summary summary for a PR.
copilot: walkthrough a detailed list of changes, including links to relevant code snippets.
copilot:poem Write a poem describing the change.
copilot:all automatically generates all of the above.
And this is just the first step.
In the future, Copilot will also add two major functions of Gentest and Ghost Text.
Gentest: When Copilot finds that the submitted Pull Request lacks enough tests, it will automatically remind the developer and suggest potential tests that can be edited, accepted or rejected according to project requirements.
For project owners, this feature also helps them develop policies around testing and supports developers in following those policies.
Ghost Text: When a developer creates a Pull Request, Copilot will dynamically obtain information about code changes and automatically give suggestions.
Additionally, other ongoing explorations and experiments at GitHub include:
Solving AI problems: It is possible to automatically describe how to solve a problem and even automatically suggest changes that need to be made.
As follows, after submitting a problem transferred to PyTorch, the new function can provide generated code suggestions and initiate a new Pull Request.
Review PR with AI: Since a large number of developers spend time on code review and preparing PR for code review, this feature can reduce the cycle time spent on review.
Copilot for Docs
Application address: https://githubnext.com/projects/copilot-for-docs
The importance of an excellent technical document to a developer is obviously self-evident.
The function of Copilot for Docs can help you quickly locate and find what you want from this document.
According to the official blog, Copilot for Docs is an experimental tool only. Programmers can use the chat interface to find content in the documentation, including questions about the language, framework, and technology used.
The more powerful place is this personalized feature.
It can output different answers according to the different programming levels of users, the degree of understanding of the document, and the content they want to find.
Programmers can also get the latest content as soon as possible, keeping existing materials in sync with the latest version of the documentation directly from the GitHub repo.
At the same time, another benefit of getting content from GitHub is that it makes it easy to answer questions about your team's private content. This feature can provide information that only your internal team can see based on the existing GitHub permissions.
In addition, Copilot can also shuttle back and forth between different third-party documents, such as React, Azure documents and MDN, can splice their content and return the result.
In the future, GitHub is also exploring other ways of indexing resources beyond documentation, such as issues, pull requests, discussions, and wikis, to give developers everything they need to answer technical questions.
Copilot for CLI
Application address: https://githubnext.com/projects/copilot-cli/
Next to code editing and pull requests, the command line terminal is where developers spend the most time.
However, even the most skilled developers need to scroll many pages to remember the precise syntax of many commands.
And the GitHub Copilot CLI can write commands and loops, and give vague find flags to satisfy programmers' queries.
Copilot CLI provides three shell commands: ??, git?, gh?.
"??" refers to the generic goto for arbitrary shell commands. It will compose commands and loops and throw vague lookup flags to satisfy individual queries.
"git?" is used to specifically search and invoke git. Compared to ??, it will be more powerful in generating Git commands, and you don't even need to explain when you are in the context of Git, and the query is more concise.
"gh?" will combine the power of the GitHub CLI command and query interface, and use AI to generate complex flags and jq expressions.
The golden age of software development! 10 times efficiency improvement
GitHub CEO Thomas Dohmke said that although automatic code completion has greatly improved developer productivity, the new Copilot X can increase developer productivity by 10 times.
With the new chat feature, Copilot will be the foundation for people learning to code in the future.
"The golden age of software development has come to chat interfaces."
“The 10x engineer is still a myth, but the 100x engineer powered by AI is more real than ever,” says Nvidia scientist Jim Fan.
Given that GitHub "mixed and matched" OpenAI's GPT-3 and GPT-4 to drive the chat function and the existing auto-completion code function.
Dohmke explained that when you type in an editor, you expect a very fast response.
Therefore, where speed is required, we use smaller models, such as code auto-completion based on Codex; and where accuracy is required, we use larger models, such as GPT-4 in chat.
Netizen: The code farmer is "dead"
Netizens have expressed that programming has been changed forever by ChatGPT.
"As a developer, this is really frustrating. In programming, the source of our joy is playing with logic and solving technical problems, rather than focusing on productivity or meeting needs."
“I’ve been programming for 30 years and I still love it. I don’t know what the world would look like without human code. I do like having CoPilot for the annoying stuff though, and ChatGPT to help me out Fix the bug."
"After the emergence of AI, there may be no need for so many code farmers. I am very anxious about this. Even though I have saved money and invested in various ways, it will take almost 9 years to "lay flat". I am a little envious Those who work in FANG, they seem to be able to achieve financial freedom easily."
"In the medium to long term, we're going to lose jobs."
Programmers finally eliminated themselves with programming...
References:
https://github.blog/2023-03-22-github-copilot-x-the-ai-powered-developer-experience/
END
Welcome to join Imagination GPU and artificial intelligence communication group 2
Join the group, please add the editor WeChat: eetrend89
(Please add company name and title)
recommended reading
Dialogue with the Chairman of Imagination China: Using GPU as the fulcrum to strengthen software and hardware collaboration to facilitate digital transformation
Imagination Technologies is a UK-based company dedicated to the research and development of chips and software intellectual property (IP). Products based on Imagination IP are used in the phones, cars, homes and workplaces of billions of people around the world. For more information on cutting-edge technologies such as the Internet of Things, smart wearables, communications, automotive electronics, and graphics and image development, welcome to Imagination Tech!