2023年3月23日 Github Copilot X 來了!Github Copilot X is here!

什麼是GitHub Copilot?

寫過程式的都知道,寫程式就是把自己頭腦的邏輯說給電腦聽,讓電腦可以聽懂,但電腦不像人,容錯率很低,所以有時候邏輯沒問題,但電腦就是不跑,可能是少寫一個字母,可能是多空一格,不管事多大多小,看不懂,不跑就是不跑。若程式行數少,慢慢找,總有找到的一天,但若程式龐大,那真是很想哭,我就曾為了寫錯一個字,找了好幾天都找不到,最後還是請別人幫忙找,才找到那個錯字,大概是因為自己打的字,怎麼看都覺得是對的,哈!

為了解決這種在寫作程式中,本不該耗掉程式開發人員時間和精力的大小事, GitHub開始嘗試將大型語言模型導入程式寫作中,看能不能解決一些問題,讓程式開發人員不但寫得快,還寫得好又輕鬆。因此他們開發了世界上第一個大規模生成式 AI 開發工具,它採用 OpenAI 的 Codex 模型,源自 GPT-3 的技術,稱作 GitHub Copilot (https://github.com/features/copilot)。

GitHub Copilot 讓 AI 透過自動補全註釋和代碼,讓程式開發人員感覺上天堂。根據統計,目前 GitHub Copilot 已經能夠撰寫 46% 的代碼,並幫助開發者提高 55% 的編碼速度,實在驚人,省下的時間都可以多睡一覺了!

什麼是GitHub Copilot X?

2023年3月23日 Github Copilot X 來了!Github Copilot X is here!

“X”通常代表著某種未知,但會希望出現的東西。對 GitHub 而言,發展出 GitHub Copilot 當然不是終點,若 Copilot 要能做得更好,那他們得在工作流程上提供更多幫助,讓自己成為開發生命週期中,那個最為可靠 AI 助手。為此,新的 GitHub Copilot X 採用新的 GPT-4 模型,還加入聊天和語音功能,讓開發人員可以在開發流程中,隨時請 Copilot X 回答相關問題,意思是這回你等於有個資深、內行的前輩站在你身旁,有什麼問題,問 Copilot X 就對了,如果你無法想像,那就把它想成是程式編碼裡的 ChatGPT。

這有什麼好處?

想像一下,若未來的軟體開發人員都能因 AI 的協助,把那些日常瑣事都交給 GitHub Copilot X 處理,只要你想得出來的創意,GitHub Copilot X都能幫你跟電腦溝通,此時,推展出更新創意,而不是學習如何和電腦溝通,成為人們最重要的目標。若人類能將所有精力都專注在創意發想上,那人類的進步速度可能會達到前所未有的速度。

要如何試用?

若想率先使用 GitHub Copilot X 即將推出的功能?那你要去這裡(https://github.com/features/preview),然後選擇你想要的功能,開始排隊。

2023年3月23日 Github Copilot X 來了!Github Copilot X is here!
2023年3月23日 Github Copilot X 來了!Github Copilot X is here!

如果你是新來的,要先註冊。

2023年3月23日 Github Copilot X 來了!Github Copilot X is here!

2023年3月23日 Github Copilot X 來了!Github Copilot X is here!

2023年3月23日 Github Copilot X 來了!Github Copilot X is here!

因為這是個未來計畫,所以排隊不代表你什麼時候一定用得到,功能可能也不完整,因為試用的都算是 beta 版,僅供嘗鮮。

如果你真的有需求,可以考慮成為個人用戶,目前每月訂閱價格為10 美元、一次訂一年比較便宜,只要100 美元。如果你是教育人士(https://docs.github.com/zh/education/explore-the-benefits-of-teaching-and-learning-with-github-education/github-global-campus-for-students/apply-to-github-global-campus-as-a-student),可以再便宜一點。

2023年3月23日 Github Copilot X 來了!Github Copilot X is here!

What is GitHub Copilot?

For anyone who has written code, you know that writing code is all about expressing your brain’s logic to a computer so it can understand your brain’s logic. However, computers don’t have the same ability to understand context or make logical leaps that humans do, which means that sometimes, a small mistake can cause a program not to run. This can be frustrating, especially when debugging large codebases. To help solve this issue, GitHub started experimenting with large language models to see if AI could be used to assist with programming. This led to the development of GitHub Copilot, the world’s first large-scale generative AI development tool powered by OpenAI’s Codex model, a descendant of GPT-3 (https://github.com/features/copilot).

GitHub Copilot assists developers by auto-completing comments and code. According to statistics, GitHub Copilot is already writing 46% of code and helps developers code up to 55% faster, which is impressive and saves valuable time.

What is GitHub Copilot X?

The “X” usually represents an unknown or something to be desired. For GitHub, developing GitHub Copilot is just the beginning, as they hope to provide more assistance in the development workflow and create the most reliable AI assistant in the development lifecycle. For this reason, the new GitHub Copilot X is using the new GPT-4 model and adding chat and voice features, allowing developers to ask Copilot X questions throughout the development process. It’s like having a senior, experienced mentor standing next to you in the coding process, ready to answer any questions you may have. If you can’t imagine it, think of it as a ChatGPT for coding.

What are the benefits?

Imagine if future software developers could use AI to handle all their day-to-day tasks with GitHub Copilot X. Instead of learning how to communicate with a computer, developers could focus on their creativity and push out new ideas. If humans could concentrate all their energy on creating innovative ideas, the progress we could achieve might be unprecedented.

How can I try it out?

If you want to try out the upcoming features of GitHub Copilot X, you can sign up to be notified here (https://github.com/features/preview). Keep in mind that this is a future plan, and signing up does not guarantee access to each feature, and there is no guaranteed timeline for access. Once you get access to preview a feature, please note that it is considered beta or a technical preview, meaning the feature may still have some kinks to work out. If you have a real need for it, you can consider subscribing as an individual user, which is currently priced at $10 per month, or $100 for a yearly subscription. If you are an educator (https://docs.github.com/zh/education/explore-the-benefits-of-teaching-and-learning-with-github-education/github-global-campus-for-students/apply-to-github-global-campus-as-a-student), it is even cheaper.

發佈留言

發佈留言必須填寫的電子郵件地址不會公開。 必填欄位標示為 *