Devin, The first-ever AI Software engineer.

Devin, The first-ever AI Software engineer.

This might be the Biggest AI news this week, and it's shocking what do the future holds. Imagine our students still learning the basics of C and HTML, ooh shit.


Yesterday, A startup called Cognition Labs just dropped the world's first AI software engineer:

? It's called Devin and it can write complete apps end to end all by itself.

? In a demo, Devin was able to complete real jobs posted on Upwork.

? It also correctly resolved almost 14% of GitHub issues found in real-world open-source projects (better than many developers)


The demo release was made just yesterday, on their account Twitter page

It got more than 18.9 Million Views in 24 Hours.

Shortly, Devin is an agent that solves software engineering tasks through the use of its own shell, code editor, and web browser.

Devin has shown outstanding performance in some few things.


  • First, contributing to open-source projects.Just a link to a GitHub repository, Devin can make setups and gathering that are required.
  • Second, Fine-tuning large language models.
  • Third, Tested and proved to complete software engineering works in Upwork.

It was even tried in real jobs and has done all those too.


Check the demo Here

Amazing performance than GPT's in SWE-Bench benchmark.

SWE-Bench is a benchmark and dataset designed to evaluate the coding capabilities of large language models on real-world software issues collected from GitHub. It consists of 2,294 Issue-Pull Request pairs from popular repositories, serving as a testbed for assessing the ability of systems to automatically solve GitHub issues. This benchmark aims to test the next generation of language models in resolving practical coding challenges within the context of complex codebases.


Being unassisted, Devin was evaluated on a random 25% of the dataset and it correctly solved 13.86% of the issues end-to-end and far exceeded the performance of other famous and publicly known AI models including GPT-4


WHO ARE THE MAKERS?

It's natural to be curious about the people behind the work. Honestly, the job was done by highly skilled and competitive software engineers. They have extensive experience working in the field of Applied AI. Here are a few of them:


Scott Wu Co-founder and CEO. He is the mastermind behind Cognition turning Ideas into reality. Did an Economics from Harvard and was previously Software engineer addepar and Co-founder and CTO of Lunchclub.

SILAS ALBERT Math Geek in AI Research | Ph.D Student in AI from Stanford University.

STEVEN HAO Human Sofware Engineer previously worked at Scale AI and student from MIT doing Computer science and Mathematics.


ECNERWALA Software Engineer || Top Competitive Programmer doing a lot of work in Github , YouTube and CodeForces.


Tony Jiang Human Software engineer previously a student at Yale University doing Computer Science and Mathematics.


Jeffrey Ling Human Software engineer previously working on AI from Google and Waymo . Done his BA in computer science and Math from Harvard University.

Last but not least,

Neal Wu Previously From Google Brain and Harvard Computer Science. Featured for competitive programming at Wired and doing competitive programming in YouTube


There's an awesome team behind Cognition who makes difficult things seem easy, above are just few. Imagine if today we have AI engineers who can solve complex computer programming tasks effortlessly, what do you think the future will look like in the next 10 years?


ABOUT COGNITION?

This is small team based in New York and the San Francisco Bay Area, an applied AI?lab focused on reasoning. They believe by solving reasoning, we can unlock new possibilities in a wide range of disciplines.

Code is just the beginning.

This was really exciting, we couldn't wait to talk about it. What do you think about the future of work in the coming years?

要查看或添加评论,请登录

社区洞察

其他会员也浏览了