I'll admit I've been using using JetBrains copilot equivalent to right click and generate unit tests lately and the unit tests it created actually favored a contextual understanding of the code enough to write tests to match my intent rather than mechanics and critically actually catch bugs in my actual handwritten code. I do worry a lot about junior devs over using AI and making themselves useless hires because of it but I think in the hands of a responsible senior+ dev or particularly thoughtful mid level dev I think a Human+AI strategy can be a real productivity gain.
Matthew Beatty’s Post
More Relevant Posts
-
🚀 Introducing Devin – The World's First Fully Autonomous AI Software Engineer! Engineers, meet your new AI partner, Devin. Designed to work independently or alongside you, Devin elevates your engineering potential, allowing teams to tackle more complex challenges by automating thousands of coding decisions and tasks. ✨ Key Features: Advanced long-term reasoning and planning Equipped with common developer tools within a sandbox environment Real-time collaboration and iterative learning Bug detection and resolution capabilities With Devin’s debut, it might just be the perfect time to consider joining farmers or tradespeople, because it seems the coding fields are about to be harvested by AI!"
Introducing Devin, the first AI software engineer
https://www.youtube.com/
-
Meet Devin AI, the groundbreaking autonomous software engineer revolutionizing coding! 🚀 What's your take on AI's role in software development? Join the conversation and share your thoughts on Devin AI's impact. #DevinAI #FutureofCoding #AIAssistance Christophe Zoghbi
-
IT Leader | Consulting, Solutions, Program, Technology and People | Deliver Data Engineering, Machine Learning, Artificial Intelligence and MLOps Programs | Generative AI | On Cloud (Azure, GCP, AWS) | Coding and Kaggle
Just this week I talked about Pair programming with AI using GitHub Copilot and Cursor and both are good; and in my view GitHub Copilot is one of the best Business Use Case of AI so far... And within a week I came across Cosine Genie - I am on the waitlist and I have not tried it yet... As per the website, "Genie is the best AI software engineer in the world by far - achieving a 30% eval score on the industry standard benchmark SWE-Bench." This is definitely an interesting space to watch... And to clarify we still need the best Solution Architects and the best Software Engineers who can understand and guide and stitch the code generated by AI to make a Business solution, else there will be tons of useless code. NB : Hope we all remember the last CrowdStrike outage by a missing a if null statement... https://cosine.sh/genie
-
Just realized something about the Gen AI coding hype... Everyone's talking about how fast we can write code now. But that's not the real bottleneck anymore. The metric that actually matters? Time-to-Preview. If your devs can't test their changes in a prod-like environment within minutes, all that AI-generated code is just creating a bigger backlog. That's exactly why we ended up creating a solution to make previews instant. Seeing teams go from 30-min preview times to under 30 seconds has been wild. Fast previews = fast learning = better code. The future of dev productivity isn't just about writing code faster - it's about learning faster. #engineering #devtools #previewenvironments
-
-
Shaping Enterprise AI operations | Cloud & AI Engineering Leader @ LexisNexis | Innovating AI/ML Automation & Cloud Infrastructure | Harvard Business Leader
Worlds first AGI software engineer is here! Which SHOCKED The ENTIRE INDUSTRY Cognition just dropped 𝐃𝐞𝐯𝐢𝐧, first AI Software Engineer 𝐇𝐞𝐫𝐞 𝐚𝐫𝐞 6 𝐢𝐧𝐜𝐫𝐞𝐝𝐢𝐛𝐥𝐞 𝐭𝐡𝐢𝐧𝐠𝐬 𝐢𝐭 𝐜𝐚𝐧 𝐝𝐨: Devin is a fully autonomous AI software engineer tackling engineering tasks with its own shell, code editor, and web browser. 1. Devin quickly masters new technologies. 2. It can contribute to mature production repositories. 3. Can build and deploy apps end to end. 4. Devin can train and fine tune its own AI models. 5. It completed a job posting on upwork as well. 6. Address bugs and feature requests in open source repositories. Their team evaluated Devin on SWE-bench, a challenging benchmark that asks agents to resolve real-world GitHub issues found in open source projects like Django and scikit-learn. Devin correctly resolves 13.86% of the issues end-to-end, far exceeding the previous state-of-the-art of 1.96%. Even when given the exact files to edit, the best previous models can only resolve 4.80% of issues. What are your thoughts on the same? #AI #softwareengineering #genai
Introducing Devin, the first AI software engineer
https://www.youtube.com/
-
👉 Meet #Devin: The Game-Changer in AI Software Engineering 🚀💻 Dive into the future of software development with Devin, the first fully autonomous #AI software engineer by #Cognition. From coding to website creation, #Devin redefines the benchmarks of software engineering with unparalleled efficiency. Swipe right to explore how Devin is transforming the landscape of AI-driven software solutions. 🚀💻 #AI #DSInsights #InnovateWithDevin #SoftwareEngineering #Innovation
-
Senior Systems Engineer - Cloud DevOps @ Infosys | Cloud & DevOps Practice | 4x AWS Certified, 1x Azure | Infosys Certified in Cloud, DevOps Professional, Prompt Engineering |
👩💻🤖 Calling all software engineers! Have you heard about Devin.ai? It's the latest AI sensation in the market. With Devin.ai around, coding, debugging, error-solving, and deployment are a breeze! It's like having a tech wizard on your team 24/7. 🌟✨ Just remember, even with all this AI magic, there's one thing Devin.ai can't do... bring you coffee! ☕️😉 #DevinAI #SoftwareEngineering #aihumor
-
Meet Devin, Fully autonomous AI software engineer. Devin is a tireless, skilled teammate, equally ready to build alongside you or independently complete tasks for you to review Devin can plan and execute complex engineering tasks requiring thousands of decisions. Devin can recall relevant context at every step, learn over time, and fix mistakes
Blog
cognition-labs.com
-
Hot take on AI (LLMs, to be nerdy) and software engineering: If an AI assistant/copilot successfully predicts the code you are going to write, probably this code shouldn't be written in the first place. This piece of code should probably be a library or a framework. Either find an existing one and stop writing this boilerplate, or refactor into one
-
Developer marketing & business development at Growth Acceleration Partners | Volunteer: NET Foundation, Microsoft Alumni Network, Women in Cloud | #womenintech #dotnet #Microsoft
Code Whisperer or Code Assistant? 🤖 Our engineers weigh in on GitHub Copilot's impact on speed, quality, and more. It's a mixed bag, but the future of coding is definitely AI-powered. Read our full report: https://hubs.li/Q02Nw07P0 #code #coding #copilot #AI
-
Building ConnectHV for the Hudson Valley / Principal Product Designer at Okta
1moI've been using Cursor to generate tests for my side project and it's been impressively good with understanding my intent rather than what I've actually written! It's caught a bunch of thorny bugs that would've taken me a while to notice. I 100% agree, though — it's effective because it's stuff I could do myself, so it's just accelerating my work rather than doing stuff that's totally new to me. The upcoming generation of engineers are going to have to fight hard against the temptation to outsource all their thinking to AI. (Your post also made me think of this line I read recently, not sure who to attribute it to: "AI is not going to replace humans, but humans with AI are going to replace humans without AI.")