Edition 8 - AI Assisted Coding

Author

Mohammed Ali Chherawalla
April 24, 2024

Sponsored by

Read time: 3 mins

AI is not just a trend; it's a powerful tool that's here to stay. My first encounter with large-scale AI and Machine Learning was back in 2016-2017, long before ChatGPT became popular. One notable project we worked on reduced 14 hours of effort to just 45 minutes in document analysis. You can read about it here.
In this project, the AI classifier identifies key paragraphs, and a human auditor selects the most relevant ones, streamlining the process significantly.

Before moving ahead, I’m very excited to share that this edition is sponsored by The Rundown.

The Rundown is the world’s fastest-growing AI newsletter, with over 500,000+ readers staying up-to-date with the latest AI news and learning how to apply it.

Our research team spends all day learning what’s new in AI, then distills the most important developments into one free email every morning.

The pattern from that project is now prevalent even in programming, massively enhancing productivity when used correctly. It applies to various aspects of what we do like;

  1. Big Features, System Design

  2. Small Features

  3. Refactoring

  4. Bug Fixing

  5. Code Quality

  6. Testing

Different AI tools excel in different areas:

  • ChatGPT:

    • Bug Fixes: Acts as a modern Stack Overflow, providing decent solutions to error messages.

    • Code Quality: Generates good documentation for small functions. It's okay at refactoring small code changes but can be unreliable.

    • Features: Good at writing small features and tests but requires very specific instructions.

  • GitHub Copilot:

    • Code Quality: Often produces high-quality code recommendations.

    • Small Features: Can generate code that matches detailed comments, but be cautious of auto-completed comments.

  • CodeRabbit:

    • Code Quality: Excellent for code quality improvement before merging. It's great at identifying inconsistencies and suggesting better approaches in PRs.

  • pre.dev:

    • Big Features, System Design: Useful for initial system design in greenfield projects, but its reliability decreases as system complexity increases.

  • JetBrains AI

    • Testing: Perhaps the best there is out there but it doesn’t really solve the itch. I lean more towards tools like Keploy, & PlayWright that record user actions and create test cases rather than depending only on AI to write test cases.

The key is to use AI to improve your coding abilities, not to replace them. If you find the right balance, you can use the power of AI-powered coding while still using your creativity and problem-solving skills that make you human.

In the end, using these technologies wisely and ethically is the best approach. The potential is huge, but you have to make sure you are in control, not the other way around.

DATA TALES

Meta has an AI tool called SapFix, which can automatically generate fixes for bugs identified by Sapienz, their intelligent testing system. SapFix can suggest multiple fixes for a bug, which developers can then review and approve, streamlining the bug-fixing process. This demonstrates AI's role in reducing the time developers spend on debugging. Read more about SapFix here.

WHAT I’M READING

I'm currently reading 'No Rules Rules' by Reed Hastings and Erin Meyer, a fascinating read on the unconventional culture and management philosophy behind Netflix's remarkable success. The book offers insightful strategies for cultivating an innovative, freedom-oriented workplace - valuable lessons for any entrepreneur aiming to build a high-performing, agile organization.

HAPPENINGS AT WEDNESDAY

  1. We hosted TechMeet Pune at our HQ. It was a banger with lots of knowledge-sharing and networking. Experience here.

  2. Our new Clutch page (in Wednesday style) is now live. Take a look here. 
    P.S. Want a free week of work? Reply to this email.

EARN REWARDS⚡️


You can get rewards for referring your network to LeadReads 👇️ 

1 referral - LinkedIn Shoutout from Wednesday
3 referrals - Our Data Engineering Onboarding Kit
5 referrals - Our Service Picker Tool (A must-have for developers)

{{rp_personalized_text}}

Or copy & paste this link to your friends: {{rp_refer_url}}

Hi, I’m Mohammed Ali Chherawalla (Mac), Co-founder & CTO at Wednesday Solutions, a specialized engineering services company with a focus on Applied AI, Data, and Application Modernization. 10% of India's Unicorns are our customers. I make it a point to read every message from my subscribers, so don't hesitate to share your thoughts with me.