Jev analyzed 1,700 PRs for 9 cents
A developer used AI (Gemini) to analyze 1,700 pull requests in 2 minutes for just 9 cents, extracting work allocation data across initiatives. This demonstrates how AI can help CTOs and CEOs quantify what percentage of engineering effort goes toward different products or projects.
Summary
The speaker describes a cost-effective analysis of pull requests using AI technology. The analysis processed 1,700 PRs in approximately 2 minutes at a cost of only 9 cents. The methodology involved creating 17,000 matching pairs from these PRs and then using Google's Gemini AI to tag themes and extract relevant data. The speaker then addresses CTOs and CEOs directly, highlighting a common challenge in executive leadership: board members, teams, and other leaders frequently request visibility into work distribution—specifically, what percentage of engineering effort is allocated to different initiatives or products. The speaker uses this example to illustrate how AI-powered analysis can provide quantifiable answers to these organizational questions, enabling data-driven insights into resource allocation and project prioritization across multiple product lines.
Key Insights
- Analyzing 1,700 pull requests with AI costs only 9 cents and completes in 2 minutes
- The analysis created 17,000 matching pairs from PRs for thematic classification
- Gemini was used to automatically tag themes and extract structured data from PR data
- CTOs and CEOs face consistent pressure from boards and leadership to quantify work distribution across initiatives
- AI analysis can answer executive questions about percentage of work allocated to different products
Topics
Transcript
[0:00] It cost me 9 cents. This took about 2 minutes. He analyzed 1,700 revision requests (PRs). He created 17,000 pairs in these PRs for matching, then the themes were tagged using Gemini, and he extracted all of this data. CTOs and CEOs , I know your board , your team, and your leadership are constantly asking you what percentage of work goes to which initiatives? Tell me the percentage for this [0:30] product and another product.
Full transcript available for MurmurCast members
Sign Up to AccessMore from How I AI
I tried Muse, Meta's new AI agent (meet Slime 🦖)
The speaker reviews Muse, Meta's new AI agent, highlighting its ability to perform browser-based tasks, connect to multiple data sources, and help users achieve personal goals through an approachable interface. Key features include connectors to email/calendar/health data, idea suggestions, artifact creation, and customizable avatars.
I’m using Jev more than Opus 5.5 or GPT-6. Here’s why.
The speaker demonstrates why they're using Jev, a fast and inexpensive decision-making model from Type-Safe AI, more than other recent models like Opus 5.5 and GPT-6. Jev specializes in classification, clustering, and real-time decision-making tasks at a fraction of the cost of traditional LLMs, enabling complex data analysis and product features that would have been prohibitively expensive before.
Warp agents open PRs to fix the factory itself
Programming agents can autonomously improve factory systems by analyzing failed launches and proposing specific updates to agent definitions. A self-improvement loop enables observer agents to detect failures and generate evidence-based modifications that prevent recurring issues, such as changing specific steps in factory agent procedures.
Humans are still the bottleneck in Warp’s AI factory
Warp discusses how human code review has become the main bottleneck in their AI-assisted software development process, with a 3.5-hour delay from PR to first human review compared to 35 minutes from launch to PR. They're evolving their workflow to reduce human dependency by allowing requesters to review agent-generated code themselves, and plan to eventually skip review entirely for low-risk tasks by treating code review as a risk management exercise.
I Quit Claude Because It Was Annoying
The speaker explains why they stopped using Claude, citing frustrations with its tendency to produce nonsensical output and communicate in an unnatural, non-human manner. They mention considering a switch to Opus 5.5 but remain uncertain about fully migrating their work tasks.