EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Aug 19·5 min read
Technical leaders should have the largest AI exhaust
TL;DR
Measuring engineers by AI exhaust, such as lines of code or tokens spent, fails to capture their true impact. Senior engineers should focus on enabling others and leveraging their influence rather than just coding output.
✦ Why It Matters
Reassess how you measure your team's contributions, focusing on impact rather than just coding output.
Key Takeaways
- Engage directly with AI tools to understand their capabilities.
- Experiment with coding agents to inform decision-making.
- Review code changes critically, even when using AI.
- Maintain a greppable codebase for effective agent context.
Full Summary
Measuring engineers by traditional metrics such as lines of code or tokens burned is misleading, as these do not equate to real impact. Senior engineers are expected to leverage their influence through mentoring and strategic decisions rather than just coding.
However, the rise of AI tools has disrupted this norm, necessitating that technical leaders gain firsthand experience with coding agents to understand their capabilities and limitations. Open questions around code review, context management, and agent autonomy highlight the need for experimentation.
The author emphasizes that high AI exhaust, resulting from rigorous testing and exploration of new tools, is a sign of effective leadership rather than a performance target. This shift in practice requires leaders to remain closely involved in the coding process to maintain their judgment and influence.