Try for free
The Merge (by CodeRabbit)
The Merge (by CodeRabbit)

We still don't understand LLMs today.

20 August 2026 29:44 CodeRabbit

Listen to episode

About this episode

How do LLMs actually work—and why do they remain black boxes?
In this episode of Merge, Shriyash “Yash” Upadhyay, co-founder of Martian, joins CodeRabbit to explore LLM interpretability, AI research, code review benchmarks, and the search for the “steam engine of AI.”

We discuss:
- Why we still don’t fully understand how LLMs work
- How Martian is researching machine intelligence
- Why code review is a crucial test of AI code generation
- How precision and recall shape AI code-review performance
- Why static AI benchmarks eventually become unreliable
- How real-world developer behavior can improve evaluations
- What more reliable and interpretable AI could unlock
- The tools and programming languages Yash uses in his own work
- How aspiring researchers can get started in machine learning

Today’s language models can generate code, solve complex problems, and power increasingly autonomous systems. But without understanding why they succeed, when they will fail, and how their internal mechanisms produce their outputs, building AI systems we can truly trust remains difficult.

Could interpretability provide the scientific foundation for the next generation of AI?

Learn more about Martian: https://withmartian.com/

Learn more about CodeRabbit: https://coderabbit.ai/

Subscribe for more conversations about AI, software engineering, code review, and the future of developer tools.

#LLM #AIInterpretability #ArtificialIntelligence

More AI podcast episodes

Browse all →

Want to find AI jobs?

Join thousands of AI professionals finding their next opportunity

We respect your inbox. Unsubscribe at any time.

© 2026 The Merge (by CodeRabbit). All rights reserved.

Common Questions

Frequently asked questions

Quick answers about how DevFound's AI matching, resumes, and referrals work.

DevFound's AI Copilot ingests your profile, goals, and live job data to deliver curated matches in seconds. Every match includes a resume variant, suggested referrals, and interview prep so you can act immediately. The more feedback you provide, the sharper the Copilot becomes.

AI-led job searches shrink the hours spent sifting through boards and formatting resumes. DevFound pairs automation with your personal outreach, so you reserve energy for interviews and negotiation. Traditional networking still matters, but AI gives you a lift before you even send a message.

Modern AI roles expect comfort with production-grade code, data fluency, and practical ML tooling. The strongest candidates pair deep technical chops with storytelling—translating model impact to product, GTM, and exec partners. Continuous learning keeps you ahead as stacks evolve.

DevFound rewards active seekers. Keep your profile fresh, respond to match quality prompts, and enable alerts so you never miss a role. The AI prioritizes companies and teams that align with your feedback, accelerating both introductions and interview invites.

High-density tech hubs continue to host the deepest AI talent pools, yet distributed teams are catching up fast. Use DevFound filters to hone in on onsite, hybrid, or fully remote roles and watch openings expand across time zones.

DevFound aggregates thousands of remote AI openings and flags the nuances—core hours, async culture, and visa needs—up front. The Copilot also recommends how to position your distributed work experience so hiring managers know you can thrive on a remote team.

Still have questions?

Contact Support