“AI’s capability improvements haven’t come from it getting less affordable” by Anders Woodruff episode artwork

EPISODE · Mar 27, 2026 · 17 MIN

“AI’s capability improvements haven’t come from it getting less affordable” by Anders Woodruff

from LessWrong (30+ Karma)

METR's frontier time horizons are doubling every few months, providing substantial evidence that AI will soon be able to automate many tasks or even jobs. But per-task inference costs have also risen sharply, and automation requires AI labor to be affordable, not just possible.[1] Many people look at the rising compute bills behind frontier models and conclude that automation will soon become unaffordable. I think this misreads the data. The rise in inference cost reflects models completing longer tasks, not models becoming more expensive relative to the human labor they replace. Current frontier models complete tasks at their 50% reliability horizon for roughly 3% of human cost, and this hasn’t increased as capabilities have improved. I define cost ratio as the inference cost of the average AI trajectory that solves a task divided by human cost to complete the same task. Using METR's data, I examine the trend in cost ratios over time. I show three things: Across successive frontier models, the cost ratio at each model's 50% reliability time horizon hasn't increased.Among tasks models successfully complete, longer tasks don't have higher cost ratios than shorter ones.Time horizon trends barely slow when capping AI spending per [...] ---Outline:(02:42) Evidence from METR(03:22) Cost ratio at models 50% time horizon isnt increasing(04:59) Time horizons improvements arent driven by expensive long tasks(06:07) Progress at a fixed cost is just as fast(08:04) Limitations of my methodology(09:42) Inference scaling will just make progress faster(11:10) Conclusion(12:12) Appendix A: Why I get different results from Ord(15:32) Appendix B: Additional cost at time horizon graphs(16:45) Appendix C: 80% affordable time horizon The original text contained 9 footnotes which were omitted from this narration. --- First published: March 27th, 2026 Source: https://www.lesswrong.com/posts/E6ELHguZFNF3Czp55/ai-s-capability-improvements-haven-t-come-from-it-getting --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

NOW PLAYING

“AI’s capability improvements haven’t come from it getting less affordable” by Anders Woodruff

0:00 17:21

No transcript for this episode yet

We transcribe on demand. Request one and we'll notify you when it's ready — usually under 10 minutes.

Accidental Accountant Regan Williams Hi, I'm Regan! I'm a CPA of 30+ years helping "accidental accountants" navigate tax & accounting issues with confidence! Here, we find solutions to common challenges bookkeepers, accountants and CPAs face. Don't see an answer to your question? Then ask! I'm here to help people like you. Two Recruiters: Zero Filter Two Recruiters At Two Recruiters: Zero Filter, we're on a mission to demystify the hiring process, share insider tips, and empower you to maneuver through the professional world with confidence. With more than 30 years of combined experience navigating the intricate web of job markets, talent acquisition, and career development, we're here to spill the tea on everything career related. But wait, there’s more! We will dive into many life topics that are interesting to us as well.  Get ready for a rollercoaster of insights, stories, and no-holds-barred advice!Join us for conversations that matter – where work, life, and authenticity collide in the most unexpected and rewarding ways. Capital Ideas Podcast Capital Group Want to learn how professional investors do it? The Capital Ideas podcast brings you the latest investment thinking from Capital Group, one of the world's largest investment management organizations. Each week we'll get inside the minds of portfolio managers, analysts and economists to break down market trends, macroeconomic forces, investing approaches and lessons learned from personal experience. Take 30 minutes and tap into the intellectual capital of Capital Group. Capital Client Group, Inc.All Capital Group trademarks mentioned are owned by The Capital Group Companies, Inc., an affiliated company or fund. All other company and product names mentioned are the property of their respective companies.For full disclosures go to capitalgroup.com/global-disclosures. My Take On It with Your Angelic Karma® Your Angelic Karma Here we take a look at how the United States measures alongside other First World Nations. + taking a deep dive into the science -The Report

Frequently Asked Questions

How long is this episode of LessWrong (30+ Karma)?

This episode is 17 minutes long.

When was this LessWrong (30+ Karma) episode published?

This episode was published on March 27, 2026.

What is this episode about?

METR's frontier time horizons are doubling every few months, providing substantial evidence that AI will soon be able to automate many tasks or even jobs. But per-task inference costs have also risen sharply, and automation requires AI labor to be...

Can I download this LessWrong (30+ Karma) episode?

Yes, you can download this episode by clicking the download button on the episode player, or subscribe to the podcast in your preferred podcast app for automatic downloads.
URL copied to clipboard!