Subvolts take AI isn't stalling; frontier labs see clear paths for rapid future progress. Oct 29, 2025 - Developers
- Learn about arguments for continued rapid AI progress and potential future capabilities.
- Business
- Understand the debate around AI's trajectory and its potential impact on labor markets.
- Everyday users
- Discover why AI development is not slowing down and what advancements to expect.
This is a solid interview discussing current AI progress debates and future outlooks from a prominent venture capital firm.
Is AI Stalling Out? Cutting Through Capabilities Confusion, w/ Erik Torenberg, from the a16z Podcast
Audio streams from Erik Torenberg, Nathan Labenz. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Erik Torenberg joins The Cognitive Revolution podcast to debate if AI progress is stalling. Host Nathan Labenz counters, citing qualitative advances like 100X context windows, real-time voice, and improved reasoning. The episode explores AI's labor market impact, protectionism, and recursive self-improvement, arguing for continued rapid AI development.
Some people worry AI is not getting better fast enough. This episode argues that AI is actually making big leaps in areas like understanding long texts and talking in real time. This matters because faster AI could change jobs and create new technologies sooner than we think.
This episode dissects the debate on AI's current trajectory, offering a counterpoint to claims of stagnation. It's for anyone tracking AI's real-world impact and future potential, particularly those interested in how frontier labs like OpenAI, Google Gemini, and Anthropic are pushing capabilities. The discussion on labor and recursive self-improvement connects directly to current industry concerns and research directions.
- HostNathan Labenz
- GuestErik Torenberga16z
- Episode Title: Is AI Stalling Out? Cutting Through Capabilities Confusion, w/ Erik Torenberg, from the a16z Podcast
- Show: The Cognitive Revolution
- Length: 98 minutes
- Context Window Expansion: 100X
Erik Torenberg joins to debate whether recent developments suggest AI progress is slowing down or stalling, addressing arguments from Cal Newport and others. Nathan counters this view by highlighting significant qualitative advances, including 100X context window expansion, real-time interactive voice, improved reasoning, vision, and AI's growing contributions to hard sciences. The conversation then covers AI's impact on the labor market, the potential for AI protectionism, and concerns about recursive self-improvement. This episode argues that AI capabilities are not stopping, with frontier developers seeing a clear path for continued rapid progress in the coming years.
Sponsors:
Tasklet:
Tasklet is an AI agent that automates your work 24/7; just describe what you want in plain English and it gets the job done. Try it for free and use code COGREV for 50% off your first month at https://tasklet.ai
Linear:
Linear is the system for modern product development. Nearly every AI company you've heard of is using Linear to build products. Get 6 months of Linear Business for free at: https://linear.app/tcr
Shopify:
Shopify powers millions of businesses worldwide, handling 10% of U.S. e-commerce. With hundreds of templates, AI tools for product descriptions, and seamless marketing campaign creation, it's like having a design studio and marketing team in one. Start your $1/month trial today at https://shopify.com/cognitive
PRODUCED BY:
CHAPTERS:
(00:00) About the Episode
(03:49) Is AI Slowing Down?
(09:15) Newport's Scaling Law Theory
(16:56) The Value of Reasoning (Part 1)
(17:17) Sponsors: Tasklet | Linear
(19:57) The Value of Reasoning (Part 2)
(24:52) Explaining GPT-5's Vibe Shift
(31:39) AI's Impact on Jobs (Part 1)
(36:50) Sponsor: Shopify
(38:47) AI's Impact on Jobs (Part 2)
(44:08) Recursive Self-Improvement via Code
(49:35) The Future of Engineers
(53:24) Economic Pressure vs. Protectionism
(58:29) Progress Beyond Language Models
(01:07:11) The State of AI Agents
(01:19:19) China's Open Source Models
(01:29:44) A Positive Vision Forward
(01:37:39) Outro
SOCIAL LINKS:
Website: https://www.cognitiverevolution.ai
Twitter (Podcast): https://x.com/cogrev_podcast
Twitter (Nathan): https://x.com/labenz
LinkedIn: https://linkedin.com/in/nathanlabenz/
Youtube: https://youtube.com/@CognitiveRevolutionPodcast
Spotify: https://open.spotify.com/show/6yHyok3M3BjqzR0VB5MSyk
- context window
- Nathan Labenz
- how much information an AI can consider at once
1 h 27AI:AM: Where's the Steering Wheel? My Agent Set the Password, Hidden Valence, Proofs Find Linux Bugs
WithCameron Berg, Everett Hildenbrandt, Grigore Rosu, Phil Maher
2 h 192 h 021 h 28AI:AM: A Level We Shouldn't Pass? Notes from The Curve + Tokens vs. Salaries & Is SaaS Cooked?
WithThomas Sohmers, swyx, Evan Miyazono, Edward Hu
1 h 03Ep 76: Sora Creators Bill Peebles, Rohan Sahai & Thomas Dimson on Their Unexpected Viral Success
WithBill Peebles, Rohan Sahai, Thomas Dimson
1 h 14All things AI w @altcap @sama & @satyanadella. A Halloween Special. 🎃🔥BG2 w/ Brad Gerstner
WithSatya Nadella, Sam Altman
19 minThe Best of 2025 (So Far) with Sarah Guo and Elad Gil
WithWinston Weinberg, Dr. Fei-Fei Li, Brendan Foody, Dan Hendrycks
Questions people ask
- Is AI progress slowing down?
- The episode argues against this, highlighting significant qualitative advances in AI capabilities like context windows, voice interaction, and reasoning.
- What are some recent AI advancements?
- Advancements include 100X context window expansion, real-time interactive voice, improved reasoning, and vision capabilities.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Erik Torenberg, Nathan Labenz.