Subvolts take AI research faces compute bottlenecks, not idea shortages, with significant geopolitical and leadership shifts predicted. Dec 18, 2025 - Developers
- Understand the core research challenges and potential future directions in AI development.
- Business
- Gain insights into market dynamics, innovation constraints, and strategic partnerships shaping the AI industry.
This episode features a deep dive with AI researchers into current research challenges and future predictions, offering valuable industry insights.
AI Vibe Check: The Actual Bottleneck In Research, SSI’s Mystique, & Spicy 2026 Predictions
Audio streams from by Redpoint Ventures. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Ari Morcos and Rob Toews join Unsupervised Learning to discuss AI research bottlenecks, model plateaus, and future predictions. They debate if breakthroughs are happening outside LLMs, why capital might limit innovation, and the effectiveness of reinforcement learning. The conversation also touches on OpenAI's vulnerabilities, SSI's mystique, and U.S. chip restrictions' impact on China's AI development. Bold 2026 predictions include leadership changes at OpenAI and Chinese open-source models leading.
Experts discuss if AI is hitting a wall or just changing focus. They explain why too much money might slow things down and predict big changes at OpenAI and in AI leadership by 2026.
This episode offers insights into the current state and future trajectory of AI research, directly relevant to frontier labs. Listeners will learn about potential research plateaus, the real constraints on innovation beyond ideas, and geopolitical impacts on AI development. It is for AI researchers, industry professionals, and investors tracking the competitive landscape and strategic decisions of major AI players like OpenAI and Meta.
- HostJacob Effron
- HostJordan Segall
- GuestAri Morcos
- GuestRob Toews
- Guests: Ari Morcos and Rob Toews
- Show: Unsupervised Learning
- Length: 78 minutes
- Topic: Debate on AI model plateaus and breakthroughs in other modalities
- Claim: Infinite capital at labs may actually constrain innovation
- Prediction: Sam Altman won't be OpenAI's CEO by year-end 2026
Ari Morcos and Rob Toews return for their spiciest conversation yet. Fresh from NeurIPS, they debate whether models are truly plateauing or if we're just myopically focused on LLMs while breakthroughs happen in other modalities.
They reveal why infinite capital at labs may actually constrain innovation, explain the narrow "Goldilocks zone" where RL actually works, and argue why U.S. chip restrictions may have backfired catastrophically—accelerating China's path to self-sufficiency by a decade. The conversation covers OpenAI's code red moment and structural vulnerabilities, the mystique surrounding SSI and Ilya's "two words," and why the real bottleneck in AI research is compute, not ideas.
The episode closes with bold 2026 predictions: Rob forecasts Sam Altman won't be OpenAI's CEO by year-end, while Ari gives 50%+ odds a Chinese open-source model will be the world's best at least once next year.
(0:00) Intro
(1:51) Reflections on NeurIPS Conference
(5:14) Are AI Models Plateauing?
(11:12) Reinforcement Learning and Enterprise Adoption
(16:16) Future Research Vectors in AI
(28:40) The Role of Neo Labs
(39:35) The Myth of the Great Man Theory in Science
(41:47) OpenAI's Code Red and Market Position
(47:19) Disney and OpenAI's Strategic Partnership
(51:28) Meta's Super Intelligence Team Challenges
(54:33) US-China AI Chip Dynamics
(1:00:54) Amazon's Nova Forge and Enterprise AI
(1:03:38) End of Year Reflections and Predictions
With your co-hosts:
@jacobeffron
- Partner at Redpoint, Former PM Flatiron Health
@patrickachase
- Partner at Redpoint, Former ML Engineer LinkedIn
@ericabrescia
- Former COO Github, Founder Bitnami (acq’d by VMWare)
@jordan_segall
- Partner at Redpoint
Jacob is an AI investor at Redpoint Ventures. He's led Redpoint's investments in companies like Abridge, Physical Intelligence & Legora. Follow Jacob on Twitter (@jacobeffron).
On Unsupervised Learning we probe the sharpest minds in AI in search for the truth about what's real today, what will be real in the future and what it all means for businesses and the world. If you're a builder, researcher or investor navigating the AI world, this podcast will help you deconstruct and understand the most important breakthroughs and see a clearer picture of reality. Subscribe to this show to stay up to date on our latest episodes.
Same guests, other shows
1 h 07AI Vibe Check: Lab Wars, Why APIs Might Vanish & Future Predictions
WithAri Morcos, Rob Toews
1 h 17
1 h 04
43 min2025 In Review, 2026 Predictions — With Reed Albergotti
WithReed Albergotti
45 minAI Enterprise - Databricks & Glean | BG2 Guest Interview
WithAli Ghodsi, Arvind Jain
Questions people ask
- Are AI models plateauing?
- The episode debates whether models are truly plateauing or if focus on LLMs obscures breakthroughs in other modalities.
- What is the real bottleneck in AI research?
- The conversation argues that the real bottleneck in AI research is compute, not ideas.
- What are the predictions for 2026?
- Predictions include Sam Altman not being OpenAI's CEO by year-end and a Chinese open-source model potentially becoming the world's best.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to by Redpoint Ventures.