DeepSeek

State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI | Lex Fridman Podcast #490

Lex Fridman

Open-weight Chinese AI company known for DeepSeek R1 and ongoing frontier open-weight models; discussed as a pivotal moment in 2025 that spurred a broader wave of Chinese model releases.

You Are Being Told Contradictory Things About AI

AI Explained

Open competitor model family used for benchmarking and synthetic-data experiments.

The 10,000x Yolo Researcher Metagame — with Yi Tay of Reka

Latent Space

A company mentioned for its contributions to Mixture-of-Experts (MoE) models.

Outlasting Noam Shazeer, Crowdsourcing Chai AI w/ 1.4m DAU — with William Beauchamp, Chai Research

Latent Space

A company praised for its innovative inference engine and cost-effective AI models, drawing parallels to Chai's own development philosophy.

How to train a Million Context LLM — with Mark Huang of Gradient.ai

Latent Space

Mentioned for its paper on multi-head latent attention, which Mark Huang found to be a novel and insightful contribution.

Best of 2024: Open Models [LS LIVE! at NeurIPS 2024]

Latent Space

Mentioned as a provider of frontier-level performance models in 2024. Also cited as needing a minimum of 50,000 GPUs for state-of-the-art pre-training.

[Paper Club] Upcycling Large Language Models into Mixture of Experts

Latent Space

A model that uses a large number of experts (128 or 160).

A Comprehensive Overview of Large Language Models - Latent Space Paper Club

Latent Space

An organization involved in AI research, with a paper mentioned for the next session.

Did you miss these 2 AI stories? A *Real* LLM-crafted Breakthrough + Continual Learning Blocked?

AI Explained

The Stablecoin Future, Milei's Memecoin, DOGE for the DoD, Grok 3, Why Stripe Stays Private

All-In Podcast

An AI model mentioned for its open weights, and also referenced in a comparison of Grok 3's quality.

Information Theory for Language Models: Jack Morris

Latent Space

Company that released a 400 billion parameter model, making its base and fine-tuned weights available.

JD Vance's AI Speech, Techno-Optimists vs Doomers, Tariffs, AI Court Cases with Naval Ravikant

All-In Podcast

A Chinese AI model, highlighted as evidence that China is rapidly catching up to the US in AI development, challenging the notion of US monopoly.

DOGE vs USAID, Crypto Framework, Google's $75B AI Spend, US Sovereign Wealth Fund, GLP-1s

All-In Podcast

An AI model or company, mentioned as an example of China's advancements in AI, suggesting they are catching up to the US.

DeepSeek Panic, US vs China, OpenAI $40B?, and Doge Delivers with Travis Kalanick and David Sacks

All-In Podcast

A Chinese AI startup that released the R1 language model, claiming it was trained for $6 million on 2,000 GPUs, sparking debate about AI development costs and open-source vs. closed-source models.

Scaling Test Time Compute to Multi-Agent Civilizations — Noam Brown, OpenAI

Latent Space

Mentioned for a finding that MCTS (Monte Carlo Tree Search) was not very useful to them.

Ray Dalio: US Debt Spiral, How to Avoid Disaster | The All-In Interview

All-In Podcast

An AI model announced, highlighting China's advancements in AI applications.

The Shape of Compute (Chris Lattner of Modular)

Latent Space

A research team that released impressive models and pushed advancements in low-precision training and PTX-level optimization, prompting industry reaction.

Trump's First Week: Inauguration Recap, Executive Actions, TikTok, Stargate + Sacks is Back!

All-In Podcast

A Chinese open-source AI model capable of running on a laptop, demonstrated to be as good as older OpenAI models but at a fraction of the cost, illustrating falling AI development costs.

Sleep-Time Compute — Letta AI (Charles Packer, Charlie Snell, Kevin Lin)

Latent Space

A company whose models (like 3.7) showed significant Pareto shifts with sleep time compute.

The Magic of LLM Distillation — Rishabh Agarwal, Google DeepMind

Latent Space

An organization whose models were used in experiments to demonstrate that distillation is possible even without access to logits, by using synthetic data.

World Leading Investing Expert: The Big Shift Is Coming! This Investment Could 15x in 5 Years!

The Diary Of A CEO

Dan Wang on What China and America Can Learn from Each Other

Conversations with Tyler

Nothing Much Happens in AI, Then Everything Does All At Once

AI Explained

⚡️Factorio Learning Environment: the ultimate Game Agent Eval — Jack Hopkins

Latent Space

An AI model that performed decently in Lab Play but struggled significantly in Open Play, often defaulting to creating excessive numbers of chests.