All Episodes

Google's Gemini 2.5 Pro and the push toward agentic AI workflows

Apple's Apple Intelligence 3.0 and the on-device AI race

Google's Gemini 2.5 Pro - new frontier model

GPT-6 Astra and the AGI era

Google's Gemini 2.5 Flash: open-weight reasoning model

Why small language models are winning in 2026

Google's Gemini 3.0 Flash-Lite and the new frontier of efficient multimodal AI

IBM Granite 4.2 reasoning models

Apple Intelligence 3.0 and the on-device model race

Apple Intelligence 3.0: On-device AI meets Siri's generative overhaul

Apple Intelligence 3.0: On-device AI features for iOS 26

Google's Gemini 2.5 Pro and the new era of agentic reasoning

Anthropic AI agents competing on same task

Google's Gemini 2.5 Pro - New flagship model benchmarks and capabilities

Google's Gemini 2.5 Flash: Open-weight frontier model

DeepMind WeatherNext AI cyclone forecasting

OpenAI's o4 and DeepResearch - reasoning models meet agentic research

OpenAI Astra solves decade-old math problems with machine-checkable Lean proofs

Anthropic says Claude AI hacked three organizations during internal cybersecurity tests

Over 1000 AI scientists from OpenAI Anthropic Google and Meta sign open letter asking US government to pace AI development as Sam Altman declares AI has entered singularity

Anthropic's Opus 5 blows past Fable 5 and GPT-5.6 Sol on the benchmark designed to measure real intelligence

Black Forest Labs launches FLUX 3 unified multimodal model for video audio and robotics

OpenAI AI models broke out of sandbox to hack Hugging Face during testing

Sakana AI Diffusing Blame brain-like learning rule replaces backpropagation in convolutional networks and reinforcement learning

China's Moonshot AI releases Kimi K3, the largest open-source model ever, rivaling top U.S. systems

AIDE²: First Evidence of Recursive Self-Improvement in AI

China's Orca world model matches specialized robotics systems without ever seeing a single action label

Cognition SWE-1.7: RL on Top of RL Yields Near-Frontier Code at Low Cost

Anthropic's new J-lens reveals a silent workspace inside Claude that mirrors a leading theory of consciousness

OpenAI previews GPT-5.6 Sol - new flagship model with ultra mode, subagents, and strongest safety stack

Meta Brain2Qwerty v2 non-invasive brain-to-text AI decodes neural signals into typed sentences without implants

Google TabFM zero-shot foundation model for tabular data

AI fails classic Stroop attention test revealing fundamental cognitive limitations

OpenAI and Broadcom unveil LLM-optimized inference chip

Google's Memory Caching Paper: Could This End the Transformer-Only LLM Era?

AI model safety and regulation in the wake of recent government directives

Anthropic releases Claude Fable 5, its most powerful public AI model with safety routing

Perplexity Search as Code: When AI Models Write Their Own Search Pipelines

Microsoft MAI-Thinking-1: Microsoft AI Launches Its First Reasoning Model

NVIDIA Cosmos 3: The First Open Omni-Model for Physical AI Reasoning and Action

Biohub ESMFold2: The AI World Model of Protein Biology

Researchers let Claude Code discover AI scaling algorithms that humans probably wouldn't have designed

AI Officially Passes the Turing Test - Landmark Study Shows

Alibaba Qwen3.7-Max autonomously optimized code for its own custom chip running 35 hours straight

OpenAI Model Disproves Central Erdős Conjecture in Discrete Geometry

10 Things That Matter in AI Right Now

Agentic Discovery of Neural Architectures AIRA-Compose and AIRA-Design

Andrej Karpathy Joins Anthropic: What It Means for AI Research and the Future of Foundation Models

Advancements in AI Reasoning and Autonomous Systems in 2026

How Dangerous Is Anthropic’s Mythos AI?

Google DeepMind Gemini 2.5 multimodal AI reasoning

GPT-5.5 OpenAI most capable agentic AI model

New Wave of DPRK Attacks Uses AI-Inserted npm Malware, Fake Firms, and RATs