November 2026
126 episodes
- #6524 Why Israel Never Finished Its Sovereign Hebrew AI
- #6523 Unpacking Israel's National AI Strategy
- #6522 Nebius and the GPU Cloud Pivot
- #6521 Multitenancy in Qdrant Vector Databases
- #6520 Mapping Multi-Tenancy Across Vector Databases
- #6519 How Smuggled GPUs Move Through Real Companies
- #6518 Why Export Bans Cannot Stop Remote Compute
- #6517 Why GPU Wiring Dictates Model Parallelism
- #6516 Why Cargo Thieves Are Targeting GPU Shipments
- #6515 Creative Schemes Behind the Global GPU Hunt
- #6514 Why Even Solo Developers Face a GPU Shortage
- #6513 Why Dual-Socket Servers Win on Dense AI Models
- #6512 Build Servers Never Actually Build Anything
- #6511 Why Voiceprints Are Not an Interoperable File Format
- #6510 Voice Isolation Before Speech Recognition
- #6509 How a Knesset Government Can Form with Under 61 Votes
- #6508 Why Speech Transducers Omit Common Symbols
- #6507 Why Broadcast Headsets Fit Commentators but Not Voiceover
- #6506 Why Generative Audio Restoration Hallucinates Highs
- #6505 Why Blanket Forts Beat Expensive Portable Vocal Booths
- #6504 Why Voiceover Studios Obsess Over Repeatability
- #6503 Why Tokenizer Vocabulary Size Is a Hardware Choice
- #6502 What a Cross-Entropy Loss of 1.8 Actually Means
- #6501 Splitting Frontier AI Across Thousands of GPUs
- #6500 Why LLM Leaderboards Need Psychometrics
- #6499 Why Simple Rules Beat Value Models in Reasoning RL
- #6498 Why You Must Read AI Traces Before Writing Evals
- #6497 Why Speech Recognition Confidence Scores Lie
- #6496 How Speech Models See Audio as an Image
- #6495 Teaching Speech Models to Choose Their Own Audio
- #6494 Where Speech Models Actually Learn
- #6493 Why Speech Models Forget Other Voices When Tuned on Yours
- #6492 What Audio Models Learn Before Fine-Tuning
- #6491 Agent Skills Break When Secrets Leave Your Laptop
- #6490 What the Evidence on Daylight Saving Actually Shows
- #6489 Why Governments Move Time Zones Off the Grid
- #6488 Why Australian Suburbs Resemble American Ones
- #6487 Turning Acoustic Physics into a Math Problem
- #6486 Why Voice Satellites Need Integrated Audio
- #6485 Choosing Audio Bitrates and Settings for Voice
- #6484 Why Narrow Speech Tasks Still Need Large Models
- #6483 What Kills Ad-Free Social Networks
- #6482 Proprietary Plugs and Formats Nobody Wanted
- #6481 Why Universal Standards Fail
- #6480 When Ready to Attack, Seeming Weak
- #6479 Why Research Agent Agreement Is Not Independence
- #6478 Why Great Dictation Headsets Come from Gaming and Broadcast
- #6477 Finding a Technical Evaluation Job Without the Sales Quota
- #6476 Bluetooth Microphone Quality and the Audio Ceiling
- #6475 USB Audio Limits on Android for Voice Models
- #6474 Why to Call a Headset Specialist Instead of Amazon
- #6473 Why Whispering Breaks Speech Recognition Physics
- #6472 What a VPN Node Can Actually See
- #6471 Why Commercial VPNs Cannot Be End-to-End
- #6470 Why Computer Science Publishes in Conferences
- #6469 Why Switching Off One Attention Head Changes Everything
- #6468 Why Tiny Neural Gates Replaced Voice Activity Gates
- #6467 Voice Activity Detectors Need Adaptive Thresholds
- #6466 Conferences Rule Machine Learning, Not Journals
- #6465 Why Modern AI Models Cut Most Attention Layers