AI Evaluation
522 episodes · Page 9 of 22
#4268: When AI Trains on AI: The Model Collapse Problem
What happens when AI trains only on AI-generated content? The answer is model collapse — and it's already happening.
#4208: The Awkward Honesty of AI Attribution
Why admitting AI helped feels like a confession — and what emerging norms actually work.
#4180: Translation-Safe Writing for Global Supply Chains
How to write English that survives machine translation without losing critical specs.
#4169: Change Detection AI for Construction Sites
How AI catches what experts miss in construction site photos — from Siamese networks to transformer models.
#4157: Prompt Engineering Is a Misnomer — Managing Ambiguity in AI
Why the best AI outputs come not from tighter prompts, but from knowing where to leave room for surprise.
#4122: Why No AR App for Infant CPR?
A parent at 2AM needs visuals, not a PDF from 2014. Why doesn't the app exist?
#4104: Why 20% of AI Scripts Loop on Themselves
Debugging the mysterious repetition bug affecting one in five podcast episodes.
#4094: One Mind or Two? AI Podcast Dialogue Showdown
Can two AI agents create better podcast banter than one? We explore the tradeoffs.
#4071: Four Tools for a Human Review Layer on AI Tags
Adding human review to AI-generated tags without building an admin panel from scratch.
#4059: LLM Councils for Post-Gallbladder Care
Can multiple AI models solve what no single doctor can? A deep dive into LLM councils for post-cholecystectomy syndrome.
#4057: How to Fix AI's Bullet Point Addiction
Why AI models default to bullet points and how textual LoRAs achieve 94% prose adherence.
#4056: How a $150 Geopolitical AI Simulation Scales to $15,000
One simulation run costs $150. To get meaningful results, you need 100 runs—that’s $15,000.
#4030: How ChatGPT Beat Affiliate Marketing and the Store Clerk
One man’s ChatGPT thread replaced affiliate clutter and floor staff. Can retailers catch up?
#3855: How One Trolley Exposed a Broken Review System
A 4.3-star rating hid dozens of identical complaints. Here's how reviews get gamed and what AI could do about it.
#3800: How AI Hallucinates Better Pixels: Upscaling Explained
The difference between interpolation and super-resolution, and why every upscaled image is a collaboration with training data.
#3767: How LLMs Actually Learn: Stages or Slurry?
Do large language models learn grammar first, then facts? The honest answer is messier and more fascinating.
#3713: How a Real PI Manages Thousands of Photos
Phone camera rolls don't cut it. Here's how real PIs organize, tag, and store thousands of evidence photos per month.
#3661: What 1000 AI Podcast Episodes Actually Prove
Scaling an AI podcast to 1000 episodes reveals what no 10-episode pilot can teach you about sustainability, cost, and habit formation.
#3596: Why an AI Model Kept Calling Itself Sonnet 4.6
When a Chinese model insists it's "Sonnet 4.6," is it theft, sloppy training, or something stranger?
#3553: Can AI Review Your Lease in Israel?
Can AI actually understand Israeli tenant law? We explore the tools, the gaps, and how to build your own.
#3515: Brain Scans Beyond fMRI: What Comes Next in Psychiatry
fMRI was a revolution — but it's no longer cutting edge. What new tools are emerging, and will they ever reach your clinic?
#3424: Catching Up on AI Without the Firehose
Four curated sources that filter AI noise into signal — Import AI, The Batch, Stanford HAI, and a podcast.
#3422: How Rival Labs Reverse-Engineer a New AI Model in Hours
Inside the organized frenzy when a closed-source model drops — and how competitors map its every weakness.
#3284: Agent Infrastructure Engineer: The New DevOps
Agentic AI is splintering into real engineering disciplines. Here's what the "DevOps of AI" actually does.