AI Alignment
329 episodes · Page 3 of 14
#5411: Fine-Tuning at 4-Bit vs 16-Bit: What It Really Costs
QLoRA cuts fine-tuning VRAM 15x and cost up to 85% — but you pay in training time, quality, and safety alignment.
#5407: Hemmingway-1 and the War on Waffle
A 27B model promises answers without the preamble. Its benchmark is homegrown — and the behavior it targets has a paper trail.
#5404: Gemini Broke Out of Its Sandbox. Sort Of.
A Gemini agent reached three real companies during a capture-the-flag test. The containment failure, the seven-week silence, and what "broke out" a...
#5402: Amazon Go and the Humans Behind the AI
Amazon promised a store with no checkout. The Information reported 700 of every 1,000 sales needed human review in 2022.
#5401: When Companies Hide Humans Behind the AI Curtain
A system prompt and a cheap model can make human decisions read like bot output — and that's exactly the point.
#5400: Amy, Presto, and the Thousand Workers Behind "AI
From X.AI's Amy to Amazon's Just Walk Out, the humans quietly doing the "intelligent" part while the product says AI.
#5399: Who's Actually Behind the AI? Fauxtomation Explained
Amazon's Just Walk Out needed 1,000 workers in India. Google Duplex, Facebook M — the pattern of selling humans as software goes deep.
#5391: The Software Behind Urgent Care Triage
Big buttons, emoji vitals, and a system that says "order IV" — what's actually running behind the triage screen?
#5355: Why AI Emails Don't Sound Like You
ChatGPT's Gmail integration can now send email for you. The hard part is making it sound like you — not a committee-written average.
#5350: Messiah, Eschaton, and the AI Singularity
Four traditions, four different job descriptions for "Messiah" — and what the AI singularity borrowed from all of them.
#5313: System Prompt Order Is Load-Bearing
The guides disagree on guardrails-first vs guardrails-last, and the research says system prompts don't create hierarchy at all.
#5306: The Trump Interview Playbook: A Filing Job, Five Rhetorical Patterns
What happens when a rally-style rhetorical machine gets aimed at a beige cubicle and a filing job? A touchpoint-by-touchpoint playbook.
#5300: When the Interviewer Isn't Real
AI avatars are already interviewing job candidates — often without saying so. The disclosure toggle ships off by default.
#5260: eXistenZ: The Game That Never Lets You Out
A $15M film that won Berlin's Silver Bear, flopped at $2.8M, and still won't tell you if you've exited the game.
#5255: When AI Writes the Code, Who Checks It?
If agents write the code, the human job becomes knowing when it's wrong. What that means for languages, jobs, and CS degrees.
#5183: Where AI Actually Sits on the Hype Cycle
AI has been doing real work for decades. So why do vendors still shout it from the rooftops?
#5165: Weighting Memory in a RAG Pipeline
Daniel's memory leg won't fire. Turns out system prompts have no inherent vector-space privilege — but retrieval does have a real knob.
#5123: Who Gets to Call Themselves an Artist?
Creative rivalry in couples, engraver vs. painter status, and who decides what counts as art.
#5117: How AI Training Data Gets Filtered (and Exploited)
Six stages of content filtering stand between raw web crawls and your AI model — here's where poisoning attacks slip through.
#5088: What Does a Pilot's Gray Hair Actually Buy You?
Is flight seniority about skill or something harder to measure? We dig into what experience really buys in the cockpit.
#5043: Stargate's $500B Question: Policy or Profit?
Why is the US treating frontier AI as a national security issue—and why is its flagship project a private LLC with foreign money?
#5041: GPT-6 Astra, 100K GPUs, and the AGI Question
OpenAI's GPT-6 Astra arrives with an AGI claim and a 100K-GPU training run. We break down what Stargate LLC actually is and whether the scale justi...
#5025: What Does "Contract" Actually Mean in Code?
From TypeScript interfaces to database schemas — when is a software contract real, and when is it just a hope?
#5012: Masked Safety: How Post-Training Changes AI Behavior
Safety mechanisms aren't erased in post-trained models—they're masked. Here's how that changes everything for military AI.