AI

Artificial intelligence, machine learning, and everything LLM

1375 episodes 11 topics RSS Feed

From the latest breakthroughs in large language models to the rise of autonomous AI agents, this channel dives deep into the technology reshaping every industry. Corn and Herman explore how AI works under the hood, debate the implications of increasingly capable systems, and try to make sense of a field that moves faster than anyone can keep up with.

#5232: Zello, PTT Buttons, and the Kitchen Intercom Question

Why your Bluetooth speaker ignores Zello's push-to-talk — and what to buy instead for a hands-full, phone-off fallback.

androidhome-safetyemergency-preparedness

#5230: Why Your Research Agent Won't Stop Searching

Most agent search runs add no new evidence — and 66-85% of the context window is just search snippets. Here's what actually stops the loop.

ai-agentsragai-orchestration

#5220: Malware vs. Virus: What's Actually Attacking Your Machine

Every virus is malware, but not every malware is a virus — and that distinction explains why your antivirus keeps missing things.

cybersecuritysocial-engineeringmalware-taxonomy

#5208: Why Tech Consortia Keep the Stack From Flying Apart

From GSM to HDMI to LTO, why competitors form standards bodies — and what happens when the commitments break.

hardware-standardsmodel-context-protocolai-orchestration

#5207: Hard Drives Are Having Their Biggest Boom in a Decade

Everyone assumes flash killed the spinning disk. Seagate's 52% margins and sold-out capacity through 2028 say otherwise.

data-storagesupply-chainhardware-reliability

#5206: Why IKEA Has No Replacement-Parts API

IKEA's catalog is community reverse-engineering. GS1 solved replacement semantics years ago. Neither reaches a purchasing manager's database.

supply-chainapi-integrationlegacy-systems

#5205: Measuring Rooms From One Photo

One photo, one reference ruler, and a geometry trick from 1999 that measures objects without any AI at all.

computer-visiongeodesycoordinate-systems

#5203: Why LTO Tape Still Backs Up the Cloud

Magnetic tape isn't dead — it's holding up the cloud. How LTO-10 broke a 25-year compatibility promise, and why AI needs tape more than ever.

data-storagedata-redundancybackup-strategies

#5198: Counting the AI Models Nobody Releases

YOLO is free — the labeled data isn't. Why the custom image models that matter most are the ones you'll never see on Hugging Face.

computer-visionsatellite-imagerytraining-data

#5188: DeepSeek's Point Release That Isn't

DeepSeek shipped a whole new architecture and called it a point release. Here's what actually changed inside the model.

transformerslarge-language-modelsai-inference

#5187: Translation Is Not a Pipe: What Machines Miss

Rubrics rate machine translation above human work — until readers pick side-by-side. What translators know that metrics can't see.

llm-as-a-judgelinguisticscultural-bias

#5184: AI Is a Number Factory, Not a Wordsmith

Under the prose, every AI model is just matrices of floating-point math. So where does the randomness actually come from?

ai-inferenceinference-parametersdrug-discovery

#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?

ai-historymachine-learning-historygenerative-ai

#5182: DeepSeek V4.1 Flash: 1M Context, 437x Smaller KV Cache

DeepSeek V4.1 Flash landed with a 1M-token window and a KV cache 437x smaller than V1. Here's what actually changed — and why the middle of your co...

context-windowtransformersai-inference

#5169: Fine-Tuning vs From-Scratch for Minor Languages

What 774 experiments and a new Armenian model reveal about the tradeoff between fluency and knowledge in low-resource languages.

fine-tuningtraining-datalow-resource-languages

#5167: The Stethoscope Isn't Dead, It's Getting AI Ears

Two centuries of the stethoscope — and why AI is now making it more useful than ever, not obsolete.

healthmedical-historyai-inference

#5166: Why Your Baby Tracker Can't Log Oxygen Saturation

Baby apps have templates for everything — except the metrics you actually need. The custom tracking framework exists. It just never shipped.

healthusabilitysoftware-development

#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.

ragprompt-engineeringlatency

#5160: Bluetooth Peripherals on One Android Phone

How many Bluetooth devices can one Android phone juggle? The seven-device limit is a stack artifact — profiles matter more.

iot-protocolshardware-standardshuman-computer-interaction

#5158: A Pulse Oximeter Said 97%. That Was the Problem.

A reassuring oxygen reading can hide a child in respiratory distress. Here's what home devices actually tell you — and what they can't.

respiratory-healthchild-developmenthuman-computer-interaction