HAI

HAI

100 videos · Open playlist on YouTube

2:08

Bridging the Pixel Gap in Browser Automation.

September 1, 2026

2:23

How One Narrow Safety Rule Can Make an AI Less Safe Everywhere Else.

September 1, 2026

2:04

Why splitting a chunk from its document makes it retrieve for the wrong question

September 1, 2026

4:20

Three You Can Take Back. One You Can't.

September 1, 2026

2:21

Why a 50-turn agent pays for the same screenshot 35 times unless it caches the pixels

September 1, 2026

1:53

Why Claude's Click Lands in the Wrong Spot — and the One Ratio That Fixes It

September 1, 2026

3:00

Mycroft Thesis Guard: Fighting Thesis Drift in Investment Research

September 1, 2026

2:18

Mycroft's GitHub MCP Assistant Part 2: The Manual Client & Debugging Harness

September 1, 2026

1:27

RAG Explained by Rishika: Open-Book vs. Closed-Book AI

September 1, 2026

3:35

Private AI Valuation Agent Week 5: Local LLM vs. Rules

September 1, 2026

2:54

Mars Rover Autonomy: Scored Paths & Target Selection

September 1, 2026

3:22

Private AI Valuation Agent Week 4: Entity Resolution & The Golden Set

September 1, 2026

4:00

Fast Radio Bursts: Real-Time Sky Triage

September 1, 2026

1:34

Open Weight vs. Closed Weight AI: Understanding the Trade-offs

September 1, 2026

2:09

Muskan's Update on Vendor Intel Guardrails: Enforcing Instructions in Code

September 1, 2026

1:55

Brief Evaluation Harness: Grading AI Outputs with Code

September 1, 2026

1:47

How to Debug AI Agent Hallucinations: Trace the Entire Pipeline

September 1, 2026

4:08

Correlation is Not Causation: Separating Four Key Analytical Tasks

September 1, 2026

9:35

Causal Couture Week 6: Launching the Causal Intelligence Layer

September 1, 2026

6:22

Causal Couture Week 5: Implementing Scenario Intelligence & Trends

September 1, 2026

4:33

Causal Couture Week 4: Preparing the Phase 4 Architecture

September 1, 2026

3:59

Causal Couture Week 3: Designing the Scenario Intelligence Layer

September 1, 2026

4:22

Causal Couture Week 2: Evaluating the Scorecard & Identifying Limitations

September 1, 2026

3:55

Causal Couture Week 1: Building the Phase 3 Analytical Prototype

September 1, 2026

5:04

Why AI Agents Fail Silently?

September 1, 2026

4:51

What Actually Makes an AI "Agentic"? (The 4 Tiers)

September 1, 2026

9:48

Mycroft Update: Running Two AI Agents Live

September 1, 2026

4:44

When AI Agents Disagree: Part 2 — Building the Code in Mycroft

September 1, 2026

4:10

When AI Agents Disagree: Part 1 — The Mechanics of Cross-Agent Validation in Mycroft

September 1, 2026

1:49

Why the same API key that shipped your prototype becomes a critical bug in production

September 1, 2026

2:03

Why web search never runs your code but your own tool always does — with identical syntax

September 1, 2026

1:55

Computer Use: Demo to Production.

September 1, 2026

2:13

The model says "I have no preferences" while assigning 74% to survival.

September 1, 2026

1:39

Persisting Progress Across Context Windows

September 1, 2026

2:29

Two Resolutions, One Click — The macOS Coordinate Roundtrip.

September 1, 2026

6:34

Verifying Fluent AI Assertions

September 1, 2026

5:59

How to Engineer an AI Agent: Pipelines vs. Autonomous Loops

September 1, 2026

3:36

What is MCP? Connect AI Apps to Your Tools in 8 Lines of Python!

September 1, 2026

3:44

Why We Built the Mycroft Log Book Before the AI Router

September 1, 2026

2:49

Why Creativity is a Bug in Financial LLMs (Understanding Temperature)

September 1, 2026

3:20

Mycroft's ECIS Ep. 4: Who Said It? Weighting Context in AI Financial Analysis

September 1, 2026

3:14

Mycroft's ECIS Ep. 3: Scaling to Three Models & Verifying AI Uncertainty

September 1, 2026

2:56

How Attention Rewrites Financial AI (And Why Hedges Matter)

September 1, 2026

6:18

Mycroft: Rebuilding Morgan Stanley's AI Tools (And Finding a Hidden Bug)

August 31, 2026

3:01

RAG Silent Failure: The 4 Invisible Gaps Killing Your AI Apps

August 31, 2026

2:27

Mycroft: Why Our AI Runway Scorer Refuses to Make Judgements

August 31, 2026

2:24

Mycroft Ep. 2: One Run, Two Customers (Human-Readable vs. Machine JSON)

August 31, 2026

7:55

RAG First Principles: 3 Best-Practice Myths Debunked by Data

August 31, 2026

2:27

Mycroft Ep. 3: The Boring (and Essential) Work of Code Auditing

August 31, 2026

2:17

Mycroft Update on Survivorship Bias: Why Excluding Failures Distorts Backtests

August 31, 2026

2:13

Why Backtesting on Revised Financial Data is "Look-Ahead Bias"

August 31, 2026

1:54

Building Regex Claim-Splitting Logic for Patent Parsing in Mycroft

August 31, 2026

2:06

Why Correlation Doesn't Equal a Tradable Signal (Spurious Correlation)

August 31, 2026

3:18

Deterministic What-Ifs in Mycroft: Scenario Analysis Without AI Guessing

August 31, 2026

3:20

What Actually Makes an AI "Agentic"? (Cutting Through the Hype)

August 31, 2026

3:17

Mycroft Update: Why Your AI Agent Ignores Your Code

August 31, 2026

3:47

Fencing the AI: Why Our Finance Agent Cannot Approve Itself

August 31, 2026

3:25

Inside an AI Agent Loop: The Only Step That Touches Reality

August 31, 2026

3:53

Refusal by Design: Implementing 15 Failure Safeguards in AI for Mycroft

August 31, 2026

3:16

How to Test Mycroft's Finance AI Agent: Breaking the Books on Purpose

August 31, 2026

3:52

How to Prove What Your Financial AI Actually Did (Auditing the Chain)

August 31, 2026

5:29

Quantum Sphere: The Extinction Paradox That Defies Everyday Physics

August 31, 2026

3:29

Building an Evidence-Driven AI Variance Engine in 3 Weeks for Mycroft

August 31, 2026

3:30

AI "Forgetting" is Actually Context Overflow (How to Code Eviction)

August 31, 2026

2:49

Mycroft Update: What Can We Actually Prove About AI?

August 30, 2026

1:53

Why partial JSON can be a valid object before the closing brace arrives

August 29, 2026

1:50

Why the server hands back an encrypted context you're only going to echo

August 29, 2026

1:42

Why wrapping the same text in XML changes the answer Claude gives

August 29, 2026

1:37

Claude, Listener Creator.

August 29, 2026

1:27

Claude, Executive Briefing.

August 29, 2026

1:37

Claude, Listener Creator.

August 29, 2026

1:27

Claude, Executive Briefing.

August 29, 2026

1:50

Why the server hands back an encrypted context you're only going to echo

August 29, 2026

1:42

Why wrapping the same text in XML changes the answer Claude gives

August 29, 2026

1:53

Why partial JSON can be a valid object before the closing brace arrives

August 29, 2026

3:30

Claude, Equipped — What a Cowork Plugin Actually Is

August 29, 2026

1:30

Claude, Action Creator.

August 29, 2026

3:39

Claude, On Call — The Support Plugin

August 29, 2026

4:12

Claude, Closing — Do You Have to Be a Salesperson to Use the Sales Plugin?

August 29, 2026

3:30

Claude, Equipped — What a Cowork Plugin Actually Is

August 29, 2026

1:30

Claude, Action Creator.

August 29, 2026

3:39

Claude, On Call — The Support Plugin

August 29, 2026

4:12

Claude, Closing — Do You Have to Be a Salesperson to Use the Sales Plugin?

August 29, 2026

3:43

Claude, Unstuck

August 29, 2026

3:43

Claude, Unstuck

August 29, 2026

2:10

Claude, Finding It

August 29, 2026

2:10

Claude, Finding It

August 29, 2026

4:31

Claude, Shipping — The Product Plugin

August 29, 2026

4:55

Claude, On Message — The Marketing Plugin

August 29, 2026

4:31

Claude, Shipping — The Product Plugin

August 29, 2026

4:55

Claude, On Message — The Marketing Plugin

August 29, 2026

5:14

Claude, In Concert

August 29, 2026

5:14

Claude, In Concert

August 29, 2026

5:33

Claude, Counsel

August 29, 2026

5:13

Claude, Scouting — The Research Plugin

August 29, 2026

5:05

Claude, By the Numbers — The Data Plugin

August 29, 2026

4:25

Claude, Configured — Installing and Customizing Plugins

August 29, 2026

4:29

Claude, In Order — The Productivity Plugin

August 29, 2026

5:33

Claude, Counsel

August 29, 2026

5:13

Claude, Scouting — The Research Plugin

August 29, 2026

Other series

Humanitarians AI Lyrical Literacy Project