
Why Would AI Want to do Bad Things? Instrumental Convergence
Robert Miles AI Safety274K views
Robert Miles explains instrumental convergence: why almost any goal leads an AI to seek resources and self-preservation.
267,244 videos · Algorithms, systems, security, machine learning and building software.

274K views
Robert Miles explains instrumental convergence: why almost any goal leads an AI to seek resources and self-preservation.

710K views
Robert Miles explains the orthogonality thesis: why intelligence and goals are independent, so a smart AI need not have 'sensible' goals.
9:40328K views
Nine documented cases where reinforcement-learning systems satisfied the letter but not the intent of their objective.
9:0630.7K views
Part 1 of "Teaching Computers to Learn": what it means for a machine to learn from data, with the history of the idea.
11:2524.4K views
Part 2: the mathematical origins of machine learning, from least squares and regression to fitting models.
12:2457.6K views
Part 3: Rosenblatt's perceptron, how it learns a linear boundary and where it fails.
13:35A 14-minute video from Jordan Boyd-Graber: “How to find and measure AI's weak spots [Research, NAACL 2025 Outstanding Paper]”.
13:35A 14-minute video from ritvikmath: “Multiclass Classification : Data Science Concepts”.
6:55A 7-minute video from AI Coffee Break with Letitia: “[Own work] MM-SHAP to measure modality contributions”.
18:18A 18-minute video from AI Coffee Break with Letitia: “Linear algebra with Transformers – Paper Explained”.
1:21:12A 81-minute video from متعة الرياضيات و الفيزياء: “كيف أثبت الذكاء الإصطناعي نافييه ستوكس ؟”.
25:39A 26-minute video from 안될과학 Unrealscience: “양자컴퓨터 원리와 구글의 양자컴퓨터가 대박인 이유 [안될과학 -랩미팅 2화 1부]”.
13:18A 13-minute video from NPTEL-NOC IITM: “Support Vector Machines Deep Dive: Kernels & Optimization”.
22:25A 22-minute video from freeCodeCamp Arabic's catalogue; the title names its topic.
5:49Intro
A 6-minute video from Greg Durrett: “Machine Learning (NLP for High School)”.
6:43A 7-minute video from Universitat Politècnica de València - UPV: “Redes neuronales y Deep Learning | 22/35 | UPV”.
6:59A 7-minute video from Greg Durrett: “Ethics: Dangers of Automation (Natural Language Processing at UT Austin)”.
1:01:04A 61-minute video from 주민철(교원-지능형ICT융합전공): “강화학습-미로찾기-Q-learning 설명추가”.
5:40A 6-minute video from Greg Durrett: “BART (Natural Language Processing at UT Austin)”.
15:20
15:48A 16-minute video from 안될과학 Unrealscience: “이세돌과 맞대결한 인공지능과 딥러닝의 원리는 무엇일까 [안될과학 -랩미팅 3화 2부]”.
39:59A 40-minute video from ScienceEtonnante: “Les 4 étapes pour entrainer un LLM”.
8:38Intro
A 9-minute video from Higgsino physics's catalogue; the title names its topic.
17:06A 17-minute video from Hemerson Pistori: “Da desordem para a inteligência artificial: por que a entropia é um conceito central em IA ?”.
12:56A 13-minute video from AI Coffee Break with Letitia: “Data BAD | What Will it Take to Fix Benchmarking for NLU?”.
7:06Intro
A 7-minute video from Neuro Symbolic: “Introduction to Propositional Logic”.
9:52A 10-minute video from AI Coffee Break with Letitia: “Training large language models to reason in a continuous latent space – COCONUT Paper explained”.
21:12A 21-minute video from Umar Jamil: “How diffusion models work - explanation and code!”.
3:09A 3-minute video from Greg Durrett: “LM Evaluation (Natural Language Processing at UT Austin)”.
8:09A 8-minute video from Universitat Politècnica de València - UPV: “Herramientas IA. Generalidades sobre herramientas de generación de imágenes a partir de tex | | UPV”.
43:42A 44-minute video from Monsieur Phi: “GPT-4 est-il incontrôlable ?”.
6:30A 6-minute video from Wandida, EPFL: “AIXI (Artificial Intelligence Crossed with Induction)”.
8:25A 8-minute video from Gate Smashers: “What is Convolution in CNN | Artificial vs Convolutional Neural Network”.
19:41A 20-minute video from Science4All: “Deep learning et fractales | Intelligence artificielle 50”.
24:27A 24-minute video from Tech With Tim: “Local AI Explained: How to Run AI Models on Your Computer”.
29:06Intro
A 29-minute video from Hugging Face: “A gentle visual intro to Transformer models”.