safety testing

Robert Miles introduces the first of several videos on the paper 'Concrete Problems in AI Safety', focusing on how AI systems can cause accidental negative side effects. It explains the problem in accessible terms.

Robert Miles goes through ten common reasons people give for dismissing AI safety, adapted from Stuart Russell's list, and answers each one. It is aimed at skeptics and newcomers.

Robert Miles examines whether the case for AI safety is a Pascal's mugging, a tiny probability of an enormous payoff. He argues it is not, because AI risk does not depend on vanishingly small probabilities.

Destin of Smarter Every Day goes behind the scenes at the 23andMe lab to show how genotyping works. He also discusses DNA testing and privacy.

Veritasium tells how Kodak film was fogged by radioactive fallout from the first atomic tests, letting the company detect nuclear testing before anyone announced it. The video also covers the deal to keep it quiet and how fallout contaminated packaging.

FortNine looks at the data on why newer pickup trucks with tall, blunt grilles kill so many pedestrians, cyclists and motorcyclists. It argues that lowering hood height would substantially cut fatalities.

Not Just Bikes explains why cars rarely crash into buildings in the Netherlands, attributing it to street design that forgives mistakes and to systematic crash investigation. It contrasts this with North American acceptance of such crashes.

Not Just Bikes argues that oversized American fire trucks and fire department requirements push cities toward wide roads and dangerous street design. He compares this with the Dutch approach to emergency response.

AVweb's Paul Bertorelli analyzes the safety record of the Cirrus Airframe Parachute System (CAPS), looking at the 90-plus deployments over 20 years of Cirrus flying. He asks whether the whole-aircraft parachute has actually saved lives.

Not Just Bikes explains why crossing streets is dangerous in the US and Canada, where roads are designed to move cars fast, and contrasts it with Dutch designs for pedestrians.

Robert Miles asks whether corporations can be thought of as artificial general superintelligences and how far the analogy holds. He argues the differences, such as being limited by employees' capabilities, matter for AI safety.

Tom Scott explains the design of the British BS 1363 plug and socket, including fuses, shutters and safety features, and why he thinks it beats other plugs.

Technology Connections explains weaknesses in the US electrical system, centering on how cord and circuit ratings can mismatch and cause fires. It covers extension cords, power strips and plug history in about 25 minutes.

A short Not Just Bikes video on continuous sidewalks in the Netherlands, where the sidewalk runs unbroken across side streets so cars must visibly cross pedestrian space. It places them within Dutch sustainable safety and Vision Zero efforts.

Tom Scott explains why Ipley Cross in England's New Forest was one of Britain's most dangerous junctions. The angled crossroads lets a car's A-pillar hide approaching traffic and cyclists because there is no relative motion in the driver's view.

Veritasium takes an IQ test and examines what IQ measures, its history including military use and eugenics, its controversies, and how it correlates with life outcomes. Experts including Cecil Reynolds and Stuart Ritchie contribute.

A doctor from Medlife Crisis explains in under three minutes why even a rich patient should not simply order every medical test, touching on pre-test probability and Bayesian reasoning. False positives make blanket testing counterproductive.

FortNine surveys motorcycle airbag technology and explains why airbags have not reached mainstream adoption. It breaks down gear such as the Helite Turtle, Dainese D-Air and Alpinestars Tech-Air.

BigClive explains why microwave oven transformer projects are extremely dangerous, since the high voltage and high current can kill. He contrasts them with spot welder builds where the secondary is replaced by low-voltage heavy wire.

Practical Engineering explains how engineers and agencies put a dollar value on a human life to weigh cost against safety. It covers the legal and regulatory basis for these calculations.

A SAAMI-produced training video for firefighters explaining what happens when sporting ammunition burns, including that unchambered cartridges lack the energy to penetrate much. It includes test footage of ammunition in fires.

Robert Miles explains instrumental convergence: why an AGI with almost any goal would tend to pursue sub-goals like self-preservation and resource acquisition. It builds on Omohundro's basic AI drives and his orthogonality thesis video.

Robert Miles explains the orthogonality thesis: how an agent can be highly intelligent while pursuing goals that seem stupid or alien to humans. He separates intelligence from goals and touches on the stamp-collector example.

Robert Miles explains reward modelling, in which an AI learns what humans want from their preference comparisons instead of a hand-written reward function. He walks through the 2017 paper on deep reinforcement learning from human preferences.

Robert Miles covers a paper where researchers ran real versions of his earlier mesa-optimizer thought experiments. They show inner misalignment, where an AI learns the wrong goal in training and misbehaves when deployed.

Not Just Bikes explains how North American speed limits are set and why posted limits matter far less than street design. It argues that drivers follow the road's cues, so safe speeds have to be built into the street.

A Medlife Crisis video by cardiologist Rohin Francis on overdiagnosis and over-screening. It explains how detecting more disease does not necessarily mean saving more lives, and covers harms such as unnecessary treatment.

Not Just Bikes explains how SUVs and light trucks have grown in size and share of US sales, and why that is dangerous for pedestrians, cyclists and city life. He covers the regulatory loopholes, such as fuel economy rules and the chicken tax, that favored them.

Benn Jordan explains how Flock Safety's license plate cameras log and share vehicle locations, and explores how adversarial AI can defeat them. It covers weak security and the surveillance risks of private police-camera networks.

Technology Connections explains that the so-called EV charger in an electric car's charging cable or wall unit is really an EVSE, a smart switch that safely delivers AC power. The actual charger is inside the car. The video covers the safety signaling and mechanisms involved.

America's Test Kitchen tests eight carbon-steel skillets to find the best one, asking whether a single pan can replace stainless, cast iron and nonstick. It names a winner and covers how the pans perform.

Not Just Bikes shows how traffic calming is built into streets across the Netherlands, from city centres to villages. Street design naturally slows cars and makes roads safer.

Adam Neely tests the claims for tuning to A=432 Hz by trying the unusual tuning systems that its proponents' numerology requires. He compares how they sound against standard tuning and temperaments.

Not Just Bikes argues that stop signs are a poor traffic control, drawing on Dutch practice and the Idaho stop to show intersections can be safer without them, especially for cyclists.

A Computerphile video in which Rob Miles warns against assuming a general AI will think like a human. He uses the thought experiment of a stamp-collecting AI to show how a goal-driven system could pursue its aim with dangerous results.

FortNine's tongue-in-cheek guide to why drivers fail to see motorcyclists and how riders can make themselves more visible. It touches on perception psychology, including the role of saccades.

Rob Miles explains for Computerphile why simple rule sets like Asimov's Three Laws cannot make robots and AI safe. Terms like harm and human are too ambiguous to specify precisely.

Technology Connections explains why North American tail light rules, which allow red turn signals and flashing brake lights, make signals ambiguous. He argues for amber signals and cites studies showing a safety benefit.

Veritasium's Derek Muller spends an hour in a silent, dark anechoic chamber at BYU to test the claim that 45 minutes is the most anyone can bear. He reports on his experience and other people's reactions.

Big Clive takes apart an IKEA Koppla USB power supply to assess its safety, looking at isolation and transformer construction.