Technology explainers

How Do Self-Custody Crypto Wallets Work, and What Can Go Wrong?

Self-custody wallets manage cryptographic keys that authorize blockchain transactions. They remove a custodian but shift recovery, transaction verification, smart-contract approval, device security, inheritance, and irreversible-error risk to the owner.

How Do Forecasters Predict a Tropical Cyclone's Track and Intensity?

Cyclone forecasts combine satellites, aircraft, radar, ocean data, coupled models, ensembles, and expert judgment. Large-scale steering makes track easier than intensity, while official warnings translate uncertain paths into wind, surge, rain, and local impact.

How Does Heat Move Through a Computer Chip?

Chip heat moves from transistor hot spots through silicon, interconnects, buried interfaces, package materials, thermal paste, spreaders, sinks, and finally air or liquid. Local power density and interface resistance can matter more than the external fan.

How Do Astronomers Track Space Debris Near the Moon?

Lunar-distance debris is tracked mainly with timed optical astrometry, brightness measurements, dynamical orbit fitting, and repeated reacquisition. Angles alone leave large range uncertainty, while weak radar echoes and complex Earth-Moon-Sun gravity complicate prediction.

How Does X-Ray Polarimetry Reveal the Physics of Extreme Objects?

X-ray polarimetry reconstructs a preferred electric-field orientation statistically from many photon interactions. Combined with spectra and timing, it probes magnetic fields, scattering geometry, accretion flows, relativity, and quantum-vacuum effects, but remains model-dependent.

How Can a Printed Object Display Information Without Electronics?

Passive printed displays mechanically reveal colors, symbols, or optical patterns as an object moves or deforms. They need no powered sensor when the physical state directly drives the message, but alignment, wear, viewing angle, and fail-safe coupling determine reliability.

Why Is the Electrolyte So Important in a Rechargeable Battery?

A battery electrolyte transports ions, blocks electrons, survives electrode voltages, and forms protective interphases. Conductivity, solvation, additives, temperature, safety, metal deposition, and full-cell compatibility determine whether the battery performs and lasts.

How Are AI Models Tested for Cybersecurity Risks Before Release?

Prerelease cyber evaluations test model knowledge, tool use, exploit development, human uplift, and long-horizon agent performance in isolated realistic tasks. Their value depends on uncontaminated tests, clear baselines, uncertainty, and thresholds tied to deployment action.

How Does Vagus Nerve Stimulation Work, and Why Does Precision Matter?

Vagus nerve stimulation sends programmed electrical pulses through a nerve containing many functional fiber bundles. Its benefit and side effects depend on which fibers are recruited, making anatomy, electrode design, timing, physiological feedback, and individualized programming central to precision.

How Can a Non-Invasive Brain-Computer Interface Turn Brain Signals Into Text?

Non-invasive brain-to-text systems measure EEG or MEG signals, learn task-linked neural patterns, and decode probable text with machine learning. They do not read unrestricted thoughts, and clinical value depends on intended-user tests, transparent uncertainty, practical hardware, and reliable meaning.

How Do Molecular Electronic Devices Work, and Why Are They Hard to Build?

Molecular devices place molecules inside electrical junctions to conduct, switch, store charge, or sense their environment. Their promise comes from chemical design, but reproducible contacts, mechanism proof, device variation, array integration, endurance, retention, and system-level performance determine whether they can leave the laboratory.

How Do Laws Make Technology Platforms Pay News Publishers?

News-payment laws change bargaining power through negotiation duties, arbitration, copyright-related rights, levies, or hybrid incentives. Their success depends on preventing avoidance, supporting diverse original journalism, preserving public access, and measuring newsroom outcomes rather than announced deal values alone.

How Can a Dead Star Reveal What a Rocky Exoplanet Was Made Of?

Polluted white dwarfs reveal destroyed planetary material when heavy elements temporarily appear in their normally clean atmospheres. Spectral abundances can indicate primitive rock, differentiated interiors, thermal processing, or possible water-bearing material, but accretion history and element sinking limit reconstruction.

How Does Artificial Light at Night Affect Insects, and What Makes a Lamp Safer?

Artificial light can disrupt insect navigation, movement, feeding, reproduction, and predator exposure. Lower-impact lighting removes unnecessary illumination first, then reduces short wavelengths, brightness, spill, and operating time while verifying both human performance and ecological outcomes.

How Can Water Utilities Protect Their Control Systems From Cyberattacks?

Water utilities secure operational technology by inventorying every connection, removing public exposure, segmenting networks, restricting and monitoring remote access, protecting controller logic, and testing backups and manual operation. Resilience must limit physical consequences and restore safe service, not only prevent intrusion.

How Does the EU AI Act Classify Artificial Intelligence by Risk?

The EU AI Act applies several filters: prohibited practices, two routes to high-risk status, transparency duties, and a separate regime for general-purpose AI models. Classification depends on intended use, deployment context, legal role, and current transition dates, while other EU laws still govern systems outside the high-risk category.

How Do Digital Watermarks Identify AI-Generated Images, and Where Do They Fail?

Digital watermarks embed detector-readable signals, while signed provenance records an asserted creation and editing history. Positive results can support attribution, but screenshots, transformation, unsupported tools, stripped metadata, and adversarial removal make negative results weak evidence. Neither mechanism proves the depicted event is true.

How Does Online Age Assurance Work, and What Can It Get Wrong?

Age assurance combines declarations, documents, trusted records, facial estimation, behavioural signals, or privacy-preserving credentials. Every method trades confidence against privacy, accessibility, friction, and boundary errors, so proportionate evidence, data minimization, alternatives, audits, and appeals are essential.

How Can an AI Music Model Reproduce Parts of Its Training Data?

AI music models generalize when they learn shared musical relationships, but memorize when optimization retains source-specific detail that prompts can extract. Dataset duplication, rare examples, model capacity, prompt specificity, and repeated sampling affect the risk; layered data, training, evaluation, licensing, and output controls reduce it.

How Should an AI Shopping Assistant Verify Product Claims?

AI shopping verification requires claim extraction, authoritative evidence, contradiction checks, calibrated status labels, visible shopper warnings, and consistent escalation across listings and enforcement. A chatbot can identify a conflict without proving which statement is true or whether a seller intended deception.