Lounge›Insight
Key concepts and the latest issues in crypto, stocks and AI, explained
Lounge›Insight
Key concepts and the latest issues in crypto, stocks and AI, explained

A plain-English guide for people covered by Korea's National Health Insurance: when dividends start counting, how dependents lose status, what regional and employee subscribers pay, and which ETF payouts are excluded. Checked against NHIS and statute pages in October 2026.

Artificial Analysis released Intelligence Index v5 on October 27, 2026. Here is the new benchmark mix, the weights, the launch-day leaderboard, and why v5 scores should not be compared with v4 numbers.

An earnings call explains numbers that are usually already public. Here is how to listen to the prepared remarks and the analyst Q&A, what guidance language to track, and how Regulation FD shapes what management can say.

ARCC is the largest publicly traded business development company. This guide explains where its dividend comes from and how to judge whether it is sustainable using core EPS, non-accrual rates, NAV and leverage from its Q3 2026 report.

The Ethereum Foundation does not own or run Ethereum, but it regularly sells ETH to fund research, grants and security work. Here is how its treasury policy works, what it sold in 2025-2026 and how to check whether a sale actually moved the market.

Infinity Ground pitches an agentic IDE that turns plain-language prompts into dApps. Here is what the AIN token does, how its supply is structured, and which market signals to check after its October 2026 surge.

Auto-Research Arena, posted to arXiv in October 2026, tests whether AI research agents can improve a small language model's training and inference code across nine targets. Here is what it measures, what it found, and what it does not prove.

Under Korean tax rules, pension savings contributions within the annual credit limit are deemed credited even if you never claimed the credit. Here is how Korean residents register those amounts as tax-excluded before withdrawing.

Spinning up more AI coding agents produces more code, but not proportionally more finished features. Here is why Brooks's law still bites, what the 2026 evidence shows, and how to design work splits and review so parallel agents actually help.