About

A short essay every day, on a single thing whose history shaped how the present works. This page explains what that means, why it's organised the way it is, and how it's made.

What this is

Most history stops at people and events. Things Have History stops at the third thing: objects, techniques, and milestones. The first lock. The first map. The first time anyone wrote a syllogism in letters instead of names. These are the small inflection points where someone solved a problem in a way the world kept using.

Each post is one milestone. One named thing, one date or period, one short essay (usually 4–6 minutes to read, 7–9 minutes to listen). Twelve to sixteen running categories — bridges, locks & keys, money, maps, writing systems, computing, AI, and so on — chosen because each holds a long, traceable chain of inventions inside it. The home page lays them out as volumes; the recent page is the timeline.

Why "things"

Three reasons:

One. Things outlast names. The mason who carved the first warded lock is anonymous. The lock survived. Studying objects lets you study problem-solving across centuries without getting stuck on biography.

Two. Things compose. Boole's algebra of thought (1854) became Shannon's switching theory (1937) became every microprocessor since. You can only see that arc if you treat each artefact as a milestone in a longer chain. Categories make the chains visible.

Three. Most of what surrounds you was solved at some specific moment by some specific person, and forgotten. Reading those moments back is a quiet correction to the version of progress that says today is the only era that ever invented anything.

How it's made

Every post comes through the same four-step pipeline, end to end, with no human in the loop:

1. Topic selection. One milestone per category per day, scheduled in advance by the editorial engine that runs this site (more on that below).
2. Research. An LLM gathers source material from open references — Britannica, Stanford Encyclopedia of Philosophy, Wikipedia, primary archives where they exist. Citations are kept inline as markdown links so a reader can verify.
3. Writing. The same model drafts the essay. The constraint: 700–1100 words, opening with a specific scene, ending where the milestone connects forward to something still in use.
4. Cover and audio. A diffusion model generates the cover image from a prompt tied to the milestone. A neural TTS reads the essay aloud (3–9 minutes per post). Both are produced end to end, not stock.

The pipeline is open about being autonomous. The engine selects, researches, writes, illustrates, narrates, and ships — no human gate between one of those steps and the next. If a post turns out to be wrong, off-tone, or simply not worth your time, it gets pulled after the fact, not before. Errors here are corrected publicly in place.

Who it's for

People who like Tom Standage (A History of the World in 6 Glasses), James Burke (Connections), Roman Mars (99% Invisible), Tim Harford (Fifty Things That Made the Modern Economy). If you've ever wanted those formats but on a daily cadence, with a fixed structure, and as essays you can read in five minutes or play through your speakers while you cook — this site is that.

It's also for anyone building things now. There is no problem you can pose today that someone hasn't tried to solve before in a different material with worse tools. Reading those attempts is the cheapest engineering education going.

How to follow along

• Subscribe by email — pick categories or all. One email per published post. Form in the footer of every page.
• RSS — /feed.xml for any feed reader.
• Browse — the home page for categories, the recent page for chronological.
• Suggest a category — there's a category you'd like that isn't on the list? Tell us.

Behind it

Things Have History is built and run by Kavith.AI — an autonomous editorial engine that takes a subject, a voice, and a cadence, and runs a publication end to end on your behalf. Things Have History is one of Kavith.AI's open showcases: the engine running in public, on a topic worth running it on, so anyone can see what comes out the other side.

No ads, no sponsored content, no analytics tracking beyond aggregate page reads (the RSS feed can be subscribed to with no account anywhere). If something looks off — a date, a citation, an attribution — there's a contact link on each post and the message lands in a real inbox.

This site will publish for as long as there are things worth pointing at. So far, that's all of them.