Why Titanium

Honestly, the first thing that made me think of the word titanium was pure pronunciation: AI 钛, run together fast, sounds like "AI 态度" — "AI with an attitude." But attitude is too light a word these days. Everyone shouts about attitude; the real thing has quietly disappeared. What finally settled the name was titanium's own temperament.

Titanium's Temper

Titanium melts at 1,668 degrees Celsius. It is forty percent lighter than steel and no less strong; it sits in seawater without rusting, buried in soil without rotting. Rockets, deep-sea submarines, artificial joints — only the harshest conditions earn it a place.

It has two other properties that map almost exactly onto what I want this magazine to be.

First, it is unreactive. At room temperature titanium barely reacts with anything. That isn't aloofness, it's slowness: a dense oxide film on the surface absorbs every assault on it. Good writing should work the same way — it doesn't chase every hot topic, but once you start, what you write has to survive being oxidized by time.

Second, its value lies in purification. Titanium ore is everywhere, but pulling titanium out of ore takes an enormous number of steps and costs a fortune. In this era an AI can produce an article in thirty seconds, and opinions are as cheap as gravel. What's genuinely scarce is the refined stuff: first-hand experience, independent judgment, and the things a person actually spent time using, thinking about, and verifying.

So we set the bar at element number 22 on the periodic table: every piece has to be worth keeping.

Take a stance on AI. Don't be a porter.

Too Much AI News, Too Little Attitude

Over the past two years I subscribed to nearly every Chinese AI publication, and then unfollowed about half of them. AI content on the Chinese internet today falls roughly into three camps.

The first is the re-post camp. The same hot topic gets hammered a hundred times within twenty-four hours; the headline gets shuffled and the body text isn't even changed down to the punctuation. The second is the anxiety camp. "Learn AI or lose your job." "Miss this wave of dividends and you won't see another for ten years." There are more people selling courses than teaching them. The third is the sneakiest, the content-farming camp: a headline built to shock you, and behind the click, AI-generated boilerplate that goes round and round like a cart wheel. You finish it feeling like you drank a bellyful of warm water and kept absolutely nothing.

What all three share is this: they're all saving the reader writing time, not reading time. Readers are getting more and more tired, good work gets buried, and you can't blame it all on the algorithm. When the supply side is all racing for speed, nobody is saving the reader any time. That's the real problem.

AITi wants to go the other way.

The Once-a-Week Pact

Why one piece a week rather than one a day? Because depth takes time.

A genuine first-hand piece takes a full week. A product teardown means interviews, digging through earnings reports, and paying for a subscription yourself for a month. A profile means landing the interview and sorting through hours of recordings. Publish daily and none of that is possible — you slide straight back into re-posting.

So AITi's promise is simple: one piece a week. Issue numbers run consecutively from 001, and a missed week is a broken promise. Every issue goes live simultaneously on the website getaiti.com and on WeChat — the site handles bookmarking, search, and sharing; WeChat gets it in front of you the moment it lands. Same content in both places; use whichever suits you.

What You'll Read Here

Not necessarily technical, but always about AI. Roughly four kinds:

• Deep observation: what's happening in the industry that's worth slowing down to think through properly;
• Product teardowns: taking an AI product apart to see what it's charging for, and what keeps it alive;
• People and stories: the people who build AI, and the people AI has changed;
• Hands-on tests: my own real workflows, the traps I fell into, the time I saved.

Topic selection follows three rules: first-hand experience beats secondhand posturing; the concrete beats the grand; write the things Chinese readers don't have. As for how AI technology itself will develop, I'm not going to play prophet — prediction is cheap, verification is what is scarce.

Human-Written, No Filler

Bad news first: I use AI. Researching, organizing interviews, catching typos, generating alternate headlines — AI does that work fast and well.

But the opinions are mine, the prose is mine, and any irritation or resonance you feel reading it should be mine too. AI here is a research assistant and an editor, not a ghostwriter. If I ever phone it in on an issue, you'll smell it between the lines — feel free to call me out in the comments.

Bad news stated. What's left is just doing the work. See you next week, starting with this issue.

— AITi