𝗧𝗵𝗲 𝗺𝗼𝗿𝗲 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗔𝗜 𝗴𝗲𝘁𝘀, 𝘁𝗵𝗲 𝗺𝗼𝗿𝗲 𝘄𝗲 𝘂𝘀𝗲 𝗶𝘁.

Sounds backwards, but it's not

In 1865, economist William Jevons had already observed something peculiar: as steam engines became more efficient at burning coal, coal consumption went up, not down

Why? Efficiency unlocked new use cases. Suddenly, coal made economic sense for industries that couldn't afford it before

That's 𝗝𝗲𝘃𝗼𝗻𝘀' 𝗣𝗮𝗿𝗮𝗱𝗼𝘅. And it's playing out now in tech

🔹 𝗖𝗹𝗼𝘂𝗱 𝗰𝗼𝗺𝗽𝘂𝘁𝗶𝗻𝗴: It didn't cut spending as the investment was cheaper. It made way for millions of startups that couldn't afford to own servers. It enabled millions of startups that couldn't afford servers

🔹 𝗦𝘁𝗼𝗿𝗮𝗴𝗲: The price per GB of hard drives has reduced. Total data stored exploded

🔹 𝗔𝗜 𝘁𝗼𝗸𝗲𝗻𝘀: As costs dropped 10x, usage went up 100x. Companies now use more tokens per task because they can afford higher accuracy

Most AI tokens in the future will be used on things we don't even do today. Software projects that wouldn't have been started. Contracts that wouldn't have been reviewed. Research that wouldn't have happened

The pattern repeats:

When mainframes cost millions, only governments used them. Minicomputers brought tens of thousands of users. PCs brought hundreds of millions. Smartphones brought billions

👉 𝗪𝗵𝗮𝘁 𝘁𝗵𝗶𝘀 𝗺𝗲𝗮𝗻𝘀 𝗳𝗼𝗿 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿𝘀:

AI won't replace programmers. It will expand what programming means

This means, if AI makes programmers 10x more productive, but creates 20x more opportunities to apply programming, we'll need twice as many developers

The "programmable surface area" of business keeps growing. Every company needs AI agents. Every workflow needs automation. Every decision needs data

𝗧𝗵𝗲 𝗿𝗲𝗮𝗹 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻 𝗶𝘀𝗻'𝘁 𝘄𝗵𝗲𝘁𝗵𝗲𝗿 𝗔𝗜 𝘄𝗶𝗹𝗹 𝗿𝗲𝗽𝗹𝗮𝗰𝗲 𝘆𝗼𝘂

It's whether you'll use it to do things that weren't possible before