I'll make this a point by point dump of what I can remember. I haven't fact checked so if I'm wrong, let me know please.
Timeline:
OpenAI wanted to build a ChatGPT “GPT” to replace programmers
They hired 2000 programmers to train their GPT
No updates for 2 years after that
OpenAI’s flagship launches 20$ per month
A new tier emerges with 200$ per month
Reports that they are still not profitable
Reports that they want to launch a 2000$ per month product to replace programmers
They intend to buy Windsurf for 3 billion, which everyone laughed at
They made an offer to buy it for just that.
Windsurf is a VS Code fork that's allegedly awesome with a very good sales team and are growing rapidly.
My experience with VS Code and VS Code Online (vscode.dev) has been nothing sort of awesome. It's a free open source product and has this extensions library which has pretty much everything. Reminds me of Chrome when I first used it 16 ago. Also the web version allows you to bypass organisational restrictions on new software installation.
Side note: Did you know the new Google Code Editor is just a VS Code fork?
Now why would OpenAI pay so much for a VS Code fork?
Because it's where the 2000$ per month paying audience is. Organisations aren't going to fork out money just like that, they will need a robust product with a lot of data, much more than 2000 developers can provide.
Also, with how awesome VSCode ecosystem is, building on top of it and improving it is going to deliver a lot of value.
Hence 3 billion dollars in funny OpenAI money sounds quite reasonable if your org can do decent job to replace a programmer down the line.