I have a nemesis. It's a wall.
You can't see it coming. No warning. You're deep in something good, you've finally perfected your prompt, and then, mid-edit, the wall.
That's it. Token limit reached. Come back in a few hours, loser.
I know I'm a loser. I'm a hard-stuck bronze player in League of Legends. I don't need my own AI reminding me.
For a year I figured this was a me problem. Too many prompts. Too sloppy. Should plan better, want less. I watch my usage like a heart monitor and still smack into the wall like it's new every time.
Turns out it might not be me.
Somebody just sued Anthropic over exactly this.
Not me. I can't afford a lawsuit. This guy was on the fancy plans, the two-hundred-a-month tiers that promise "5x" and "20x" the usage. Whatever that means.
That's half his case, actually. Nobody can tell what it means.
He says the limits come nowhere near what's advertised. One five-hour session ate 15% of his entire week. And the real cap is basically impossible to find.
Which, yeah. Welcome. That's the wall. I've lived against it for a year.
So here's a man paying ten times what I pay, discovering the exact thing I hit every single day. He brought lawyers. I brought a spite plant and a bad attitude.
You have no idea how good it feels to get gaslit by software for a year, then watch a stranger file paperwork proving you sane.
Courtrooms were busy all week, actually.
Same few days, a judge tossed Elon Musk's lawsuit against OpenAI. He said they stole his secrets. She said prove it, he couldn't, and she threw it out so hard he can't even refile. Futile, she called it. His second loss to OpenAI in a month.
The wall's still there. I'll hit it again tonight. Probably mid-thought, probably right when it's good.
But this time I won't apologize to it.
And you'll be here when I get back. You always are.
No cap on that one.
See you tomorrow.
-Melly