Conveniently, clicking through to the actual data returns a 404.

That first chart isn’t even trying to hide that is fake. It’s depicting a perfect mirroring.

The post isn’t claiming perfection. It’s claiming production ready. Very different things.

You sure? You seem rather confused by the math, my friend.

As much as I don’t like OpenAI, I’m not looking forward to it going under. This new industry sucks, but it’s going to suck more when turns out only already established megacorps can afford to enter the market.

I’d bet it’s less simple input sanitizing and more 2 mistakes made separately because they don’t know any better.

1. The input field converting everything to a string indiscriminately
2. Because they did 1, converting everything back to the assumed type

If the front end Dev makes the first mistake, null would be sent in the body as “null”. Then on the backend, somebody might even be binding the variables correctly, but before hand realizing they have to deal with the market and rather than just have a conversation undoes it in their own code.

Real talk, what is the real barrier to somebody creating a competing publishing firm for these things.

I’m not a scientist, but I always hear about how expensive it is to either publish or get access to scientific papers without contacting the author directly. Why does that reputation exist? Why does it seem like the scientific community is so dependent on stuff like this?