Competitive programming has always changed. Problems got harder. People got faster. Editors got better. The community got stronger. But I think something fundamentally different happened recently. AI changed the environment.

A few years ago, I had a relatively simple idea of what Codeforces ratings meant. If someone could consistently solve A, B, and C in contests, you could roughly expect them to eventually reach a certain level. The exact number was never universal, but there was a shared feeling of what different ratings represented.

Today, that intuition feels much less reliable. And I'm not talking only about cheaters getting high ratings. I'm talking about what happens to the people who don't cheat.

If some participants can use AI or other prohibited assistance to solve problems substantially faster or solve problems beyond their actual level, then everyone else is competing in a different environment. Codeforces is a relative ranking system, so this doesn't just create a few suspicious accounts at the top. It can affect the ratings of everyone around them.

There are already attempts to quantify this effect. One recent analysis estimated around 150–190 points of rating deflation among sampled likely-legitimate contestants over the period studied, although the author also discusses important limitations of the methodology. And this is where I started wondering:

What does 1600 actually mean anymore?

Maybe the more important question is not whether today's 1600 is "better" or "worse" than yesterday's 1600. Maybe it's simply a different measurement in a completely different era.

P.S. I used AI for some of the writing because of my English/language limitations. The ideas and opinions are still mine :)