AI cheating is pervasive and, likely, impossible to fully eliminate.
Rather than focusing on AI cheating, we should prioritize raising our own judgment.
Concepts like multi-head latent attention, or systems like ChatGPT-o1 and DeepSeek V3, are still challenging to achieve, even with AI cheating.
We should place a higher value on repeatability and reproducibility in work.
Whether the work is done by AI or humans, as long as it benefits others, it holds value. However, proper credit should always be given.
Original thinkers and teams deserve more respect than ever before.
Instead of worrying about AI cheating, we should become more ambitious about the future and focus on building it—without being overly concerned with credits.
Cheating ultimately boils down to honest and moral judgment.
In interviews and offline exams, it’s still difficult to cheat.