Why did ChatGPT get ‘lazy’ in December?
And what can we do if it happens again?
What happened?
People are reporting (mostly on social media) that ChatGPT is either refusing to perform difficult and resource-intensive tasks or is telling the user how they can do it themselves.
Coders have said that ChatGPT now refuses to write more than a few dozen lines of code. Data analysts say that ChatGPT refuses to perform “lengthy tasks” such as generating a table with a dozen columns. Basically, if you give it a complex task, instead of doing it (which it could in the past) it does part of the work and gives you instructions on how to do the rest.
However, in mid-December, some people posted on X that ChatGPT has gotten better and faster if you force it to do the work. But other people are still complaining that it is slow and moody. So, who knows how long this issue will persist?
What has OpenAI said?
If they know why this is happening, they are not admitting it.
