For several months, there have been complaints that OpenAI’s artificial intelligence has become a bit lazy and inaccurate. This is an endemic problem of AI and its nature. To solve it, the only recipe is to update and correct the code periodically.
In the official blog of OpenAI, they have just announced that GPT-4 Turbo has been updated to “complete tasks such as code generation more thoroughly than the previous preview model and aims to reduce cases of “laziness” where the model does not complete a task”.
Some ChatGPT users recently complained that the chatbot frequently refused to complete the tasks they requested and attributed the lack of updates to GPT-4.
However, the update from OpenAI is for GPT-4 Turbo, a more advanced version of GPT-4, an extended version of GPT-4 that was trained with information as recent as April 2023 and is only available in a preliminary version. Those who are using GPT-4, which learned from data available before September 2021, may still experience the same issues of sluggishness.
OpenAI stated in the post that more than 70% of users who use GPT-4 through its API have switched to GPT-4 Turbo due to its more updated knowledge base.
The company said there will be more updates to GPT-4 Turbo in the coming months, including the general availability of GPT-4 Turbo with vision. This will allow users to perform multimodal queries, such as text-to-image generation.
Habrá un día en que la IA se arregle a sí misma, or at least that’s what giants like Meta, Google, or Microsoft are aiming for. But that day has not yet arrived.