Through a post that appeared on the official blog of OpenAI was confirmed as GPT-4 Turbo has been updated, allowing the platform to generate code in much more depth than before.
The developers’ efforts, specifically, have focused on optimizing the use of GPT-4 Turbo by making it less “lazy” compared to the past. Despite this, the company Sam Altman it didn’t go into much specifics about implementations.
The update we are talking about seems to offer a concrete response to ChatGPT users who have recently complained about a chatbot who often refuses to complete requested tasks. However, it should be underlined that the update concerns GPT-4 Turbo (trained with data up to April 2023) and that the introductions do not affect GPT-4, potentially still subject to refusal to delve deeper into some prompts.
Launching improved GPT models in the API, as well better and cheaper embedding models (e.g. one model is 5x cheaper than our previous embedding models with stronger performance): https://t.co/rFlFLFFJ4e
— Greg Brockman (@gdb) January 25, 2024
GPT-4 Turbo updates: more active chatbot and the introduction of embeddings
OpenAI focuses heavily on GPT-4 Turbo and, in this regard, confirmed in the same post as beyond 70% of ChatGPT users have left the previous version aside, given the more up-to-date data offered. On the other hand also Microsoft Copilotsince last December, now bases its operation on GPT-4 Turbo.
Regarding the update, the company has made it clear that it will not be the last, with the possible future introduction of a system for generating images through text.
Recent OpenAI implementations also include some smaller AI models, referred to as embedding. OpenAI describes them as a “A sequence of numbers that represents concepts within content such as natural language or code“. These should support applications that use the increased recovery generationa type of AI that draws from a database to provide answers, without having to create textual output from scratch.