It looks like the reports of ‘laziness’ on the part of the ChatGPT AI bot were quite accurate, as developer OpenAI has just announced a fix for the issue – which should mean the bot takes fewer shortcuts and is less likely to fail do something halfway through the attempt.
The latest update to the ChatGPT code is “intended to reduce instances of ‘laziness’ where the model fails to complete a task” according to OpenAI. However, it’s worth noting that this only applies to the GPT-4 Turbo model, which is still in limited preview.
If you’re a free user on GPT-3.5 or a paid user on GPT-4, you might still notice a few issues in terms of ChatGPT’s capabilities – although we’re assuming the upgrade will eventually make its way to the other trickle down models too.
In December, OpenAI cited a lack of updates and “unpredictable” behavior as reasons why users may notice poor ChatGPT performance, and it appears that work to resolve these issues is still ongoing.
More thorough
One of the tasks that GPT-4 Turbo can now perform “more thoroughly” is code generation, OpenAI said. More complex tasks can also be completed from a single prompt, while the model will also be cheaper for users to work with.
Many of the other model upgrades mentioned in the OpenAI blog post are quite technical, but the bottom line is that these AI bots become smarter, more accurate, and more efficient. Many improvements concern “embeddings,” the numerical representations that AI bots use to understand words and the context around them.
ChatGPT recently got its own app store, where third-party developers can showcase their own custom bots (or GPTs). However, there are rules that prohibit certain types of chatbots, such as virtual girlfriends.
It also looks like OpenAI is busy pushing ChatGPT forward on mobile, with the latest ChatGPT beta for Android bringing the ability to load the bot from any screen (just like you could do with Google Assistant or Siri).