OpenAI and Meta Could Make ‘Reasoning’ AI

Currently, all large-scale companies in the world are in some way artificial intelligence We see that he is trying to get a share of his cake. Some, like Google, take advantage of every opportunity that comes their way, while others, like NVIDIA, see opportunities and make other companies their customers. As a result, DALL-E and DALL-E, which we are used to using now, ChatGPT Tools like these appear.

But these tools have a major shortcoming. They cannot think. They are fed huge amounts of data and give us output based on that data. Although these outputs seem to have a logic behind them, the truth of the matter emerges when the cornering questions come. But OpenAI with Metaseems to overcome this logic barrier.

The next GPT will level up problem solving.

Speaking to the Financial Times, OpenAI COO Brad Lightcap said that the next GPT they will release will be ‘in problem solving’.reasoningHe said that he will make great progress in difficulties such as ‘. Brad stated that the ability of artificial intelligence models to reason has only recently been discovered.

A similar statement came from Meta managers. The executives in question made similar statements about the Llama 3 model that we will see in the coming weeks. For example, Joelle Pineau, who is in charge of Meta’s AI research, explains how the artificial intelligence model talks and make senseHe said that they thought he could plan and have memory.

This shows that language models, which largely fulfill the role of assistant, will be assigned to the meaning and reasoning side in the next stage. What do you think the consequences of this will be?

We say “Artificial intelligence will now make logic.” While he rejoices, the actions of artificial intelligence are: