Open AI: GPT Development Slows Down. False Promises of Breakthrough?

Open AI: GPT Development Slows Down. False Promises of Breakthrough?

 Open AI is probably conning their investors as the rate of development slows down. Mircosoft has just sanctioned 14 billion dollars for the Open AI research and development. Other giant corporations like Apple and Google are also interested in AI but Microsoft is very deeply involved in it. No other corporation is taking as big of risk in AI than Microsoft. Even though Apple has been working on its Apple Intelligence but they are taking a more risk-averse approach by not overselling its capabilities. They don't want to lose their Iphone sails with this new trend. AI is useful at times but it's not the end for Apple. Microsoft is more ambitious when the topic of AI is brought up.

For a time, it looked like a big leap forward with each new iteration of their AI model GPT. From GPT 1.0 to GPT 3.0, it felt like an utterly revolutionary rate of development. Since the release of GPT 3.0, things have been a bit stale from their end.

 

 

Next Flagship AI model under development is given the project name Orion and the anticipation has been great for it. It's still an improvement over the GPT 4.0 but the leap forth will not be as huge as before. 

GPT 5.0 is not even coming in 2024 as revealed by Open AI Sam Altman (C.E.O). Here for more details.

Instead, Open AI will be prioritising the codename project Strawberry, a.k.a GPT-o1. This one is different from the previous models as the main focus of this project is deal with mathematical and scientific problems. A more nuanced and methodical approach to problem solving will be employed as the foundation technology for this model. Accuracy and Logical Reasoning for this model will be polished up.

5 Websites to Make your Day Brighter

Reportedly, Open AI and Sam Altman are resorting to fresher strategies to tackle the slowdown in the development of newer versions of GPT. Some areas may even see stagnation like coding, for instance. 

One technique, developers at Open AI are looking to employee is to train the newer model on the synthetically generated data by the existing models to hone their computational capabilities. Post-training processes are also expected to go under newer algorithms. Company has no plans of releasing the new model, Codenamed Orion this year in 2024. 

 

Submarine Cables and Internet Connectivity