OpenAI has launched GPT-4, an update over its original ChatGPT tool, which the company says features advnaced securithy and the ability to handle sensitive requests, making it closer to a human experience.
GPT-4 can analyze images and write code in all major programming languages, with many developers being able to create basic computer games in less than a minute using the tool.
While the previous version of ChatGPT was not able to reason and learn new things, with a lot of hallucinations, inappropriate or incorrect responses that is, the new generation relies on a ‘multimodal model’ that can accept longer text inputs of up to 25,000 words, as opposed to the earlier 3000 words, and is trained to be safer and more factual, OpenAI said.
Microsoft last month released the Bing AI chatbot using GPT-4, giving users a preview of the powerful tool’s capabilities.
So powerful is GPT-4 that it can take a mock website and turn it into an actual website by genertaing HTML and Javascript code.
Many industry leaders believe has GPT-4 has created the kind of disruption that browsers did in the early 1990s, in the nascent stages of the world wide web.
GPT-4 has proved its mettle in providing cogent answers for tech support, helping lawyers with lawsuit prep as well providing doctors with early diagnosis procedures.
With its wide display of knowledge, the GPT has also gone a ways in creating anxiety in people, understandably as humans gear up to compete with artificially-trained machines.
Real-time updation is another concern as GPT-4 still pulls information from a database that is limited to its last updation date.
The resulting time-gap could make accurate information a challenge.
At the moment, there is a waitlist to try out GPT-4 and users have to subscribe to ChatGPT Plus, a $20 monthly subscription, for premium access to the service.
GPT-4 currently has a cap of 100 messages every four hours, as it is staggered into a full release.
Stay updated with all the insights.
Navigate news, 1 email day.
Subscribe to Qrius
