GPT-4 announced by OpenAI: it’s outstanding | | | | Turtles AI

  OpenAI, the research laboratory founded by tech giants including Elon Musk, Reid Hoffman, Peter Thiel, and Sam Altman, has launched its latest version of the Generative Pre-trained Transformer (GPT). GPT-4, which is available from March 14, 2023, is now being integrated into ChatGPT Plus and select 3rd party partners via its API. OpenAI announced that GPT-4 has made significant improvements over its predecessor, GPT-3.5, by accepting image prompts and offering video and animation features in the future. GPT-4 has demonstrated exceptional reasoning abilities by scoring in the 90th and 99th percentile for the bar exam and Biology Olympiad, respectively, placing it on par with the brightest human students. OpenAI stated that GPT-4 can now understand and output text and images as well. It can accept inputs consisting of interspersed text and images and generate text outputs over a range of domains, including text and photographs, diagrams, or screenshots. For instance, when an image of balloons floating in the sky is uploaded and users ask GPT-4 what would happen if the strings were cut, the model can reason that the balloons will fly away into the sky. Another example provided by OpenAI showed that GPT-4 could even detect what's strange in a deliberately strange image that depicted a man ironing clothes while attached to the back of a taxi. GPT-4 has also demonstrated its improved reasoning abilities. For example, it can find the open window for when three people with different schedules are available for a 30-minute meeting, which can be complicated even for a human to figure out quickly. GPT-4 was able to offer a proper solution that works for all three, unlike GPT-3.5, which failed in suggesting an appropriate time because it conflicted with one person's schedule. It can even turn a hand sketch into a working HTML website. Furthermore, OpenAI has stated that GPT-4 will be integrated into Microsoft services, including Bing. Microsoft has invested over $10 billion in OpenAI, which highlights how serious it is about the coming generative AI revolution and in going after Google. In response, Google made a clumsy release announcement for its own AI-powered search engine called Bard, which at the moment looks underwhelming, to say the least. In conclusion, the launch of GPT-4 has marked a significant milestone in the development of chatbots and AI technology. The ability to accept image prompts and provide video and animation features in the future means that the model has the potential to become even more versatile in its application. With its exceptional reasoning abilities, GPT-4 is set to revolutionize the way we interact with technology and improve the quality of our lives.