ChatGPT: I believe that many netizens have heard of it before using it, and some even mentioned that this kind of generative artificial intelligence has the opportunity to make everyone unemployed, or even exterminate human beings! Recently, a website visited OpenAI’s CEO: Sam Altman, who talked about the ChatGPT roadmap. ChatGPT is amazing today, will it be even more Amazing in the future? You know the words of the moderator.
ChatGPT upgrades extremely fast
OpenAI’s ChatGPT generative dialogue artificial intelligence robot, since its launch in November last year (2022), has not only become a new technology that changes the world, but its rapid progress will scare you to death. Upgraded from GPT-3 to GPT-4 core, and recently added Bing, making it capable of surfing the Internet and processing data beyond 2021; adding Plugin makes ChatGPT like a smartphone, with different functions, such as moderator trial Cantonese speech can be generated from text with one click, and video can be generated from text with one click.
Sam Altman is good to talk
This time I interviewed Open AI CEO: Sam Altam’s organization is called humanloop, which itself is a company that develops artificial intelligence. In a blog on its website, it was mentioned that the author recently had the opportunity to talk to Sam Altam and about 20 artificial intelligence developers. The author describes Sam Altam as quite open and happy to share his views on AI with you. This time, the moderator also used ChatGPT 4 to translate the main content of this interview into Chinese. Sam Altman mainly mentioned the future roadmap of Open AI. After reading it, everyone can know how useful the future generative dialogue artificial intelligence machine robot can be.
OpenAI’s problems and future roadmap explained by Sam Altman
1. OpenAI is in a situation right now where it seems like many of their projects are stuck in a narrow single lane, which is the GPU. This caused OpenAI’s service to be unstable at times, and the speed was not fast enough, and it also affected some new features that OpenAI planned to launch.
2. OpenAI has a lot of plans, and they hope to provide cheaper and faster GPT-4 in 2023 (Editor’s Note: At present, you need to use ChatGPT 4 and Plugin services through the OpenAI website and App, and the monthly fee is US$40, ie Does that mean it will be cheaper in the future?), it’s like putting a more powerful engine in a more affordable car.
3. Sam Altman believes that many AI developers are very interested in ChatGPT’s Plugin development, but Sam believes that it will take a while for many Plugins to enter the market successfully. Just like some people may very much want to match the new smart watch accessories on their watches, but in fact those accessories may not be mature enough, and it may take time to adapt to the needs of the market. And Sam mentioned that while many people think they want ChatGPT in their mobile app, what they really want may be to have ChatGPT part of their app, just like the new smartwatch feature is the watch itself Part of it, not just a plug-in accessory.
4. The goal of OpenAI is to avoid competing with their customers, just like a restaurant does not want to open the same restaurant as its own customers. Sam said that Open AI should focus on ChatGPT, just like a restaurant chef, focusing on the dishes he is best at and making it the best. Their ideal is to make ChatGPT a super assistant for everyone at work, just like everyone has a super practical personal secretary.
5. Sam Altman hopes that future AI models will be subject to some regulation, just as he believes that future self-driving electric vehicles need to be properly regulated. Sam does not think that the existing AI model will pose a threat, just like the current electric vehicles can safely walk on the road under appropriate regulations.
6. Sam also emphasized the importance of Open Source, just like a famous chef willing to publish excellent recipes, so that more people can enjoy the food. He also said that OpenAI is considering open source GPT-3.
6. Although some people think that today’s AI models seem to have reached their limits, just like some people think that current sports cars are fast enough. But OpenAI’s data shows that they still have the potential to “take it to the next level”, but Sam tries to think that the development of future AI models may not be as sudden and fast as it was in early 2013 this year, and will be a little slower.
Article Source:humanloop.com