'today’s AI models are the stupidest they’ll ever be’

Bill gates AND sam altman on chatGPT, AI and regulations

Bill Gates sits down with the returned OpenAI CEO Sam Altman for the sixth episode of the Microsoft founder’s podcast, Unconfuse Me. In the episode, the duo dives into the past, present, and future of ChatGPT as a language model powered by artificial intelligence, including the regulations it may face soon. Bill Gates greets Sam Altman with, ‘I didn’t expect ChatGPT to get so good,’ referencing the model and its AI’s quick changes.

What started as a machine learning and language model has evolved into an AI medium that can hear, see, and speak. When OpenAI released ChatGPT-3, it did so with an Application Programming Interface (API) so that third-party apps and users could integrate the language model into their products and develop an entirely new application or help explore the advantages and limits of the technology. Soon, the smarter version of the language model came up, even prompting its branch of apps from OpenAI to face issues, including copyright and ethical concerns in art.

Bill Gates and Sam Altman during the sixth episode of Unconfuse Me with Bill Gates | image via GatesNotes

OpenAI has already released GPT-4, which it describes as more creative and collaborative than its predecessors since this one can generate, edit, and iterate with users on creative and technical writing tasks, including composing songs, writing screenplays, learning a user’s writing style, and more. During his conversation with Bill Gates for the sixth episode of the Unconfuse Me podcast, Sam Altman looks into the future of ChatGPT but says that their AI systems are on a continuous learning curve. The stuff that we’re seeing now is very exciting and wonderful, but I think it’s worth always putting it in the context of this technology that, at least for the next five or ten years, will be on a very steep improvement curve. These are the stupidest the models will ever be,’ he tells Bill Gates during their podcast interview, a few moments before he shares his outlook for his language model, one of which has already happened.

Bill Gates | video stills/images via GatesNotes and Youtube

The future of chatgpt according to sam altman

In the next two years, as Bill Gates specified, Sam Altman can see ChatGPT being multimodal, able to reason well, reliable, and customizable. The first one has already taken place with the launch of ChatGPT being able to generate videos and audio. The ChatGPT team is looking into expanding GPT-4 to break free from its limited reasoning, and they’re also focusing on reliability. The OpenAI CEO explains it as getting the best response out of the most questions asked about 10,000 times. ‘Customizability and personalization will also be very important. People want very different things out of GPT-4: different styles, and different sets of assumptions. We’ll make all that possible, and then also the ability to have it use your data. The ability to know about you, your email, your calendar, how you like appointments booked, connected to other outside data sources, all of that. Those will be some of the most important areas of improvement,’ he tells Bill Gates in the podcast interview.

Bill Gates interviews returned OpenAI CEO Sam Altman on his Unconfuse Me podcast

Regulations around AI discussed

Bill Gates steers the conversation toward regulations surrounding AI. Sam Altman says it would be easy to restrict AI with regulations but adds that if the technology goes beyond its means and impacts society, including the geopolitical balance of power, then a global agency to govern the issue is required, stating the International Atomic Energy Agency as an example for ensuring the peaceful use of nuclear technology. For this to happen, he sees an AI or system that exceeds far beyond what GPT-4 can do, ‘something with 100,000 or a million times the compute power of that,’ he tells Bill Gates to which the Microsoft founder responded with ‘I do think AI, in the best case, can help us with some hard problems, including polarization because potentially that breaks democracy, and that would be a super-bad thing. Right now, we’re looking at a lot of productivity improvement from AI, which is overwhelmingly a very good thing.’

Sam Altman on ChatGPT and its future, AI and the regulations surrounding artificial intelligence

So far, the EU Parliament has reached an agreement on its Artificial Intelligence Act, which regulates the use of AI, including banning social scoring, using AI to manipulate or exploit users, and the right of the users to launch complaints and receive ‘meaningful’ explanations. While this and the US White House released an Executive Order by President Joe Biden on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence within the country, implementing these AI regulations and rules, including the personnel monitoring the application, may still require concrete and necessary actions.

the full episode of Bill Gates interviewing Sam Altman on ChatGPT and AI is here