Sam Altman spoke at a conference on future advances in AI at Stanford University. The opportunity to talk about the development of GPT-5 and the boss of OpenAI promises: the next evolution of ChatGPT will be big.
Last year, GPT-4 was launched, the latest version of ChatGPT’s LLM, its engine in a way, the one that makes everything work. If OpenAI remains discreet about GPT-5, the company openly declares that this version of its AI is in development. Its CEO Sam Altman even revealed some details to Stanford University.
GPT-5 is “much smarter than GPT-4”
OpenAI’s plans seem bold, as Sam Altman stated that “GPT-4 is the stupidest model any of you will ever use again, by far“. He had already questioned his LLM’s reasoning ability and that seemed to be his priority. It’s hard to believe that he’s talking about competition from Google, for example: he explicitly implies that he’s talking about the future of GPT-4, namely GPT-5.
For him, GPT-4 is “a bit rubbish“, when GPT-5 should be “really good, of the materially better kind” And “much smarter than GPT-4“. In the ambitions previously discussed, Sam Altman had declared that he wanted ChatGPT to deal with video, the idea being to have a multimodal tool. Meanwhile, the video generation tool Sora has arrived. Also, Sam Altman really wanted to add “the ability to know you, your email, your calendar, how you want to make your appointments, connection to other external data sources“. The CEO even took the opportunity to talk about GPT-6, which will logically be better than GPT-5. No further details however.
OpenAI can spend lavishly and does not shy away from it
At the start of 2023, we learned that Microsoft had spent no less than 10 billion dollars in OpenAI, in order to obtain rights to use its artificial intelligence technologies. A nice loot that complements other investments: enough to spend without counting. OpenAI needs money: the company reportedly spends $700,000 per day just to run ChatGPT. In 2023, OpenAI will have spent more than $520 million on its chatbot.