In an interview with The Verge, OpenAI co-founder Ilya Sutskever explained the company's decision to keep details about GPT-4 confidential, citing competitiveness in the industry along with safety concerns.
Despite its name and initial goal to be a collaborative and open research organization, OpenAI has declined to disclose GPT -4's architecture and specific methods and data used to create the system, among other details.
Speaking to The Verge, Sutskever said many companies are seeking to develop large language models like GPT-4, which "took pretty much all of OpenAI working together for a very long time to produce."
- "On the competitive landscape front — it's competitive out there," he said.
- OpenAI's chief scientist also said AI models will eventually become so potent that it will become quite easy to cause "a great deal of harm."
- "And as the capabilities get higher, it makes sense that you don't want to disclose them," Sutskever added.
- When he was asked why OpenAI has altered its approach to open source, Sutskever replied, "We were wrong. Flat out, we were wrong. If you believe, as we do, that at some point, AI — AGI — is going to be extremely, unbelievably potent, then it just does not make sense to open-source. It is a bad idea."
- In a few years, he expects that it will be "completely obvious to everyone that open-sourcing AI is just not wise."
