Dario Amodei

I am very concerned about deploying such systems without a better handle on interpretability. These systems will be absolutely central to the economy, technology, and national security, and will be capable of so much autonomy that I consider it basically unacceptable for humanity to be totally ignorant of how they work. (source)

From a business perspective, the difference between open and closed is a little bit overblown. From a security perspective, the difference between open and closed models is, for some intents and purposes, overblown. The most important thing is how powerful a model is. If a model is very powerful, then I don’t want it given to the Chinese by being stolen. I also don’t want it given to the Chinese by being released. If a model is not that powerful, then it’s not concerning either way. (source)

Predicting the exact timeline for achieving a general AI system more capable than any living human is extremely difficult. While significant progress is being made, the complexities and unknowns in AI research make it challenging to set concrete dates. Instead, we should focus on making AI safe and beneficial.

While expanding large language models is an interesting direction, there are still many challenges to achieving artificial general intelligence. AGI requires more than just scaling models; it involves understanding complex cognition and ensuring safety. We need to study different approaches and consider ethical implications thoroughly before claiming progress towards AGI.