Venture capitalists Marc Andreessen and Ben Horowitz also expressed doubts about the continued rate of AI model improvements, citing a lack of new high-quality human data and energy sourcing issues. If the current method for improving AI models is no longer effective, the attainment of AGI may be delayed. Despite this, Altman remains optimistic, possibly due to OpenAI's potential deal with Microsoft if AGI is achieved.
Key takeaways:
- There are growing doubts about the ability of AI companies to continue improving models at the same rate as they have in recent years, with signs that gains from adding more data, computing power, and training time have slowed down.
- OpenAI CEO Sam Altman remains optimistic about achieving artificial general intelligence (AGI) by 2025, despite these challenges and skepticism from others in the industry.
- Some factors holding back AI model improvements include a lack of new high-quality human data and problems sourcing the extra energy needed to power AI data centers.
- Other technology trends, like Moore's Law, have also stopped working after years of steady progress, which could be a warning sign for the AI industry.