A naukri.com initiative
Insider
4w
102
Image Credit: Insider
AI improvements are slowing down. Companies have a plan to break through the wall.
- Some AI experts see a slowdown in AI model improvement, while others believe there is no wall and creativity is the key. The bottleneck in public data is one of the critical blocks for AI development, and AI progress requires more data. Using multimodal data, including visuals and audios, and private data are the solutions to overcome the bottleneck in public data. Researchers are focusing on enhancing the quality of data that an LLM is trained on, creating smaller and more efficient AI models. Companies will need to consider how much more expensive it will be to create the next versions of their highly-prized AI models.
- The rate of AI model improvement appears to be slowing, but some tech leaders say there is no wall.
- This is a trillion-dollar question for the tech industry.
- Some of the ways to keep the wheels of AI progress turning are tapping into new types of data, building reasoning into systems, and creating smaller but more specialized models.
- The internet is only so large.
- Private data will offer a path forward in the scramble to overcome the bottleneck in public data.
- Enhancing the quality of data that an LLM is trained on rather than just the quantity.
- AI reasoning has been an active area of research.
- Scaling laws work on a logarithmic scale rather than a linear one.
- AI leaders maintain that there are plenty of levers to pull to ensure models continue to improve.
Read Full Article
6 Likes
For uninterrupted reading, download the app