However, a study by Epoch AI suggests that tech companies are set to run out of public data used to train large language models between 2026 and 2032. This could lead to a bottleneck, hindering the scaling up of models, which has been a crucial aspect of expanding their capabilities and improving output quality. Despite this, Andreessen believes that the current situation is leading to the hiring of thousands of programmers, doctors, and lawyers to handwrite answers to questions for the purpose of training their AIs.
Key takeaways:
- Marc Andreessen, a key figure in Silicon Valley and general partner of venture-capital firm Andreessen Horowitz, believes that the rise of AI is creating a hiring boom, particularly for software engineers, rather than leading to unemployment.
- AI is facing a potential bottleneck as tech companies are projected to run out of public data to train large language models between 2026 and 2032, according to a study by Epoch AI.
- Andreessen suggests that the scarcity of data is leading companies to hire experts in various fields to manually craft answers to questions for the purpose of training their AI models.
- Nvidia CEO Jensen Huang asserts that while AI can take on some of a worker's responsibilities, it cannot fully replace anyone, indicating that human expertise is still crucial in the tech industry.