Are AI Models Getting Worse, or Are the Economics Catching Up?

AI models are improving on benchmarks, but many users still feel that the products are becoming less reliable.

In Episode 7 of System Prompt, we examine whether AI model quality is actually degrading, how API costs shape product decisions, and what growing competition means for businesses building on top of frontier providers.

The conversation explores the pressures facing companies such as OpenAI and Anthropic as they balance model performance, infrastructure costs, subscription limits, reliability, and market growth.

For enterprise users, the issue is larger than which model ranks highest.

Businesses need to understand the cost of each completed task, the risk of depending on a single provider, and what happens when pricing, rate limits, model behavior, or product access changes.

WHAT WE DISCUSS

• Why users may feel that AI model quality is declining

•