In a recent statement, Sam Altman shared that he feels uncomfortable not using the smartest settings or models available, even if their speed is very slow.
Background
The current AI race is split into two paths: one optimized for speed (such as Small Language Models - SLMs) and the other optimized for complex reasoning. Altman suggests that the industry should focus more on the trade-off of cost vs. speed compared to cost vs. intelligence. This perspective reinforces OpenAI's direction of developing frontier models like GPT-4 and subsequent generations.
Why It Matters
Altman's statement highlights OpenAI's strategic direction: intelligence remains the number one priority. For Vietnamese users and businesses, this means we will continue to face "heavy" and slow models in the near future if we want the highest processing capabilities. However, cost-efficiency (price/speed) is also becoming an equally important topic to enable widespread AI adoption.