Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

OpenAI launched its best new AI model in September. It already has challengers, one from China and another from Google.

Dec 20, 2024 - businessinsider.com
In September, OpenAI introduced the o1 model, a breakthrough in AI that uses inference-time compute to tackle complex questions by breaking them into manageable tasks. This approach allows the model to backtrack and correct errors, mimicking human problem-solving. By November, a Chinese AI lab released DeepSeek, a similar model that reveals its reasoning steps to users, unlike o1. DeepSeek's ability to solve complex problems and show its thought process was praised by AI researchers.

In December, Google launched Gemini 2.0 Flash Thinking, another model using inference-time compute, which also displays its reasoning steps. This rapid development of similar AI models highlights the competitive and fast-paced nature of the AI industry, where innovations quickly become commoditized. The proliferation of such models raises questions about the sustainability of investing large sums in developing new AI technologies, as the cost of accessing these tools has significantly decreased.

Key takeaways:

```html
  • OpenAI's o1 model, introduced in September, was a breakthrough in AI, using a new approach called inference-time compute to tackle complex questions.
  • By November, a Chinese AI lab released DeepSeek, a similar model that shows users each step of its thought process, unlike OpenAI's o1.
  • Google launched Gemini 2.0 Flash Thinking, another competitor that also reveals its reasoning steps, highlighting the rapid commoditization of AI innovations.
  • The swift proliferation of similar AI models raises questions about the sustainability of high investment costs in developing new AI technologies.
```
View Full Article

Comments (0)

Be the first to comment!