ChatGPT maker OpenAI Just Announced it Latest Model
OpenAI just shocked the AI world with ‘o1’, its new AI model which the company claims "thinks" before responding — and it outperforms humans on PhD-level science questions.
The new model, o1, marks a major step in AI scaling by increasing the compute time for inference, allowing the model to process questions more thoroughly before answering. This approach enhances reasoning capabilities, especially for logic-based tasks, by spending more time "thinking" according to the company.
While not universally superior to previous models like GPT-4o, o1 is designed for scenarios where deeper logical processing is valuable. OpenAI has also released o1-preview, a trial version, and o1-mini, a cost-effective model targeting STEM (Science, Technology, Engineering, and Mathematics) applications. These models outperform GPT-4 in science and math; for example, they scored 83% in the International Math Olympiad qualifier, while GPT-4 managed only 13%. Both are available now for ChatGPT Plus and Team users, with broader access planned soon.
Our Thoughts:
First of all, it's a bit worrying to see more and more anthropomorphism, but we can't underestimate the significance of this moment.
The idea of a model capable of doing more than just recognizing patterns, but actually thinking before providing answers is pretty cool.
If this model lives up to the hype, it has the potential to unlock breakthroughs in fields like medicine, engineering, and possibly other challenges we haven’t been able to solve.
The future is bright.!!!
Google’s robots master shoelace-tying
Google DeepMind has just unveiled two new AI systems that advance robot dexterity. Now, machines can perform complex tasks that require dextrous movement, such as tying shoelaces and hanging shirts.
领英推荐
The details:
These systems employ diffusion methods similar to image generation models to anticipate robot actions from random noise. In simulations, robots achieved a 98% success rate, and in real-world tasks such as cube reorientation, they achieved up to 97% success.
Why it matters:
Although robots still can't do things as well as humans, these improvements are a big step toward making robots that can do everyday tasks better. Using image techniques in robotics also shows that getting better in one part of AI can help other parts improve too.
QUICK HITS
Trending AI Tools
THAT’S A WRAP
Co-Founder of Altrosyn and DIrector at CDTECH | Inventor | Manufacturer
2 个月The focus on "solving complex problems" is narrow. What about models that foster creativity or ethical reflection? Consider the recent backlash against AI-generated art; does o1's emphasis on utility neglect these concerns? How might o1 be used to bridge divides rather than simply solve technical challenges?