Anthropic’s New Chatbot Fails to Impress: A Closer Look
Anthropic is a leading AI startup backed by industry giants Google and Amazon. It freshly revealed its latest offering: a family of models known as Claude 3. The company has made bold claims about the competencies of these models. They were affirming that they exceeded the performance of OpenAI on various standards. However, a recent review by TechCrunch proposes that these claims may not hold up under inspection.
Claude 3 Opus is the most capable model in the family. It is available through a subscription to Anthropic’s Claude Pro plan. This multimodal model is trained on a mix of public and exclusive text and image data. It claims a large context window equivalent to about 150,000 words. Despite these impressive stipulations, the model’s performance was found to be lacking in TechCrunch’s custom test.
The test included questions on a variety of subjects from politics to healthcare. It exposed that the model’s responses were not up to the mark. This unsatisfactory performance suggests that the academic standards used to evaluate these models may not precisely reflect the average user’s experience.
This raises important questions about the efficiency of current AI models and the standards used to evaluate them. While academic standards are useful for comparing models in a controlled environment, they may not capture the complexities of real-world use.
In conclusion, while Anthropic’s Claude 3 Opus may be a step forward in terms of technical specifications, it appears that there is still a long way to go before these models can deliver on their promises. As AI continues to evolve, it will be interesting to see how companies like Anthropic address these challenges to meet the high prospects of users.
Read More: Zora NFT AI Earning Model: Revolutionizing the AI Market
Read More: Anthropic Claude 3 AI outperforms OpenAI’s GPT-4
