Better than GPT-4? AI startup company Anthropic launches chatbot Siasun Robot & Automation Claude 3.

This week, the company Anthropic released Claude 3 for the first time, which is an artificial intelligence model. The company claims that this is the fastest and most powerful artificial intelligence model to date. The new tools are named Claude 3 Opus, Sonnet, and Haiku. According to the company, in industry benchmark tests, the performance of the most powerful new model, Claude 3 Opus, exceeds that of OpenAI's GPT-4 and Google's Gemini Ultra in areas such as undergraduate-level knowledge, graduate-level reasoning, and basic mathematics. This is the first time that Anthropic is offering multimodal support. Users can upload photos, charts, documents, and other types of unstructured data for analysis and answers. Other customized models such as Sonnet and Haiku are smaller and have lower prices than Opus. According to Anthropic, Sonnet and Opus will be available in 159 countries starting next Monday, while Haiku will be available soon. The company declined to disclose the time or cost spent on training Claude 3, but mentioned that companies like Airtable and Asana helped with A/B testing on these models. This time last year, Anthropic was considered a promising generative artificial intelligence startup founded by former OpenAI research executives. They have completed Series A and Series B funding rounds, but only launched the first version, the chat Siasun Robot&Automation, without any consumer visits or large-scale promotions. Twelve months later, it has become one of the hottest artificial intelligence startups with support from Google, Salesforce, and Amazon, competing directly with ChatGPT in both enterprise and consumer fields. In the past year, this startup has completed five different funding transactions with a total amount of approximately 7.3 billion US dollars. According to PitchBook data, the field of generative artificial intelligence has seen explosive growth in the past year, with a record-breaking total investment of 29.1 billion US dollars in nearly 700 transactions in 2023, an increase of over 260% compared to the year before. It has become a popular term on quarterly earnings conference calls for companies. Academics and ethicists have expressed great concern about the bias in the spread of this technology, but even so, it has quickly entered schools, online travel, healthcare, online advertising, and other fields. Daniela Amodei, co-founder of Anthropic, said in an interview that 60 to 80 people are engaged in the development of core artificial intelligence models, and 120 to 150 people are working on the technical side. Anthropic says that Claude 3 can summarize about 15,000 words, or the length of a fairly thick book (approximately in the range of Moby-Dick or Harry Potter and the Deathly Hallows). Its previous version could only summarize 75,000 words. Users can input large datasets and request summaries in the form of memos, letters, or stories, while ChatGPT can handle only about 3,000 words. Amodei also stated that compared to the previous version, Claude 3 has a better understanding of dealing with risks. In pursuit of a highly harmless model, Claude 2 sometimes over-rejected requests, tending to be more conservative in answering when encountering sharp topics or touching on trust and safety barriers. According to Anthropic, Claude 3 has a more nuanced understanding of prompts. However, multimodal and increasingly complex artificial intelligence models also bring more potential risks. Recently, Google pulled its artificial intelligence image generator offline because users found content that was historically inaccurate and problematic replies, which spread widely on social media. Anthropic's Claude 3 does not generate images, only allowing users to upload images and other files for analysis.