On Monday, Anthropic debuted the Claude 3, a set of AI models said to be the fastest and most powerful yet. The new instruments are called Claude 3 Opus, Sonnet and Haiku.

The company said the most capable of the new models, the Claude 3 Opus, outperformed OpenAI’s GPT-4 and Google’s Gemini Ultra on industry benchmarks such as undergraduate knowledge, higher-level reasoning and basic math.

This is the first time Anthropic has offered multimodal support. Users can upload photos, charts, documents, and other types of unstructured data for analysis and answers.

The other models, the Sonnet and Haiku, are more compact and less expensive than the Opus. Sonnet and Opus are available in 159 countries as of Monday, while Haiku will follow soon, according to Anthropic. The company declined to say how long it took to train Claude 3 or how much it cost, but said companies like Airtable and Asana helped A/B test the models.

This time last year, Anthropic was seen as a promising generative AI startup founded by former OpenAI research executives. It had completed Series A and B funding rounds, but had only released the first version of its chatbot without any user access or much fanfare.

Twelve months later, it’s one of the hottest AI startups around, with backers including Google, Salesforce and Amazon, and a product that competes directly with ChatGPT in both the enterprise and consumer worlds. In the past year, the startup closed five different financing deals totaling about $7.3 billion.

The generative AI field exploded last year with a record $29.1 billion invested in nearly 700 deals in 2023, a more than 260% increase in deal value from a year earlier, according to PitchBook. It’s become the buzziest phrase in corporate earnings conversations quarter after quarter. Scientists and ethicists have expressed considerable concern about the technology’s tendency to spread bias, yet it has quickly made its way into schools, online travel, the medical industry, online advertising, and more.

Between 60 and 80 people worked on the core AI model, while between 120 and 150 people worked on its technical aspects, Anthropic co-founder Daniela Amodei said in an interview with CNBC. For the latest iteration of the AI ​​model, a team of 30 to 35 people is working directly on it, with about 150 people in total supporting it, Amodei told CNBC in July.

Anthropic said Claude 3 can summarize up to about 150,00 words or a large book (think: around the length range of “Moby Dick” or “Harry Potter and the Deathly Hallows”). Its previous version could only summarize 75,000 words. Users can enter large data sets and request summaries in the form of a note, letter or story. ChatGPT, on the other hand, can handle around 3,000 words.

Amodei also said that Claude 3 has a better understanding of risk in its responses than its previous version.

“In our quest to have an extremely harmless model, Claude 2 sometimes went overboard with the rejection,” Amodei told CNBC. “When someone comes up against some of the more spicy topics or the trust and the safety fences, sometimes Claude 2 would be a little conservative in his answers to those questions.”

Claude 3 has a more nuanced understanding of prompts, according to Anthropic.

Multimodality, or adding options like photo and video capabilities to a generating AI, whether you upload them yourself or create them using an AI model, has quickly become one of the hottest use cases in the industry.

“The world is multimodal,” OpenAI COO Brad Lightcap told CNBC in November. “If you think about the way we as humans process the world and engage with the world, see things, hear things, say things – the world is much bigger than the text. So for us it always felt incomplete for text and code to be the single modalities, the single interfaces that we could have to how powerful these models are and what they can do.”

But multimodality and increasingly complex AI models also lead to more potential risks. Google recently took its AI image generator, part of its Gemini chatbot, offline after users discovered historical inaccuracies and questionable answers that went viral on social media.

Anthropic’s Claude 3 does not generate images; instead, it only allows users to upload images and other documents for analysis.

“Of course, no model is perfect, and I think that’s a very important thing to say upfront,” Amodei told CNBC. “We’ve tried very hard to make these models the intersection of as capable and safe as possible. Of course, there will be places where the model still comes up with something from time to time.”

Clarification: Anthropic clarified with CNBC that Claude 3 can summarize about 150,000 words, not 200,000.

https://www.cnbc.com/2024/03/04/google-backed-anthropic-debuts-claude-3-its-most-powerful-chatbot-yet.html