Meet Claude 2, OpenAI & Shutterstock and more

Pivot 5: 5 stories. 5 minutes a day. 5 days a week.

1. Anthropic's Claude 2 is the New Contender in Chatbot Arena

Anthropic, a startup in the tech industry, has unveiled Claude 2, a chatbot now available to the public in the US and UK. This new entrant stands as a competitor to established chatbots like OpenAI's ChatGPT and Google's Bard. Claude 2, an upgrade from its predecessor, offers enhanced conversational abilities and longer response generation.

The chatbot, powered by a large language model, is designed to assist with various tasks, including summarizing text, writing documents, and coding. However, like its counterparts, Claude 2 has its limitations, including the potential for incorrect answers and non-existent source citations. To mitigate this, Anthropic has iterated on Claude 2's safety, reducing the likelihood of harmful outputs.

Claude 2's distinguishing features include a more conversational tone and a sense of humor. It's also guided by a set of principles, or a "constitution," for self-revising responses. As it becomes globally available in the coming months, Claude 2 is set to make a significant impact in the realm of AI chatbots.

Read the full story here

2. OpenAI Extends Partnership with Shutterstock

OpenAI has extended its partnership with Shutterstock for another six years, allowing the AI company to train its models using Shutterstock's extensive library of images, videos, music, and metadata. This partnership, which began in 2021, has been instrumental in training OpenAI's text-to-image model, DALL-E.

Shutterstock has integrated OpenAI's image generator directly into its website and has expanded this integration to allow users to "edit and transform any image in the entire Shutterstock library." The company also plans to bring AI features to Giphy, the GIF-making platform it acquired from Meta earlier this year.

Despite concerns from artists about their work being used to train AI models, Shutterstock is fully embracing AI. The company has launched a "Contributor Fund" to compensate artists when their work is used to train OpenAI's models, and it has banned the sale of AI-generated images that weren't created using its built-in DALL-E tool.

Read the full story here

3. Prolific Secures $32M to Enhance AI Models

Prolific, a London-based startup, has raised £25 million ($32 million) to expand its operations. The company has built a system that uses a network of 120,000 human participants to inform and stress test AI models. This funding round was co-led by Partech and Oxford Science Enterprises (OSE).

Founded in 2014, Prolific already counts organizations like Google, Stanford University, the University of Oxford, King’s College London, and the European Commission among its customers. The company uses its network of participants to test new products, train AI systems, and determine whether their human-facing AI applications are working as intended.

Prolific pays its freelance participants a minimum of $6-$8 per hour, and it has paid out some $100 million to them to date. The company plans to use the new funding to expand its product and range of participants.

Read the full story here

4. Slack's Vision: Empower Everyone to Automate

Slack, the messaging software company, sees massive potential in generative AI and large language models for improving workplace productivity and efficiency. Steve Wood, Slack’s SVP, Product Management, shared this vision at the VentureBeat Transform 2023 conference.

Slack has embraced automation technologies and leveraged large language models from OpenAI and Anthropic to efficiently summarize busy channel activity. It also generates new workflows from the contextual information found in a company’s online discussion.

Wood highlighted the move towards “flexible, modular building blocks,” that would allow low- or no-code individuals to automate features of their favorite apps. He believes that the integration of outside information held in LLMs with the unique data found in individual companies’ Slack channels could be key to quickly unlocking bespoke business intelligence.

Read the full story here

5. Google Halts Development on Gen Z-Targeted Chatbot to Prioritize Bard

Google has quietly "deprioritized" its efforts on an AI-powered mobile chatbot app for Gen Z users. The app, called "Bubble Characters," featured interactive digital characters that would engage in conversations with users.

The company had been working on the app since Q4 2021, with the conversations powered by large language models. The app was designed to offer "human-like" conversations that "take action" and are "interesting for GenZ." However, work on the app has ceased as part of an internal reorganization at Google.

The shift in focus comes amid a broader reorganization within Google's Assistant organization, which has prioritized the development of the ChatGPT-competitor Bard. Some of the Bubble Characters team members were asked to pause their work on the Gen Z app to work on Bard ahead of its launch.

Read the full story here