Author: IBL News

IBL News | New York

Today, Anthropic, the creator of the Claude 2 chatbot announced that Amazon / AWS will invest up to $4 billion in the company and become the primary cloud provider.

The e-commerce group said it will initially invest $1.25 billion for a minority stake in Anthropic, which also counts Google Spark Capital, Salesforce, Sound Ventures, Menlo Ventures, and Zoom among its investors. Anthropic has raised a total of $2.7 billion to date.

As part of the deal, Amazon said it has the option to increase its investment in Anthropic to a total of $4 billion.

Simultaneously, Amazon Web Services (AWS) will make Anthropic AI accessible to its customers while collaborating on the development of future Trainium and Inferentia technology.

On its side, Anthropic will expand its support for Amazon Bedrock, providing secure model customization and fine-tuning the service to optimize Claude’s performance.

“Organizations will be able to use Claude 2 for a wide range of tasks, from sophisticated dialogue and creative content generation to complex reasoning and detailed instruction. Our industry-leading 100,000 token context window will help them securely process extensive amounts of information—including technical, domain-specific documents for use cases across finance, legal, coding, and more,” said Anthropic.

Companies already building with Anthropic models on Amazon Bedrock are LexisNexis Legal & Professional (for conversational search, insightful summarization, and intelligent legal drafting capabilities), asset management firm Bridgewater Associates (for developing an investment analyst assistant to generate charts, compute financial indicators, and create summaries of the results), and Lonely Planet (for synthesizing its decades of travel content to deliver cohesive, highly accurate travel recommendations.)

Experts say this partnership signals growing enterprise demand for safer, more controllable AI.

What’s going on here?
Amazon will invest up to $4 billion in Anthropic and become their primary cloud provider, gaining access to their AI assistant Claude.

What does this mean?
Amazon’s investment enables Anthropic to scale its compute resources and continue advancing AI safety… https://t.co/nM8OnZ00nv

— Ben Tossell (@bentossell) September 25, 2023

IBL News | New York

OpenAI, this week, unveiled an upgraded version of its text-to-image tool, DALL-E 3, that uses ChatGPT to help fill in prompts or descriptions of images in this case.

It will be issued for premium ChatGPT users in October. The company didn’t say when it plans to release a free web tool, as it did with DALL-E 2 and the original DALL-E model.

Subscribers of paid ChatGPT Plus and ChatGPT Enterprise will be able to type a few words for an image and hone it through conversations with the chatbot, making it more descriptive, and providing way a more guidance to the DALL-E 3 model.

The upgraded DALL-E 3 also generates higher-quality images that more accurately reflect prompts, OpenAI says, especially when dealing with longer prompts.

In terms of security, the company said that artists can now opt out of having certain — or all of — their artwork used to train future generations of OpenAI text-to-image models.

The launch of DALL-E 3 comes as OpenAI faces competition from Midjourney and Stability AI.
.

IBL News | New York

Character.ai, the app that lets users design their own characters, reached 4.2 million monthly active users in the U.S., compared with nearly 6 million monthly active users of ChatGPT’s mobile apps, according to Similarweb. That’s notable growth since its launch in May 2023.

Character.ai is attracting a younger demographic audience, drawing in nearly 60% of its audience from the 18- to 24-year-old age bracket.

Other AI providers, such as Perplexity.ai, Midjourney, Anthropic, and Bard’s, also attract young population demographics

Backed by an investment of over $150 million led by Marc Andreessen’s venture capital firm Andreessen Horowitz (a16z), Character.ai was valued at $1 billion in March 2023.

The startup was created by Noam Shazeer and Daniel De Freitas, AI experts who previously led a team of researchers at Google that built LaMDA (Language Model for Dialogue Applications), a language model that helps power conversational AI experiences.

IBL News | New York

Microsoft, yesterday at an event in New York, announced a new unified Copilot experience in Windows 11, starting on September 26th as a free update (version 22H2), labeled by the company as “your everyday AI companion.”

Designed to navigate any task, assist, and deliver contextual information, the Microsoft Copilot tool will be available across Windows 11 desktop, Microsoft 365, Outlook, the Edge browser, and Bing.

Applications like Paint, Cocreator, Photos, Bing Image Creator, and Clipchamp will be featured with AI capabilities, too.

• Bing will include the latest DALL.E 3 model from OpenAI, with a better understanding of prompts and more accurate images.
• Bing Chat will contain personalized answers as the assistant will include responses from the user’s chat history. It will have Microsoft Designer integrated.
• Bing Chat Enterprise will benefit from a new update. It will be available in the Microsoft Edge mobile app.
• Bing and Edge will be connected to Microsoft Shopping.
• Microsoft 365 Copilot will be generally available for enterprise customers on Nov. 1, 2023, along with Microsoft 365 Chat, a new AI assistant.
• New Surface devices that bring all these AI experiences, too.

IBL News | New York

Google, yesterday, released an upgraded version of its chatbot Bard, which can access information from Google apps and services.

• The new Bard can now retrieve and help users work with real-time info from Maps, YouTube, Hotels, and Flights, extensions that don’t leverage personal data. These extensions are enabled by default, and you can disable them at any time.

• In addition, users can enable Bard to interact with information in English from Gmail, Docs, and Drive to summarize and answer questions across personal content. At this point, Google ensured that this Google Workspace data won’t be used to train Bard’s public model and can be disabled at any time.

• Another feature is based on double-checking Bard’s AI-generated responses with Google Search (with the “Google It” button) to see if there’s content on the web to substantiate the answer. When a statement can be evaluated, users will see it highlighted in Bard’s response and will be able to click to learn more. But if the AI is unsure, the sentence may be highlighted in orange to indicate that it knows this part of the answer might be wrong. This should help users better understand when the AI is “hallucinating” providing a response based on false information. The feature will also help the AI improve as it learns what it gets wrong from user feedback and then uses that to create a better model.

• Building off of conversations shared, when someone shares a Bard conversation through Bard’s public lin- sharing feature, that author can continue that conversation in his account and build off of what they started.

• Users can upload images with Google Lens, get Google Search images in responses, and modify Bard’s responses to be simpler, longer, shorter, more professional or more casual in all supported languages.

A New York Times analyst examined all these features, and the results were mixed, not as promising as Google advertised.”Google’s Bard Just Got More Powerful. It’s Still Erratic,” Kevin Roose wrote.

Our own IBL News testing found that Bard still hallucinates in large part when retrieving information.

 

IBL News | New York

Anthropic, the AI company backed by Google, launched a paid plan for its Claude.ai chatbot in the US and UK this month.

Priced at $20 (or £18) per month, the new Claude Pro option offers priority access during high-traffic periods, early access to new features, and the ability to send 5x more messages than the free tier.

“This means you can level up your productivity across a range of tasks, including summarizing research papers, querying contracts, and iterating further on coding projects—like this recent demo of building an interactive map [see it below],” said the company.

IBL News | New York

OpenAI this month introduced the Canva plugin on the ChatGPT store, letting paid users create visuals like logos and banners.

This integration is currently exclusive to ChatGPT Plus subscribers, who are charged a monthly fee of $20.

These users can use Canva with ChatGPT, but the process requires a lengthy process with many clicks.

In the chatbot’s prompt box, ChatGPT’s customers describe the visuals they want to generate. For example: “I am a smartphone enthusiast active on Facebook. Create a banner.”

Users will then be taken to Canva, where they can edit and download the AI-generated visual by tapping on the ‘Share’ button and choosing the ‘Download’ option.
.

IBL News | New York

At its Dreamforce customer conference in San Francisco this week, Salesforce.com announced Einstein Copilot Studio, a tool that consists of three elements to build custom generative AI applications with low-code knowledge: prompt builder, skills builder, and model builder. It will be available as a pilot this fall, according to the company.

• The prompt builder allows customizing and creating templates for clients’ product or brand offerings that aren’t available out of the box with Einstein GPT or Einstein Copilot.

• The skills builder enables companies to add actions to prompts. Some examples include running a competitor analysis or objection handling.

• The model builder lets enterprise companies bring their model or use one of the supported third-party offerings like Anthropic, Cohere, Databricks, Google Cloud’s Vertex AI, and OpenAI.

Einstein Copilot Studio says that it makes accessible all the components from across the Salesforce platforms.

To deal with hallucinations — when a model makes up an answer or provides inappropriate responses — and help companies manage security, governance, and data privacy, Salesforce is building a system called “the Einstein Trust Layer.”
.

• Salesforce News

IBL News | New York

Stability AI announced the launch of Stable Audio, a model that uses generative AI techniques to deliver music and sound effects via a web interface, this week.

Audio tracks, songs, and sound effects, at high-quality 44.1 kHz, are generated by the user by writing a text with a descriptive prompt and the desired length of audio.

The underlying model was trained using music and metadata from AudioSparx, a leading music library. It was trained on 800,000+ tracks and effects.

For instance, “Post-Rock, Guitars, Drum Kit, Bass, Strings, Euphoric, Up-Lifting, Moody, Flowing, Raw, Epic, Sentimental, 125 BPM” can be entered with a request for a 95-second track, and it would deliver this track:

The company provided more samples of generated tracks on its announcement page.

Stable Audio’s free version offers to create and download tracks of up to 45 seconds, while the ‘Pro’ subscription – at $12 per month — delivers 90-second tracks that are downloadable for commercial projects.  Users can try the model at www.stableaudio.com.

According to Techcrunch, in the Stable Audio terms of service agreement, customers agree to indemnify Stability in the event intellectual property claims are made against songs created with Stable Audio.

This London–based company has raised $125 million and achieved a valuation of  $1 billion.

IBL News | New York

Slack will introduce this winter its AI tool which will instantly generate highlights and summaries of the conversations. These new generative AI capabilities were announced by the parent company, Salesforce, this month.

AI-generated summaries will also list the next steps users can take based on concerns, comments, and suggestions posted by people in the conversation.

In addition, Slack AI will produce elaborate answers when people search for information on the platform.

At the moment, searching on Slack can only bring up messages, files, and channels with the keyword.

Users will be able to integrate their AI language model of choice, using partner-built apps from OpenAI’s ChatGPT or Anthropic’s Claude.

Only 27% of companies are currently using AI tools, according to the latest State of Work research.