Gemini Live’s Innovative Features Poised to Challenge Apple Intelligence
Google’s upcoming AI assistant, Gemini Live, is generating buzz with its unique capabilities that could surpass Apple’s Intelligence. Unveiled at Google I/O 2024, Gemini Live promises “natural conversations” and a seamless integration with users’ daily activities. The latest reports reveal Gemini Live’s “background mode,” allowing users to continue conversations while using other apps or with their screen locked. This feature effectively turns the AI experience into a phone call, enabling users to multitask while interacting with the assistant.
Furthermore, Gemini Live is set to offer quick access through the smartphone’s lock screen, with extensions for Google Maps, Flights, Hotels, and YouTube. Users may be able to ask for directions, book travel, or control smart home devices directly from the lock screen.
While details are still scarce, these innovative features could give Gemini Live an edge over Apple Intelligence, which requires constant user interaction. The background mode and lock screen integration could make Gemini Live more accessible and user-friendly.
The release timeline for Gemini Live remains uncertain, but it is expected to arrive in the coming months for Gemini Advanced subscribers. As always, industry experts advise taking these leaks with a grain of salt, as plans can change.
Ex-Googler and Filmmaker Launch DreamFlare, a Platform for AI-Generated Video
A new startup called DreamFlare AI, co-founded by former Google employee Josh Liss and filmmaker Rob Bralver, aims to help creators make and monetize short-form AI-generated videos. Unlike other tools, DreamFlare serves as a studio where creators collaborate with storytellers to produce content using third-party AI.
The platform will offer two types of AI-animated content: “Flips,” comic book-style stories, and “Spins,” interactive choose-your-own-adventure films. Despite concerns from Hollywood artists, DreamFlare insists it’s creating new opportunities for creators to earn revenue, not replacing jobs.
DreamFlare has raised $1.6 million and claims partnerships with entertainment executives. The platform has a review process to ensure content doesn’t infringe copyrights. Creators can earn money through revenue sharing, ads, tips, and a merchandise marketplace. DreamFlare’s premium membership costs $2.99/month or $24/year, with a limited-time $9.99 annual offer.
Anthropic’s Claude Adds Prompt Playground to Streamline AI App Development
Anthropic has released new features for its language model, Claude, to help developers create more effective applications through prompt engineering. The company has introduced the Claude 3.5 Sonnet tool, which allows developers to generate, test, and evaluate prompts within the Anthropic Console.
Developers can compare the performance of different prompts side-by-side and rate sample answers, helping new users get started with prompt engineering and saving experienced developers time. Anthropic CEO Dario Amodei has emphasized prompt engineering’s importance for enterprise AI adoption, stating that “30 minutes with a prompt engineer can often make an application work when it wasn’t before.”
The new tools are part of Anthropic’s efforts to make Claude more accessible and useful for developers building AI-powered applications. By providing a prompt playground, the company aims to streamline the process of creating effective AI experiences.
Quora’s Poe Launches Previews, Allowing Users to Create and Share Interactive Web Apps
Quora’s subscription-based AI chatbot aggregator, Poe, introduced a new feature called Previews, enabling users to build interactive applications directly within chat conversations.
With Previews, Poe users can create data visualizations, games, and even drum machines by issuing prompts like “Analyze the information in this report and turn it into a digestible and interactive presentation.” These apps can leverage multiple chatbots, such as Meta’s Llama 3 and GPT-4, and can draw on uploaded files, including videos. The created apps can then be shared with others via a link.
Previews is similar to Anthropic’s Artifacts, but with a key difference – it supports HTML output, including CSS and JavaScript functionality, from any chatbot, not just Anthropic’s models. Quora states that Previews works best with chatbots that excel at programming, such as Claude 3.5 Sonnet, GPT-4, and Google’s Gemini 1.5 Pro.
While the feature requires a $20 per month premium Poe subscription, the few publicly available demos showcase the potential for users to create interactive applications directly within chat conversations.
Former Humane Execs Launch AI Fact-Checking Startup Infactory
Two former Humane employees, Brooke Hartley Moy and Ken Kocienda, have founded Infactory, an AI-powered fact-checking search engine. Hartley Moy and Kocienda, who served as Humane’s Strategic Partnerships Lead and Head of Product Engineering, respectively, are taking a more transparent approach compared to Humane’s pre-launch secrecy.
Infactory will leverage AI, including large language models, to create a natural language interface, but will not rely on AI-generated summaries. Instead, the platform will pull information directly from trusted sources, with citations included, to avoid the hallucination issues of current search engines.
Targeting enterprise customers like newsrooms and research facilities, Infactory will focus on data-driven subjects at launch, rather than more subjective topics. The founders emphasize the importance of carefully selecting data partners to ensure accuracy.
