Reka AI Raises $58 Million to Customize LLMs for Enterprise

Based on the premise that it is impractical to deploy an all-purpose LLM for specific use cases, a group of researchers from Google, Baidu, DeepMind and Meta founded Reka AI in July 2022. A year later the company has emerged from stealth mode with news of $58 million in Series A funding led by DST Global and Radical Ventures. Strategic partner Snowflake Ventures also participated, along with angel investor Nat Friedman, former CEO of GitHub. The Sunnyvale, California-based startup says it is building “enterprise-grade state-of-the-art AI assistants for everyone, regardless of language and culture.” Continue reading Reka AI Raises $58 Million to Customize LLMs for Enterprise

Meta’s MusicGen AI Works with Language and Song Prompts

Meta Platforms has debuted what’s being called “ChatGPT for audio.” MusicGen is an AI music generator that can create tunes from natural language or song snippets. The company says MusicGen was trained on 20,000 hours of music, including 10,000 hours of “high-quality” licensed songs and 390,000 instrumental tracks. Meta released MusicGen on GitHub this past weekend, and is currently demoing the app on Facebook’s Hugging Face page. Visitors can generate tunes by describing the sound they want. Among Meta’s prompts: “80s driving pop song with heavy drums and synth pads in the background.” Continue reading Meta’s MusicGen AI Works with Language and Song Prompts

Senators Question Meta Platforms About Recent LLaMA Leak

Meta Platforms CEO Mark Zuckerberg received a letter this week from Senators Richard Blumenthal and Josh Hawley of the Subcommittee on Privacy, Technology & the Law that took the executive to task for an online leak of the company’s LLaMA artificial intelligence system. The 65-billion parameter language model, which is still under development, was open-sourced in February. Available on request through Meta’s GitHub portal, it wound up on 4chan and BitTorrent “making it available to anyone, anywhere in the world, without monitoring or oversight,” the senators wrote. Continue reading Senators Question Meta Platforms About Recent LLaMA Leak

Twitter Community Notes Aim to Curb Impact of Fake Images

Twitter is emphasizing crowdsourced moderation. The launch of Community Notes for images in posts seeks to address instances where morphed or AI-generated images are posted. The idea is to expose altered content before it goes viral, as did the image of Pope Francis wearing a Balenciaga puffy coat in March and the fake image of an explosion at the Pentagon in May. Twitter says Community Notes about an image will appear with “recent and future” posts containing the graphic in question. Currently in the test phase, the feature works with tweets featuring a single image. Continue reading Twitter Community Notes Aim to Curb Impact of Fake Images

Windows Gets AI Boost and Bing Is ChatGPT’s Search Engine

The focus was on artificial intelligence at Microsoft Build 2023, where the software giant showcased AI enhancements across Windows 11, Microsoft 365, Edge and more. The company also said Bing is now the default search engine for OpenAI’s ChatGPT chatbot. A Copilot AI button is being integrated into the Windows 11 taskbar, making it the first personal computer operating system to integrate AI assistance to help users seamlessly connect across apps and improve productivity. “For example, you cannot only copy and paste, but also ask Windows Copilot to rewrite, summarize or explain your content,” Microsoft points out. Continue reading Windows Gets AI Boost and Bing Is ChatGPT’s Search Engine

Stability AI Debuts Open Source StableLM Foundation Model

Stability AI has released StableLM, an open source language model that will compete with OpenAI’s GPT-4 to create apps like ChatGPT. The Alpha version of StableLM is available in 3 billion and 7 billion parameters, and the company promises 15 billion to 65 billion parameter models to come. “With the launch of the StableLM suite of models, Stability AI is continuing to make foundational AI technology accessible to all,” the London-based company said. The StableLM models can generate text and code to power various downstream applications with appropriate training. Continue reading Stability AI Debuts Open Source StableLM Foundation Model

Auto-GPT Generates Social Sizzle, Ushers in Era of AI Agents

Auto-GPT, an open source app that uses OpenAI’s text-generating models, is currently generating a great deal of social media attention. The program can act somewhat autonomously in that it creates its own feedback loop, asking itself a series of questions to help build a more nuanced and complete response to a text prompt. In short, something that would take a user multiple prompts to produce the desired information using ChatGPT could be accomplished using a single request of Auto-GPT, which could independently explore a subject before spitting back a comprehensive response. Continue reading Auto-GPT Generates Social Sizzle, Ushers in Era of AI Agents

Google Cloud Partners with Replit to Develop AI Coding Tools

Google has teamed with San Francisco startup Replit in a bid to challenge Microsoft’s GitHub and the Github Copilot code generator launched in conjunction with OpenAI. Under the new partnership, Replit developers will get access to Google Cloud infrastructure, services, and foundation models via Replit’s software development AI, called Ghostwriter, while Google Cloud and Workspace developers will get access to Replit’s collaborative code editing platform. Replit, which says 20 million developers use its platform, launched Ghostwriter in the fall and in January added a conversational AI interface for generating code and debugging. Continue reading Google Cloud Partners with Replit to Develop AI Coding Tools

Amazon Opens Low-Power Sidewalk Network to Developers

Amazon has opened Sidewalk, its low-power, wide-area network (LPWAN), to developers. Introduced in 2019, Amazon’s early focus was using Sidewalk to optimize Ring cameras and Level smart locks. The company quietly extended coverage to what it says is 90 percent of the U.S. population, and indicates it is now ready to connect “the next billion devices.” Sidewalk can be useful for a multitude of IoT devices that need Internet connectivity yet can’t always rely on a robust four-bar signal, or just don’t require a lot of bandwidth or power. Continue reading Amazon Opens Low-Power Sidewalk Network to Developers

Epic Debuts Unreal 5.2 and Expands ‘Fortnite’ Profit-Sharing

Epic Games introduced Unreal Engine 5.2 at GDC 2023, demonstrating new levels of photorealism and physics designed to facilitate the creation of real-time interactive worlds for the metaverse. Procedural content generation tools and a material framework called Substrate for intricate surfaces with detailed refraction and reflection are new to the 5.2 preview build, available now at the Unreal Engine Marketplace and on GitHub. The economy around Unreal Editor for “Fortnite” has expanded to include a new revenue-sharing plan that lets creators keep 40 percent of funds generated from “islands” they create for the popular game. Continue reading Epic Debuts Unreal 5.2 and Expands ‘Fortnite’ Profit-Sharing

GitHub Is Testing New Security Tools for Open-Source Code

Cloud-based code hosting service GitHub wants to make open-source material more secure. The Microsoft service is expanding safety features with two new offerings in beta. Secret scanning alerts are now free for all public repositories while push-notifications for custom secret patterns are also being made available. Open-source code is now incorporated into a whopping 97 percent of applications, according to Synopsys, which says 90 percent of organizations rely on it to varying degrees. Yet the very access that contributes to its popularity also leaves it vulnerable to malicious actors, as emphasized by the SolarWinds, Log4j and other breaches. Continue reading GitHub Is Testing New Security Tools for Open-Source Code

Deepmind’s AlphaCode AI Can Program Like Human Coders

DeepMind researchers have trained an AI to solve computer coding challenges as well as the average person. In a paper published last week in the journal Science, the group from Google’s AI division described how AlphaCode performed when pitted against human programmers, ranking in the top 54.3 percent in simulated tests, commensurate with “approximately human-level performance.” “This performance in competitions approximately corresponds to a novice programmer with a few months to a year of training,” according to Science, which says about half the humans who compete in coding contests could outperform the AI. Continue reading Deepmind’s AlphaCode AI Can Program Like Human Coders

CodeCatalyst Automatically Sets Up Developer Environments

Amazon has announced a preview release of CodeCatalyst, a unified software development and delivery service that the company says enables software teams to plan, build and deliver applications on AWS with reduced friction throughout the development lifecycle. At AWS re:Invent 2022 in Las Vegas, Amazon VP and CTO Werner Vogels detailed how CodeCatalyst offers a “single place” where developers can collaborate to create an app on Amazon Web Services. CodeCatalyst is designed to make it simple to marshal resources and toggle between different application development environments, Vogels explained from the stage at his Thursday keynote. Continue reading CodeCatalyst Automatically Sets Up Developer Environments

Microsoft Integrates DALL-E 2 into Designer and Creator Apps

Microsoft announced it is integrating OpenAI’s DALL-E 2 into its new Microsoft Designer app, as well as its Microsoft Edge browser and the Image Creator tool in its Bing search engine. Microsoft provides cloud computing services to OpenAI and has partnered with OpenAI in AI commercialization efforts including the Azure OpenAI Service, now in preview, and GitHub Copilot. The Designer web app can be used to create designs for posters, presentations, invitations and other graphics that can be printed and used for display or shared on social or business media. Continue reading Microsoft Integrates DALL-E 2 into Designer and Creator Apps

Nvidia Debuts New AI Model That Quickly Generates Objects

Nvidia Research is introducing a new AI model that largely automates the process of creating virtual worlds, making it easier for developers to populate games and VR experiences with a diverse array of 3D buildings, vehicles, characters and more. Trained using only 2D images, GET3D generates 3D shapes with high-fidelity textures and complex geometric details. GET3D can generate “a virtually unlimited number of 3D shapes based on the data it’s trained on,” according to Nvidia, which says the objects can be used in 3D representations of buildings or the great outdoors, in games or the metaverse. Continue reading Nvidia Debuts New AI Model That Quickly Generates Objects