Homechevron_rightArchiveschevron_rightTechnology

Archives

Explore Stories by Date

Today

GitHub to train AI models on Copilot user data unless you opt out by April 24

Microsoft's GitHub will begin using customer interaction data from Copilot Free, Pro, and Pro+ users to train AI models starting April 24. The policy change affects code snippets, inputs, and outputs from private repositories. Users can opt out through GitHub account settings, though the company encourages participation to improve code suggestions and bug detection.

2 Sources

Google's TurboQuant AI breakthrough sends memory stocks tumbling on reduced storage demand

Memory stocks fell sharply Wednesday after Google unveiled TurboQuant, a new compression algorithm that could slash data storage needs for AI systems by at least 6x. SanDisk dropped 5.7%, Micron Technology fell 3%, and Western Digital declined 4.7% as investors reassessed the outlook for decreased demand for memory chips in the AI sector.

2 Sources

Google's TurboQuant slashes AI memory usage by 6x, sends memory chip stocks tumbling

Google Research unveiled TurboQuant, a compression algorithm that reduces large language model memory footprint by at least 6x while delivering an 8x performance boost on Nvidia H100 GPUs. The breakthrough targets the key-value cache bottleneck without sacrificing accuracy, but memory chip stocks including Micron and Western Digital dropped as investors reconsidered future demand for AI hardware.

7 Sources

Normal Computing raises $50M to solve the AI energy crisis with thermodynamic chip technology

Normal Computing secured $50 million in Series B funding led by Samsung Catalyst to address the escalating cost and power consumption of AI chips. The startup is developing thermodynamic computing technology and an EDA platform already used by half of the world's top semiconductor companies. CEO Faris Sbahi warns data centers will hit an energy wall by 2030.

2 Sources

Yesterday

Spotify lets artists manually approve releases to combat AI fakes flooding music platforms

Spotify is beta-testing Artist Profile Protection, a new feature that lets musicians review and approve releases before they appear on their profiles. The tool addresses the surge of AI-generated songs and fake tracks that have been wrongly attributed to real artists, affecting everyone from Drake and Beyoncé to indie acts like King Gizzard & The Lizard Wizard.

2 Sources

Domo unveils AI agent builder to connect enterprise data with external AI platforms

Business intelligence firm Domo introduced an AI agent builder and orchestration framework at its Domopalooza conference in Salt Lake City. The new tools use Model Context Protocol to connect enterprise data directly to external AI platforms like OpenAI ChatGPT, Google Gemini and Anthropic Claude, enabling organizations to deploy custom AI agents that automate tasks across business workflows.

2 Sources

Google launches Lyria 3 Pro to generate three-minute songs with enhanced creative control

Google unveiled Lyria 3 Pro, a music generation model that creates tracks up to three minutes long, a significant leap from the 30-second limit of its predecessor. The model offers enhanced creative control over song structure and integrates across Google's ecosystem, including Gemini, Vertex AI, and ProducerAI, while embedding SynthID watermarks to identify AI-generated content.

6 Sources

Reddit will require fishy accounts to verify they are human in new bot crackdown

Reddit CEO Steve Huffman announced that accounts exhibiting automated or fishy behavior must verify humanness using methods like passkeys, Face ID, or World ID. The move aims to combat bots flooding the platform as AI bot traffic threatens to overtake human activity online. Reddit already removes 100,000 bot accounts daily.

4 Sources

Apple gets full Gemini access to build on-device AI models for smarter Siri

Apple's partnership with Google runs deeper than expected. The company has complete access to Google's Gemini model in its own data centers, allowing it to create smaller, distilled AI models that run locally on devices. This process enables Apple to build efficient models requiring less computing power while maintaining Gemini-like performance for Siri improvements.

2 Sources

Claude AI suffers another widespread outage affecting thousands of users globally

Anthropic's Claude AI chatbot experienced another significant outage on March 25, with Downdetector reports spiking from 150 to over 5,318 within minutes. Users across multiple regions encountered error messages preventing them from sending prompts or receiving responses. Anthropic confirmed the service disruptions on its status page, noting elevated errors under investigation, but provided no timeline for restoration.

2 Sources

Meta smart glasses blocked in Europe as battery rules and AI regulations stall launch

Meta's new Ray-Ban Display glasses won't reach European markets anytime soon. The European launch delay stems from EU regulations requiring removable batteries by 2027 and restrictions on AI features. While Meta cites supply shortages, the company is seeking exemptions for AI wearables as regulatory challenges mount in both Europe and the U.S.

2 Sources

Intel Arc Pro B70 and B65 GPUs Target AI Workloads with 32GB Memory at Half NVIDIA's Price

Intel unveiled its Arc Pro B70 and Arc Pro B65 workstation graphics cards featuring 32GB of memory for AI inference workloads and professional applications. Priced at $949, the Arc Pro B70 undercuts NVIDIA's RTX Pro 4000 by nearly half while offering more memory capacity. The cards mark Intel's first deployment of its larger Battlemage GPU architecture, though aimed at AI and pro markets rather than gaming.

5 Sources

Intel launches Panther Lake vPro platform with AI capabilities to reclaim commercial PC market

Intel officially released its Core Ultra Series 3 Panther Lake processors with vPro certification, targeting the commercial PC market where AMD has gained ground. The chipmaker also launched Xeon 600 workstation CPUs and introduced new vPro features including a certification program for apps and accessories, promising improved power efficiency and AI-driven security capabilities.

2 Sources

Samsung unveils Exynos 1680 SoC with AI boost, 200-megapixel camera and 144Hz display support

Samsung has introduced the Exynos 1680 chipset, the successor to last year's Exynos 1580, featuring a 4nm process and significant upgrades across CPU, GPU, and AI capabilities. The new processor delivers up to 16 percent GPU performance improvement and supports 200-megapixel cameras, with the chipset expected to power the upcoming Galaxy A57 5G smartphone.

2 Sources

SLB and Nvidia expand partnership to build AI infrastructure for energy industry

Oilfield services giant SLB is deepening its collaboration with Nvidia to develop artificial intelligence infrastructure and models for the energy sector. The expanded partnership will create an AI Factory for Energy platform and modular data centers, helping oil and gas producers process massive volumes of operational data while cutting costs and reducing emissions.

3 Sources

Get all updates into your inbox 💌, Stay Up-to-date for Free 🥳

Your Daily Dose of Curated AI News

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Spend less time searching for the latest in AI and get straight to action.

The Outpost

News

About