Get hand-picked daily summaries of the best, most informative AI articles from around the web. Browse recent summaries, then jump to the original source when you want the full article.
OpenAI has launched o3-pro, its most advanced AI model yet, replacing o1-pro for ChatGPT Pro and Team users. This "reasoning model" excels in technical fields and outperforms competitors in math and science benchmarks. It offers web search, code execution, and visual reasoning, but lacks image generation and some temporary features. The model is available via API for developers at $20-80 per million tokens.
Nvidia CEO Jensen Huang dismissed claims that AI will eliminate half of entry-level white-collar jobs in five years. He believes AI will transform all jobs but create more opportunities through increased productivity. Huang...
Amazon Ads launched its Video Generation platform for all US customers, offering AI-generated video ads in minutes. The tool creates six different videos from product images, adding motion and context. It aims to help small...
Sam Altman revealed that a typical ChatGPT query uses about 0.000085 gallons of water and 0.34Wh of energy, equivalent to one-fifteenth of a teaspoon of water and an oven's energy use for just over a second. As AI's energy...
Meta's V-JEPA 2 AI model, trained on vast video data, helps robots predict action outcomes and make logical connections, reportedly outperforming Nvidia's Cosmos model in speed and potentially transforming robotics and physical...
Docker Desktop 4.42 adds native IPv6 networking, built-in Model Control Plane, improved model packaging, and Docker Model Runner updates, enhancing AI capabilities and streamlining development workflows for local AI model...
This archive is built from SMRTR newsletter stories. Subscribe for hand-picked stories without the extra noise.
Pinterest's new AI "auto-collages" feature creates shoppable, grouped product collages from advertisers' catalogs, aiming to increase user engagement and simplify ad creation, with early tests showing promising results for Gen...
Starbucks unveils "Green Dot Assist," an AI assistant for baristas and managers. The tool, built on OpenAI and Microsoft Azure, will help with drink recipes, equipment troubleshooting, and scheduling. A trial at 35 locations...
OpenAI CEO Sam Altman announced a delay in releasing the company's first open model in years, now set for later this summer. The model aims to match OpenAI's proprietary reasoning capabilities and outperform competitors like...
OpenAI released o3-pro, an upgraded simulated reasoning model, to ChatGPT Pro and Team users. It offers improved capabilities in math, science, and coding, along with significant price reductions. However, new studies question...
Verse AI's novel machine learning approach, Axiom, demonstrates impressive efficiency in mastering simple video games. Inspired by human brain function and the free energy principle, it uses prior knowledge and active inference...
OpenAI and Google, former rivals in the AI race, have formed a strategic partnership. OpenAI will now use Google's cloud computing services to support AI model training, diversifying beyond its primary backer Microsoft. This...
Meta is expanding its AI capabilities with new video editing features in Meta AI and its Edits app. Users can now apply preset AI prompts to 10-second videos, changing costumes, locations, and styles. The company is rolling out...
Artistly is an AI platform for generating and editing high-quality images from text prompts. It offers unlimited image creation, consistent character generation, and editing tools for a one-time fee. The platform is user-friendly...
AI adoption brings benefits but also risks, requiring businesses to prioritize ethical use, transparency, and data privacy to maintain trust, comply with regulations, and avoid penalties while delivering high-quality products.
Wikipedia briefly tested AI-generated summaries on mobile articles but stopped due to editor objections, citing concerns about trust, reliability, and conflicts with the platform's human-centric content creation and verification...
A humble Atari 2600, with its 1.19 MHz processor, has trounced the mighty ChatGPT in a game of chess. Citrix engineer Robert Jr. Caruso pitted OpenAI's chatbot against an emulated version of "Video Chess," the only chess title...
A new AI-powered travel planning framework combines large language models with mathematical solvers to create realistic, logical itineraries. Developed by MIT and IBM researchers, the system achieves over 90% success rates in...
AI can boost education in developing nations through translation tools, culturally relevant content, and mobile learning centers, but must overcome language barriers and cultural challenges to be effective for underserved...
AI adoption faces challenges despite its potential to streamline operations and enhance decision-making. Only 26% of companies have achieved tangible results beyond proofs of concept. To bridge the adoption gap, businesses need...
1X Technologies' Redwood AI model enables NEO robots to autonomously perform household tasks, using whole-body manipulation and onboard processing to generalize chores, interact with the environment, and respond to voice commands.
Apple has updated its AI models powering various features across its devices and platforms. However, the company's own benchmarks reveal that these models underperform compared to older models from competitors like OpenAI,...
Mistral AI's Magistral, their first reasoning model, is available in open-source and enterprise versions. It excels in multi-step logic across languages, performing well on complex math problems and offering faster reasoning...
A new study reveals people's attitudes towards AI are nuanced, depending on the context. Researchers found AI is preferred when perceived as more capable than humans and personalization is unnecessary. However, people favor human...