Homechevron_rightArchiveschevron_rightTechnology

Archives

Explore Stories by Date

22 Jul 2026

SkyPilot raises $20M to unify fragmented AI compute across clouds and optimize GPU usage

SkyPilot emerged from stealth with $20 million in seed funding to solve AI compute fragmentation. The UC Berkeley spinout, co-founded by Databricks' Ion Stoica, promises to help companies squeeze 10% more utilization from GPU spending—translating to $10 million in savings for firms spending $100 million annually. Backed by CEOs from Databricks, Vercel, Replit, and Hugging Face, the startup targets a market expected to grow from $14 billion to over $60 billion by 2034.

2 Sources

Bank of America upgrades EricaAssist with generative AI to cut call times, boost service speed

Bank of America has equipped more than 18,000 customer service representatives with enhanced generative AI capabilities through EricaAssist, delivering real-time contextual guidance in under three seconds. The upgrade reduces average call times by nearly one minute while enabling personalized service that combines human judgment with AI guidance.

2 Sources

Meta tests StoryKit, an AI storytelling app that generates bedtime stories for exhausted parents

Meta is piloting StoryKit, an AI bedtime story app that creates personalized children's stories complete with custom characters, settings, and moral lessons. Available in select countries like Mexico, the iPhone app uses AI to generate narratives, illustrations, and narration—but raises questions about privacy concerns and the role of human creativity in parenting.

2 Sources

21 Jul 2026

Substack launches AI detection tool to fight 'Claudefishing' and restore trust in newsletters

Substack has partnered with Pangram to introduce an AI detection tool that lets readers scan posts, notes, and comments to see how much content was written by humans versus AI. The move aims to combat 'Claudefishing'—when readers unknowingly consume AI-generated content—while helping writers who disclose AI usage maintain credibility. Available now on web and iOS, the tool marks a bold bet that human authorship will become a valuable differentiator in an internet increasingly flooded with AI slop.

9 Sources

Weka launches NeuralMesh 6 to cache 100% of tokens and solve GPU memory bottlenecks in AI inference

Weka unveiled NeuralMesh 6, its biggest software overhaul, alongside its first custom-designed hardware line WEKApod. The platform introduces augmented memory grid technology that caches 100% of pre-calculated tokens, preventing GPUs from wasting compute on redundant calculations. Benchmark tests show 10x higher token throughput and 10x more concurrent users compared to standard memory systems.

2 Sources

OpenAI launches ChatGPT for small businesses program to accelerate AI adoption nationwide

OpenAI unveiled a comprehensive ChatGPT for small businesses program designed to transform entrepreneurs into AI power users. The initiative offers virtual training, in-person AI academies across the U.S., and partner integrations with companies like Shopify and Intuit. This comes as ChatGPT Work and Codex reach 10 million users, doubling the user base from earlier in July.

4 Sources

Nvidia unveils Vera CPU specs and benchmarks, positioning itself to challenge AMD and Intel

Nvidia released detailed specifications for its Vera CPU, featuring 88 custom Olympus cores designed specifically for agentic AI workloads. The chip is part of the company's Vera Rubin platform and represents its first custom core design for data centers. Early customers including OpenAI, Microsoft, and Oracle are already receiving systems, marking Nvidia's aggressive push into a CPU market traditionally dominated by AMD and Intel.

10 Sources

Google launches Gemini 3.5 Flash Cyber to find and fix software vulnerabilities faster

Google DeepMind unveiled Gemini 3.5 Flash Cyber, a lightweight AI security model designed to identify and patch security vulnerabilities quickly and affordably. Available initially to governments and trusted partners through CodeMender, the model discovered 55 unique issues in the V8 JavaScript Engine—outperforming larger models while operating at significantly lower cost.

3 Sources

Samsung Health Assistant launches in beta to translate your health data into actionable insights

Samsung has rolled out its AI-powered personal health assistant in beta for eligible US users. The new feature addresses a persistent challenge in digital health: transforming scattered metrics from sleep, activity, and vitals into meaningful guidance. Unlike standalone AI chatbots, Samsung's assistant integrates directly into the Samsung Health app, connecting data from Galaxy Watch and Galaxy Ring devices to deliver physician-validated recommendations.

7 Sources

Tesla robotaxi mileage drops 36% as Elon Musk shifts tone on autonomous ride-hailing rollout

Tesla's robotaxi service drove 700,000 miles in Q2 2026, down from 1.1 million in Q1—a 36% decline despite expanding to six cities across Texas and Florida. CEO Elon Musk admitted the company needs to collect Cybercab-specific data before scaling, marking a shift from years of claims that millions of customer vehicles were training the system. The slowdown raises questions about Tesla's ability to compete with Waymo's four million weekly miles.

7 Sources

Halliday G2 smart glasses ditch cameras for AI-powered meeting intelligence at $599

Halliday unveils its second-generation smart glasses, the G2, priced at $599 with shipments planned for September 2026. The camera-free design targets professional environments with Meeting Flow AI that tracks conversations, identifies action items, and provides real-time translations in 45 languages. After a rocky debut with its first-generation glasses, the company has rebuilt its approach with upgraded binocular waveguide displays and workplace-focused AI features.

7 Sources

Cisco Antares AI models hunt bugs faster than Gemini and GPT, cost under $1 to scan 500 repos

Cisco unveiled Antares, a family of small language models designed for vulnerability detection that outperform Google Gemini and rival OpenAI GPT systems. The open-weight AI models run locally, scan 500 repositories in 15 minutes for less than $1, and keep proprietary code on-premises. Cisco is gating access due to dual-use risks, as tools that find bugs can also help attackers exploit them.

6 Sources

Microsoft funds Mistral's European AI expansion in multibillion-dollar infrastructure deal

Microsoft has committed billions to fund French AI startup Mistral's computing infrastructure across Europe, expanding Azure's presence while addressing sovereignty concerns. The deal avoids new equity stakes, instead focusing on compute capacity and model distribution through Foundry and Copilot Studio as European nations seek alternatives to U.S.-controlled AI infrastructure.

9 Sources

Mobileye partners with Stellantis to bring cloud-powered ADAS technology to future vehicles

Mobileye announced a partnership with Stellantis to integrate its Cloud-Enhanced Advanced Driver Assist technology into select vehicles starting in 2027. The collaboration will leverage Mobileye's REM technology, which already covers over 95% of public roads in the US and Europe, to enable intelligent lane keeping and hands-free driving capabilities in future Stellantis models.

2 Sources

Claude AI automation unlocks hidden features in Code and Cowork to save hours every week

Anthropic's Claude AI extends far beyond basic chatbot functions with specialized tools like Claude Code for software development and Claude Cowork for workflow automation. Users are discovering hidden features that automate email routines, clipboard monitoring, and coding tasks, though privacy concerns remain when connecting AI to personal data.

5 Sources

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

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.

Explore

News Categories

Technology Business Policy Startups Health Science Entertainment

Terms Privacy Content Contact Us