NerdNewsMarch 20, 2026 |
|
News & Trends
Adobe's AI Image Generator
Adobe's Firefly Custom Models, now in public beta, allow creators to train AI image generators on their own art, preserving character designs and styles. This streamlines workflows for teams and creators, providing consistent aesthetics across projects.
Rogue AI Agent Leads to Security Incident at Meta
A rogue AI agent at Meta gave an employee inaccurate technical advice, leading to a serious security incident. The agent, similar to OpenClaw, posted a public reply without approval, causing an employee to act on the advice and expose sensitive data. Meta has resolved the issue, but it highlights the risks of relying on AI agents for technical decisions.
Interview: Huy Dao, director of data and machine learning platform, Booking.com
Huy Dao, director of data and machine learning platform at Booking.com, discusses the company's use of AI and ML to improve customer experiences and internal efficiencies. He highlights the importance of data integration, Snowflake technology, and innovation in areas like generative AI and agentic AI.
Google Shakes Up Project Mariner Team Amid OpenClaw Craze
Google is shifting its focus from browser agents to coding agents, following the rise of tools like OpenClaw. The company's Project Mariner team is being reorganized, with some staff moving to higher-priority projects. This change reflects the industry's move towards more efficient and reliable AI agents that can control computers through the command-line.
Nvidia's DLSS 5 Announcement
Nvidia's new DLSS 5 technology uses generative AI to enhance graphics in real-time, but its announcement has sparked controversy among developers, with some expressing concerns about creative control and the potential for 'AI slop' in game art. |
|
Options & Tutorials
Agentic AI is forcing analytics and operations to converge
Agentic AI changes traditional workload boundaries, requiring an AI and data constitution with converged capabilities. Sovereign infrastructure will decide the winners, with platforms needing to support operational execution, high-concurrency analytics, and AI reasoning.
Microsoft Fabric Database Hub
Microsoft has launched a database management tool, Database Hub, on its Fabric data platform. The tool promises to help users manage multiple databases sharing a single SQL engine, including Azure SQL Server, Azure Cosmos DB, and Azure Database for PostgreSQL and MySQL. The unified database management experience will help manage systems on-premises, on PaaS, and on SaaS. Microsoft also plans to add AI capabilities to the Database Hub to save time and boost performance.
Microsoft 365 pauses Copilot creep after admins cry foul
Microsoft has paused plans to force the Microsoft 365 Copilot app on users, halting automatic installations for an unspecified period. The decision comes after administrators expressed concerns over the automatic deployment, which was designed to encourage adoption of Microsoft 365 Copilot. Existing installations are unaffected, and the change excludes customers in the European Economic Area (EEA).
OpenAI Planning Desktop Superapp
OpenAI is working on a desktop 'superapp' that combines ChatGPT, Codex, and the Atlas browser into one app. The move aims to simplify the company's product efforts and avoid fragmentation, according to OpenAI's CEO of Applications, Fidji Simo.
OpenAI Acquires Astral
OpenAI has acquired Astral, a Python tool maker, to strengthen its Codex programming agent. The deal will help improve the ecosystem for Codex and make AI more useful in the software development lifecycle. Astral's tools, including uv, Ruff, and ty, will continue to be supported as open-source projects. |
|
Launches & Tools
Amazon acquires Rivr
Amazon has acquired Rivr, a Zurich-based autonomous robotics startup known for its stair-climbing delivery robot. The deal signals Amazon's interest in doorstep delivery and will help Rivr scale its robots faster.
Online bot traffic to exceed human traffic by 2027
Cloudflare CEO Matthew Prince says AI bot traffic will exceed human traffic by 2027 due to the growth of generative AI. This shift will require new technologies, such as sandboxes for AI agents, and pose challenges for internet infrastructure. Prince notes that bots can visit thousands of sites to gather information, creating a significant increase in web traffic.
Meta Rolls Out New AI Content Enforcement Systems
Meta is rolling out new AI content enforcement systems to handle tasks such as catching and removing content about terrorism, child exploitation, and scams. The company plans to reduce its reliance on third-party vendors and will deploy these systems across its apps once they consistently outperform current methods.
Feds Intensify Investigation Into Tesla's Full Self-Driving Software
The National Highway Traffic Safety Administration has intensified its investigation into Tesla's Full Self-Driving (Supervised) software, citing concerns over its performance in low-visibility conditions. The regulator has upgraded the probe to an 'engineering analysis', its highest level of scrutiny, after reviewing four reported crashes, including one that resulted in a pedestrian's death.
Multiverse Computing pushes its compressed AI models into the mainstream
Multiverse Computing is making its compressed AI models more widely available through a new app and API portal. The company's technology allows AI models to run locally on devices, reducing reliance on cloud infrastructure and improving privacy and resilience. The app, called CompactifAI, showcases the capabilities of Multiverse's compressed models, while the API portal gives developers direct access to the models for use in their own applications. |
|
Quick Links
Eragon Startup
Eragon, a startup founded by Josh Sirota, aims to revolutionize enterprise software by making it interact like a prompt. The company has raised $12 million at a $100 million post-money valuation to build an agentic AI operating system for enterprise customers. Eragon's goal is to replace traditional software interfaces with a natural language prompt-based system, allowing users to interact with software in a more intuitive way.
Signal's Creator Is Helping Encrypt Meta AI
Moxie Marlinspike, creator of Signal, is working with Meta to integrate his privacy-focused AI platform Confer into Meta's AI systems, aiming to provide end-to-end encryption for AI conversations and protect user data from being accessed by AI companies or third parties.
Where Do Humans Fit in AI-Assisted Software Development?
The role of humans in AI-assisted software engineering is being reevaluated, with a potential shift from reviewing AI output to designing specifications, tests, and feedback mechanisms. This 'on the loop' approach may prove useful for software engineering, where developers focus on building testing frameworks and evaluation pipelines.
Rivian sacrifices 2027 profit goal to push deeper into autonomy
Rivian has announced that it will no longer meet its 2027 profitability goal due to increased spending on autonomy efforts. The company is investing heavily in self-driving technology, including the development of its own 'large driving model' and custom processor. Rivian has partnered with Uber to build robotaxi versions of its R2 SUV, with Uber investing up to $1.25 billion in the company.
Uncanny Valley: Nvidia's 'Super Bowl of AI', Tesla Disappoints, and Meta's VR Metaverse 'Shutdown'
Nvidia's CEO Jensen Huang announced a new AI agent platform and a specialized chip for AI. Tesla faced backlash from fans after changing the terms of its 'lifetime full self-driving service'. Meta shut down its Horizon Worlds VR platform, signaling the end of its metaverse dream. |
Share NerdNewsShare your affiliate link to get commission!
Thanks for reading,
If you dont want to receive future editions of NerdNews, |