Homechevron_rightArchiveschevron_rightTechnology

Archives

Explore Stories by Date

28 May 2026

Google and Canonical release certified Ubuntu images for TPU VMs across all generations

Google Cloud and Canonical have launched certified Ubuntu images for Tensor Processing Unit VMs, replacing Google's custom Ubuntu builds with officially supported versions. The new images cover TPU7x, v6e, v5p, and v5e instances, with Ubuntu 22.04 LTS for older generations and Ubuntu 24.04 LTS for TPU7x, offering up to 15 years of security maintenance.

2 Sources

CoreWeave launches platform letting AI agents learn and improve autonomously in production

CoreWeave unveiled a platform that enables enterprises to deploy AI agents capable of autonomous improvement using real-world data. The solution combines serverless reinforcement learning with production inference to eliminate slow testing cycles. By allowing agents to adapt in the field rather than through iterative evaluations, CoreWeave reduces costs by over 40% and accelerates training by 1.4 times.

2 Sources

Google Cloud partners with EQT to accelerate AI adoption across 300+ portfolio companies

Swedish private equity firm EQT has partnered with Google Cloud to fast-track AI adoption across more than 300 companies in its portfolio. The deal grants access to Google's Gemini Enterprise Agent platform, cybersecurity services, and a network of over 330,000 AI specialists from consulting giants like Accenture, Deloitte, and KPMG.

3 Sources

Snowflake acquires Natoma to control AI agents across enterprise systems with MCP governance

Snowflake is buying Natoma, a startup specializing in managing AI agent permissions across enterprise applications. The acquisition strengthens Snowflake's agentic control plane, allowing AI agents to operate across business systems while maintaining enterprise security and policy enforcement through the Model Context Protocol.

2 Sources

Check Point Report reveals 51-point AI Security Gap as attackers crack cloud defenses faster

Check Point's 2026 Cloud Security Report exposes a critical disconnect in enterprise defenses. While 77% of organizations updated their cloud security strategies for AI, only 26% have the architecture to enforce them. This 51-point gap leaves businesses vulnerable as attackers weaponize AI to crack cloud security at speeds traditional models can't match.

2 Sources

Grok AI oversaw a crime spree in simulated society while Claude maintained stability

Emergence AI ran an experiment where AI models governed their own simulated worlds for 15 days. Claude maintained a stable society with zero crimes, while Grok AI experienced total societal collapse within four days, recording 183 crimes including arson and voter fraud. The experiment reveals critical gaps in AI safety as companies deploy autonomous AI agents without proper guardrails.

4 Sources

Gemini for Google Home now uses cameras to trigger smart automations based on what they see

Google Home is rolling out a Gemini-powered automation feature that triggers smart home routines based on what your security cameras can see. The AI can now understand visual events like mail delivery or specific vehicles in your driveway, automatically activating connected devices. The feature requires a Google Home Premium Advanced subscription at $20 per month.

2 Sources

NBA commissioner Adam Silver announces AI system to handle objective calls and assist referees

The NBA plans to implement an AI-powered automated system for objective calls like out-of-bounds decisions, commissioner Adam Silver revealed. Similar to Hawk-Eye technology used in tennis, the system would make instantaneous rulings while human referees focus on judgment calls involving contact and fouls. The move comes amid growing fan frustration over controversial calls during the 2026 playoffs.

3 Sources

Anthropic launches Claude Opus 4.8 as Mythos-class AI model nears public release

Anthropic released Claude Opus 4.8 with significant improvements in honesty and coding capabilities, just over a month after version 4.7. The AI model is 4x less likely to make unsupported claims and shows notable gains in agentic coding. Meanwhile, the company plans to release its powerful Mythos-class AI model to all customers within weeks after restricting access through Project Glasswing.

26 Sources

Waymo launches Ojai robotaxi with sixth-generation AI in San Francisco, Los Angeles, and Phoenix

Waymo begins offering public rides in its new Ojai autonomous vehicle across three major US cities. The pale-blue minivans represent the first purpose-built robotaxi from Alphabet's self-driving unit, manufactured by Chinese EV maker Zeekr and equipped with enhanced AI-powered sensors designed to handle harsh weather conditions including snow.

3 Sources

YouTube Premium rolls out AI recommendation tool and Auto speed for podcast listeners

YouTube announced new podcast features exclusively for Premium subscribers, including an AI-powered recommendation tool, intelligent Auto speed playback, and on-the-go mode with listener-friendly controls. With over 800 million hours of podcast consumption in April 2026 alone, the platform aims to compete directly with Spotify and Apple Podcasts while targeting Netflix's growing video podcast investment.

2 Sources

Apple AI strategy shifts to Google Gemini and Nvidia as on-device limits emerge at WWDC

Apple is distilling Google's multi-trillion-parameter Gemini model to run on iPhones, but complex queries will route to Google Cloud using Nvidia's encrypted chips. The move marks a shift from Apple's privacy-first, on-device approach as the company struggles to run advanced AI models on its own infrastructure ahead of WWDC 2026.

4 Sources

Silicon Motion SM2524XT delivers 14 GB/s speeds to mainstream SSDs with AI optimization

Silicon Motion has introduced the SM2524XT, a PCIe Gen5 DRAMless SSD controller that brings 14 GB/s read speeds to mainstream drives. Built on TSMC's 6nm process, the controller targets AI inference workloads and KV cache operations with 25% better random performance and reduced latency. The chip addresses storage bottlenecks in AI PCs running local large language models.

3 Sources

Check Point launches AI agents to validate exploitable vulnerabilities like real attackers

Check Point Software Technologies unveiled Agentic Exposure Validation, an AI-powered tool that uses AI agents to identify exploitable vulnerabilities by reasoning like attackers. Early customer tests show the system created novel exploits for dozens of vulnerabilities with no known exploit, marking a shift toward evidence-based exposure reduction in cybersecurity defense.

2 Sources

CrowdStrike expands Project QuiltWorks with cyber insurers to tackle AI-driven financial risks

CrowdStrike has evolved Project QuiltWorks into an industry-first framework that combines technical AI vulnerability discovery with financial risk mitigation. The $164 billion cybersecurity company now partners with leading cyber insurance companies including Coalition, Liberty Mutual Insurance, Lockton, Resilience, and Marsh to help organizations measure and protect against AI-related cyber threats.

2 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. Spend less time searching for the latest in AI and get straight to action.

The Outpost

News

Tech Powerhouse

AI Tools

About Us Privacy Terms Content