All Updates
Page 855 of 1663
April 23, 2026
Claude Code Cuts Dev Time 70%, Releases 24-to-5 Days
Anthropic's Claude Code is gaining traction over OpenAI among developers. Companies like Rakuten and Money Forward report 70% dev time reduction and new feature releases shortened from 24 to 5 days. The article details their usage cases.
Anthropic Fixes Claude Degradation Mystery
Anthropic revealed three harness changes caused Claude's recent degradation: lowered default reasoning effort, verbosity prompt tweaks, and a caching bug. They reverted the reasoning and verbosity changes and fixed the bug in v2.1.116. Benchmarks showed Claude Opus 4.6 accuracy dropping from 83.3% to 68.3%.
US Accuses China of Industrial-Scale AI Theft
The US has accused China of 'industrial-scale' theft of AI technology. China rejects the claims as 'slander.' Potential huge sanctions could disrupt the Trump-Xi summit.
Carbon Nanotubes Challenge Copper Wiring
New advances bring carbon nanotube wiring closer to rivaling copper in electronics. The material shows promise but degrades over time. This could lead to superior alternatives for high-performance chips.
Anthropic Mythos Sparks AI Alarm
Anthropic's Mythos AI model is sparking fears that AI is advancing faster than safe deployment capabilities. The controversy may boost Anthropic's valuation toward $800 billion.
SoftBank Ramps Up AI Battery Production
SoftBank's mobile unit will convert its Osaka factory into a major production line for large-scale batteries. The move aims to power its own AI data centers. This positions SoftBank deeper into AI infrastructure.
Sierra Acquires YC-Backed Fragment
Bret Taylorβs Sierra, an AI customer service agent startup, announced the acquisition of Fragment, a YC-backed French AI startup. This move integrates Fragment's technology into Sierra's platform.
Korean Chip Bonuses Hit $900K Amid Boom
South Korea's chip sector offers bonuses potentially reaching $900,000 due to bullish forecasts. This boom exacerbates concerns over economic inequality. The K-shaped recovery highlights widening gaps.
Guzman y Gomez Plans Agentic Kitchens
Guzman y Gomez is developing agentic kitchens to dynamically adjust to foot traffic variations. The plan aims to optimize restaurant operations using autonomous AI agents.
Claude Opus 4.7 Overzealous Safeguards Frustrate Devs
Anthropic released Claude Opus 4.7 last week with enhanced safeguards to prevent misuse. However, these measures are now blocking legitimate developer queries, causing rising refusal rates from the Acceptable Use Classifier. Customers complain of paying for a model that often refuses to respond.
First Ransomware Confirmed Quantum-Safe
A ransomware family is the first confirmed to use post-quantum cryptography (PQC). Despite no practical benefits yet, attackers adopt it preemptively. This signals rising quantum threats.
Claude Code Fuels Mass Credential Harvest
An attacker integrated Claude-generated code into a credential harvesting operation. The compromised Bissa scanner impacted over 900 organizations.
Bezos AI Lab Hits $38B Valuation
Jeff Bezos-led Project Prometheus, a physical AI lab with Google veteran Vik Bajaj, closed a $10 billion funding round. The deal values the company at roughly $38 billion.
Zscaler Demand Surges on AI Risks
Zscaler CEO Jay Chaudhry states unprecedented demand for services amid cybersecurity risks from AI like Anthropic's Mythos. Enterprises struggle to protect against these threats.
SAP Cloud Beats Estimates via AI
SAP SE's cloud services revenue exceeded analyst estimates following AI agent integration. Europe's largest software firm is pushing AI enhancements aggressively.
Intel AI Forecast Beats Estimates
Intel issued a strong sales forecast for the current period. This reflects benefits from AI infrastructure build-out. The chipmaker is showing signs of recovery.
GPT-5.5 Announced vs Claude Mythos Preview
OpenAI has announced GPT-5.5. The article examines differences between it and Anthropic's Claude Mythos Preview. Published by ITmedia AI+.
Meta Cuts 8K Jobs for AI Funding
Meta is laying off about 8,000 employees (10% of workforce) starting May 20, canceling 6,000 open roles, with more cuts planned for H2 2026. The structural changes reorganize teams into AI-focused pods amid $115-135B AI infrastructure spend. Survivors must train their AI replacements.
LLM Agents Win Kaggle Competition
In March 2026, three LLM agents generated over 600,000 lines of code and ran 850 experiments to secure first place in a Kaggle playground competition. Combined with GPU acceleration, they dramatically compressed the idea generation, testing, and iteration loop in modern machine learning competitions.
AI Agent Luna Overorders Candles in SF Store
Andon Market in San Francisco's Cow Hollow is run entirely by AI agent Luna, acting as CEO. Luna selects products like Bananagrams and books, sets prices, but famously overordered candles. This experiment highlights AI's role in retail management.