AI News Roundup — April 6, 2026
Anthropic scrambles after Claude Code usage limits drain faster than expected, Perplexity faces privacy accusations for sharing chat data with ad platforms, and Goldman Sachs forecasts a sharp AI-driven semiconductor revenue surge.
Anthropic: Claude Code Usage Limits Draining Way Faster Than Expected
Anthropic acknowledged that Claude Code users are hitting usage limits far faster than intended, breaking automated workflows and frustrating developers. Quota drain is now a stated top priority for the engineering team, with Pro and Max plan users seeing session limits consumed much faster than before, especially during peak hours.
Perplexity Accused of Secretly Sending Chat Data to Google and Meta Ad Platforms
Perplexity AI faces accusations of violating user privacy by secretly sending chat data to Google and Meta's advertising platforms without user consent. The company faces potential fines of $5,000 per violation, adding to a growing wave of AI privacy enforcement actions.
AI Making Crypto Hacks Cheaper and Easier, Ledger CTO Warns
Ledger CTO Charles Guillemet warned that AI is drastically lowering the cost and effort of attacks on cryptocurrency platforms. Tasks that once required skilled researchers months, like reverse engineering software or chaining exploits, can now be done in seconds with the right prompts, forcing a rethink of crypto security architecture.
Goldman Sachs: AI-Led Demand to Drive Sharp Semiconductor Revenue Surge
Goldman Sachs reports that AI adoption is fueling strong semiconductor investment momentum, with revenue growth continuing sharply through 2026. Studies suggest a 23% average productivity uplift from AI, while company-level data points to even larger efficiency gains of around 33%.
AI Virtual Try-On Startups Target $850 Billion Return Problem in Retail
A wave of AI startups is deploying virtual try-on technology to tackle retail's return crisis. The US National Retail Federation estimated that 15.8% of all retail sales were returned in 2025, totaling $849.9 billion, with online sales reaching 19.3% returns. Generative AI has finally made the technology good enough to meaningfully improve retailer margins.
Bloomberg Explains Vibe Coding, the AI Trend Fueling a New Kind of FOMO
Bloomberg published a deep dive on vibe coding, the practice of building software by describing what you want to AI rather than writing code yourself. The trend is fueling anxiety among both developers worried about job security and non-developers eager to build their own tools for the first time.
AI-Assisted Cyberattacks Surge Across Crypto, Enterprise, and Open Source
A weekend security roundup highlights a five-fold rise in AI-assisted cyberattacks across sectors. Incidents include AI-assisted malware targeting crypto wallets, enterprise API abuse, and continued fallout from the Claude Code npm supply chain incident that affected users during a narrow window on March 31.