Cursor engineers have slashed unexpected tool call errors by 10x in a single sprint by combining AI-driven evaluations, A/B testing, anomaly detection, and custom model tuning to sharpen their agent harness and prepare for a multi-agent future.
Google DeepMind launches ProEval, a free open-source tool that slashes generative AI evaluation costs by up to 100x using Bayesian Quadrature techniques, achieving ±1% accuracy with a fraction of typical samples while proactively identifying model failure patterns across major benchmarks.
Perplexity launches a massive enterprise expansion of its AI agent platform at Ask NYC, introducing 50+ prebuilt workflows, Databricks and Snowflake integrations, financial data connections via Morningstar and PitchBook, and a 1Password integration — all while claiming $2.8 billion in labor-equivalent work completed since launch.
AI startup Rogo Technologies hits a $2 billion valuation after two former Wall Street bankers build a platform that automates the tedious grunt work of investment banking, including spreadsheet crunching and slide deck formatting.
GPT-5-powered agentic search systems are crushing baseline scores with an NDCG of 0.453 versus a 0.289 BM25 baseline on Amazon ESCI, but hit a hard wall when facing knowledge-gap tasks where LLMs cannot evaluate information they don't already know.
Anthropic's Batch API slashes token costs by 50%, but a new analysis reveals the savings only make sense at fleet scale — single-agent use suffers from up to 24-hour latency, and surprisingly, cheaper Haiku models take longer to batch than pricier Sonnet or Opus.
Anthropic launches built-in memory for Claude AI agents in public beta, enabling cross-session learning that has already slashed first-pass errors by 97% and reduced latency by 34% for early adopters like Rakuten, while Wisedocs reports 30% faster document verification.
Microsoft's Copilot agentic AI is now live in Word, Excel, and PowerPoint, delivering multi-step, app-native actions that are already driving massive user engagement gains, with Excel alone seeing a 67% engagement spike, 50% boost in new user retention, and 65% jump in satisfaction ratings.
Vercel confirms hackers accessed customer data before its early-April breach, with infostealer malware suspected as the attack vector targeting tokens and API keys, as investigators warn the incident is broader in scope with more victims expected to emerge.
Google launches sweeping AI upgrades to Workspace at Google Cloud Next, introducing Gemini-powered tools that promise up to 9x faster spreadsheet building and smarter document drafting, as the tech giant intensifies its battle against Microsoft and Apple for dominance in AI-driven workplace productivity.