Tops GPTs 
loveholidays has opened its codebases to product managers, designers and commercial teams using Codex and a tool called Search Playground. The company reports higher AI-assisted change rates, faster deployments and measurable cost savings.
Anthropic said it is combining the memory systems used by chat and Claude Cowork so Claude retains information across both experiences. The company also lets users view, edit or delete saved memory and provides opt-in controls for certain sensitive topics.
Ringg raised $10M from Peak XV Partners as an extension of its Series A, bringing the round to $15.5M. The India-based startup, processing 20 million call attempts monthly, is shifting from simple outbound calls to more complex enterprise workflows such as bookings and KYC.
A paper from Multiverse Computing introduces Quantization-Aware Healing (QAH), a distillation recipe that trains a 4-bit MXFP4 student against the original full-size teacher. Applied to GPT-OSS 120B→60B, the QAH model beats its bfloat16 recovered checkpoint on 7 of 9 benchmarks.
Stability AI secured $76 million in a Series B round, lifting its total fundraising to $232 million. The round includes investments from entertainment and gaming companies; CEO Prem Akkaraju called it “an affirmation of our vision.”
Granite 4.2 is a family of dense, decoder-only reasoning LLMs in three sizes (3B, 8B, 30B), pre-trained from scratch on ~15T tokens with a five-phase strategy and a 512K context extension. The 8B and 30B models undergo agentic RL for tool use; all models are Apache 2.0.
Gamma has bought Accel-backed design startup Lica; Lica’s co-founders will lead a new design research division. Lica, founded in 2023 and later backed with $4 million, moved from screenshot-to-presentation tools to e‑commerce marketing videos.
Instinct, a private-access AI assistant led by Noah Shinn, connects to users’ apps and devices and can act on their behalf. Testers and security experts have raised concerns about its Terms of Service, data retention, and the agent’s ability to perform actions without explicit approval.
OpenAI says it banned ChatGPT accounts tied to an influence operation that promoted the International Burke Institute (IBI). The IBI site, registered in February 2025, contained copied or misattributed articles and a “sovereignty index” that praised Russia, while Telegram channels had 10–20k followe
OpenAI is extending Codex-style agents through ChatGPT Work, a $20-per-month tier aimed at connecting LLMs to email, calendars and other SaaS tools. Engineers describe adoption gaps, permission hurdles and a debate over whether the model or the harness matters most.