One thing that keeps bugging me: every provider says the same line — “you own your data, we own the machine.”AWS, Azure, Google… they all frame it that way. But in practice, I’m stuck asking myself: - If 80% of breaches come from misconfigs, am I really the “owner” — or just the one carrying the blame? - If AI tools quietly log my interactions unless I dig for the opt-out, what exactly am I consenting to? - If my data stays local but models trained elsewhere still learn my patterns… what does “ownership” even mean? This feels like a blind spot when picking tools — ownership vs. liability. Curious: do you treat this as a major risk factor when evaluating AI platforms, or do you prioritise speed and features and deal with the rest later?

https://blog.google/products/google-one/google-ai-plus-expands/

• OpenAI launched “OpenAI Grove(networking with OpenAI researchers and peers.),” 5 weeks program at OpenAI San Francisco HQ (in-person) designed for individuals at the very start of their company-building journey, OR preparing to launch a company, especially in AI. Link : https://openai.com/index/openai-grove/ • OpenAI and Microsoft are continuing their partnership to create the best AI tools, focusing on safety and helping people. OpenAI, which started as a nonprofit, will continue to guide its work while also having a stake in its Public Benefit Corporation (PBC) worth over $100 billion. https://openai.com/index/joint-statement-from-openai-and-microsoft/ • As part of this effort, OpenAI is launching a $50 million grant program to support nonprofits and community organizations in the U.S. in AI education, community projects, and creating economic opportunities. • OpenAI is working to expand economic opportunities with AI: ○ AI Jobs Platform: Connects AI-skilled people with companies, including support for local businesses and governments. ○ OpenAI Certifications: Provides proof of AI skills, from basic usage to advanced prompt engineering. ○ OpenAI Academy: Free online learning platform already helping 2+ million people; now expanded with certifications. Goal: Certify 10 million Americans by 2030, giving more people the skills to use AI, earn better jobs, and create new opportunities. https://openai.com/index/expanding-economic-opportunity-with-ai/ • OpenAI acquired Statsig for 1.1 billion dollar (Indian Vijaye Raji is joining OpenAI as CTO of Applications). Statsig helps test and improve products safely and quickly. Statsig is like a tool that checks if new features or ideas work well for users before releasing them fully. • OpenAI is locking in huge cloud and chip deals ($300B with Oracle, $10B with Broadcom) to secure the compute power and design its own AI chips, it needs for the next wave of AI.

• Claude (Anthropic’s AI) can now directly create and edit real files like Excel spreadsheets, Word documents, PowerPoint presentations, and PDFs- not just give text answers. ○ ⚠️ Caution: Because Claude needs limited internet + computer access to make files, there are data security risks, so users are advised to monitor closely. ○ 👉 In short: Claude can now act like a digital co-worker who builds Excel models, PowerPoints, reports, and PDFs for you - not just suggest ideas. ○ Available to Max, Team, and Enterprise plan users. Pro users will get access in the coming weeks. ○ Link : https://www.anthropic.com/news/create-files ------------------------------------------------------ • Anthropic’s announced that it is endorsing SB 53. ○ SB 53 is a proposed California law that sets safety and transparency rules for companies (like Anthropic, OpenAI, Google DeepMind, Microsoft) building the most powerful AI models. ○ forces frontier AI companies to be transparent and accountable about safety. • Anthropic tightening its sales restrictions to prevent misuse of Claude in regions that pose national security risks, especially authoritarian nations like China. • Anthropic deepening its AI safety work through formal partnerships with US CAISI (Center for AI Standards and Innovation) and the UK’s AISI (AI Security Institute). To find and fix vulnerabilities in Claude, creating stronger defenses against jailbreaks and misuse. https://www.anthropic.com/news/strengthening-our-safeguards-through-collaboration-with-us-caisi-and-uk-aisi ------------------------------------------------------- • Claude now remembers your work context like a team assistant with long-term memory while giving you full control (and incognito mode when you want a clean slate). https://www.anthropic.com/news/memory

- Google launched “Nano Banana” in Gemini - After OpenAI’s Ghibli-style images went viral, this new trendy tool turns photos into stylized 3D figurines and creative videos. - ByteDance debuted Seedream 4.0 - An ultra-fast 2K resolution AI image generator with multi-image style reference. - Google updated Veo 3 - The upgraded video generation tool now supports vertical format (9:16), 1080p resolution, and lower cost per second. - Grok Imagine - Supports very fast, high-quality image and short video generation from text or image prompts. - Perplexity - Now supports video generation for Max, Pro, and Enterprise Pro plans.