OpenAI opens up

OpenAI & Microsoft Scrap Their Exclusivity Deal

Microsoft and OpenAI just rewrote the terms of their partnership. They’re moving to a non-exclusive relationship, where OpenAI's models are no longer locked to Microsoft's cloud.

What changed: Microsoft first invested in OpenAI in 2019 and has poured over $13 billion since, embedding OpenAI's models into Azure, Office, and Bing. Now, Microsoft's license to OpenAI's intellectual property, which runs through 2032, is non-exclusive. 

• OpenAI can sell its models directly through Amazon Web Services, Google Cloud, and any other provider. Amazon just integrated OpenAI's latest models into AWS, including AI-building service Bedrock.

• In exchange, Microsoft will stop paying a revenue share to OpenAI; OpenAI will continue giving Microsoft a portion of its revenue through 2030, but subject to a cap.

Why this happened: Tension had been building since OpenAI signed a $50 billion cloud deal with Amazon—which the Financial Times reported Microsoft was weighing legal action over. Monday's announcement resolves that standoff. An internal OpenAI memo called demand on AWS "staggering" and admitted exclusivity had capped its enterprise reach.

Musk & Altman Head to Court Over OpenAI's Future

After nearly two years of buildup, Musk v. Altman is finally in court. Opening arguments kicked off Tuesday. Just before trial, Musk dropped his fraud claims against OpenAI, Altman, and co-founder Greg Brockman—narrowing a sprawling 26-count suit down to two: unjust enrichment and breach of charitable trust. 

What’s at stake: Even scaled back, the stakes are huge. Musk is seeking:

• Up to $134 billion in damages 

• A court order restoring OpenAI's nonprofit status

• Removing Altman and Brockman from their roles

How this started: Musk was an original OpenAI co-founder and helped fund its early days as a nonprofit AI safety lab before leaving the board in 2018. His argument, detailed in the original lawsuit, is that Altman and Brockman diverted the organization from its founding mission (AI for the benefit of humanity) into a commercial enterprise that primarily enriched themselves.

China Blocks Meta's $2 Billion Manus Deal

This week, China's top economic planning body vetoed Meta's acquisition of AI agent startup Manus, unwinding a deal that had been months in the making and had already upended 100+ employees' lives.

Some Manus history: Founded in 2022, Manus specializes in agentic AI. Meta announced the acquisition (valued between $2-3 billion) in December 2025. It would fold Manus's agents directly into Meta AI as part of Mark Zuckerberg's broader push into the agentic AI space. 

Why China blocked it: China's National Development and Reform Commission issued the prohibition with no detailed explanation, stating only that it had "made a decision to prohibit foreign investment in the Manus project." But it’s a clear signal that Beijing remains wary of Chinese-origin technology landing in US hands. 

• Complicating things further: roughly 100 Manus employees had already moved to Meta's Singapore offices by March, and Manus' CEO and chief scientist are now reportedly under exit bans that prevent them from leaving China.

Build a Live Stream of Your Data

Every Monday you open Stripe, then Mixpanel, then a spreadsheet, then Slack to share the numbers. By the time you're done, the data is already old.

What if you built that dashboard once - and it just stayed live?

OptiDev connects to your tools and turns them into live applications. Not static charts. Not exported PNGs. Real apps with real data flowing through them, always current.

• Connect once, live forever: your data stays current automatically

• Databases & tools: PostgreSQL, Snowflake, Stripe, Mixpanel, and more

• Share with your team: deploy with a custom domain in one click

• Full backend included: auth, database, encrypted connections

• New connectors shipping regularly: more tools, more data, one dashboard

Build it once. Never copy/paste a number again.

Multilingual AI agents for customer interactions

ConvoZen.AI deploys autonomous, multilingual AI agents across voice, chat, WhatsApp, email, and social channels—keeping context across sessions and giving live agents real-time assistance. 

How you can use it

• Use low-latency, culturally adapted models trained on real contact-center data

• Get real-time next-best-action recommendations during live customer calls

• Automatically evaluate 100% of interactions for sentiment, compliance, and quality

Pricing: Paid

Backend infrastructure from your frontend code

Butterbase plugs into AI coding tools via MCP, then reads your frontend code and provisions a production-ready Postgres database, authentication, typed CRUD APIs, real-time sync, and globally deployed edge functions. 

How you can use it

• Generate schemas, migrations, and API routes from natural-language or in-editor prompts

• Add auth (JWT/OAuth) and row-level security without writing backend code

• Ship from frontend to database in minutes, with standard exportable Postgres/REST artifacts

• Integrate directly into your existing AI coding workflow via one line in your MCP config

Pricing: Free and paid plans