Today in AI — April 10, 2026

📖 3 min read

AI news moved fast again, and today’s theme is obvious: the industry is shifting from flashy demos to real product packaging, safer deployment, and bigger infrastructure bets.

OpenAI quietly adds a new $100/month Pro tier

OpenAI has introduced a new ChatGPT Pro plan at $100 per month, a midpoint between its mass-market $20 offering and the much steeper $200 power-user tier. On the surface this looks like pricing cleanup, but it is really a signal that AI vendors are now carving up the market with much finer precision.

The bigger takeaway is monetization pressure. Frontier labs need more revenue without killing adoption, and a middle tier gives heavy users, builders, and small teams a more realistic upgrade path. If you follow the AI tool business closely, this is exactly the kind of packaging move that tends to ripple across the whole market.

Source: TechCrunch

Anthropic locks in a massive Google and Broadcom compute deal

Anthropic says it has signed a new agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity, expected to start coming online in 2027. The company also said its run-rate revenue has now passed $30 billion, which is an eye-popping number even in today’s AI market.

This matters because compute is now strategy, not just infrastructure. The winners in AI will not only be the companies with the best models, but the ones that can reliably secure enough chips, cloud capacity, and redundancy to keep training and serving at scale. Expect more headlines like this as the AI arms race moves deeper into energy and hardware economics.

Source: Anthropic

Microsoft starts removing Copilot buttons from Windows 11 apps

In one of the more telling product reversals of the week, Microsoft is beginning to remove some of the extra Copilot buttons it had pushed into Windows 11 apps like Notepad, Snipping Tool, Photos, and Widgets. The AI features are not disappearing, but the branding overload is getting toned down.

That is a subtle but important shift. The first wave of AI product strategy was “put AI everywhere.” The second wave looks more like “keep the capability, remove the clutter.” Good AI UX is starting to mean fewer gimmicky entry points and more contextual usefulness, which is probably healthier for mainstream adoption.

Source: The Verge

Tubi becomes the first streaming service with a ChatGPT app

Tubi has launched what is being described as the first streaming service app inside ChatGPT. Users can call up the app with @Tubi, describe the kind of movie or show they want, and get curated recommendations they can immediately watch on the platform.

This is a small story with big implications. AI assistants are rapidly becoming a discovery layer, not just a chatbot. That means distribution could increasingly flow through conversational interfaces instead of classic search boxes or app menus. If you publish, review, or promote AI products, that shift is worth tracking closely. For that angle, our friends at AiToolCrush.com are in the right lane.

Source: The Verge

Sierra says the era of clicking buttons is ending

Sierra’s Bret Taylor is making a bold bet that traditional click-heavy software is on the way out. The company’s new Ghostwriter product is designed as an “agent that builds other agents,” letting users describe a task in natural language and then deploy a specialized agent to handle it.

That vision lines up with a broader shift toward autonomous workflows, agent wrappers, and natural-language interfaces replacing rigid UI patterns. It is also why people building in this space, including the OpenClaw and AI agent crowd, keep getting more attention. If you want the business side of these shifts, BetOnAI.net remains a useful read.

Source: TechCrunch