Dr. Dennis Weekly AI Blog - May 12, 2026

AI agents can now think, act, and even spend money without you

Author

Dr. Dennis Addo
May 12, 2026

Happy New Week, Everyone!

Have you ever felt like you're running on a treadmill that keeps speeding up? That's exactly what the AI landscape feels like right now, with model release cycles compressing from 18 months down to just 4 months. But don't worry—I've done the heavy lifting to distill the noise into the signals that actually matter for your business. This week, we're looking at a massive shift: AI is moving from a tool we use to an agent that acts on our behalf, and the biggest players are racing to build the infrastructure to support it.

The Trend We're Loving

The most significant trend this week isn't just a new chatbot—it's the rise of "Agentic AI" capable of autonomous action and transactions. We are moving beyond AI that simply answers questions to AI that can execute complex, multi-step workflows and even spend money on your behalf.

This matters because it fundamentally changes the digital economy. When AI agents can autonomously purchase APIs, book travel, or manage supply chains, businesses need to optimize not just for human consumers, but for machine intermediaries.

Key Implications:

• Machine-Mediated Commerce: AI agents will increasingly act as purchasing intermediaries, requiring brands to rethink their sales and marketing strategies.

• New Payment Infrastructure: Traditional payment rails are too slow and expensive for AI micropayments, driving the adoption of stablecoins and specialized protocols.

• Governance and Security: As agents gain autonomy, new frameworks are urgently needed to manage risks, ensure accountability, and prevent unauthorized actions.

Deep Dive

The AI That Dreams to Improve

Anthropic just introduced a fascinating new capability for its AI agents: "dreaming." Unveiled at their developer conference, this technique allows AI systems to review their past behavior, identify patterns, and refine their working memory between active sessions. Imagine an employee who spends their off-hours analyzing their mistakes to ensure they never repeat them.

This is a major step toward self-improving AI. By running evaluations in the background, these agents can establish better ways of working and significantly reduce errors over time. Anthropic is pushing hard to expand these autonomous capabilities beyond software engineering into high-stakes sectors like finance and law, where accuracy and continuous improvement are paramount.

OpenAI's Push for the Enterprise Wallet

The race for enterprise AI dominance is heating up, and OpenAI is making aggressive moves to capture corporate budgets. This week, they rolled out GPT-5.5 Instant, a new default model for ChatGPT that boasts reduced hallucinations in sensitive areas like law and medicine, alongside improved context management. But the real story is their push into services and advertising.

OpenAI is reportedly launching a massive $10 billion joint venture with private equity firms to deploy enterprise AI services, mirroring a similar $1.5 billion move by Anthropic. Furthermore, OpenAI launched a self-serve advertising platform for ChatGPT, aiming for $2.5 billion in ad revenue this year. They are rapidly transforming from an AI research lab into a full-stack enterprise services and commercial advertising juggernaut.

Apple's "Choose Your Own AI" Adventure

Apple is reportedly preparing a major shift for iOS 27, internally dubbed "Extensions," which will allow users to select third-party AI models to power Apple Intelligence features. Instead of being locked into a single ecosystem, iPhone users might soon be able to choose Google's Gemini, Anthropic's Claude, or others for tasks like text generation and image editing.

This move signals Apple's strategy to accelerate its AI competitiveness by opening its platform rather than building everything in-house. For the broader industry, it means the battle for AI supremacy will increasingly take place directly on the devices in our pockets, with providers competing fiercely to be the default choice for billions of users.

Dr. Dennis' Picks

Here are 5 AI tools that launched or got major updates this week:

• 🔥 GPT-5.5 Instant – OpenAI's new default model with reduced hallucinations and better context memory. Try it here

• 🔥 NotebookLM (Auto-Labels Update) – Google's note-taking AI now automatically organizes and labels your sources. Try it here

• 🔥 Amazon Bedrock AgentCore Payments – Lets developers build AI agents that can autonomously pay for services using stablecoins. Try it here

• 🔥 OpenAI Realtime Audio Models – New models for live voice interactions, translations, and speech-to-text tasks. Try it here

• 🔥 CellSAM – Caltech's new AI tool that automatically identifies and segments individual cells in biological images. Try it here

Takeaways


The era of the autonomous AI agent is here. From Anthropic's "dreaming" agents to AWS enabling machine-to-machine payments, AI is moving from passive assistance to active execution.

Insight: "We are witnessing the productization of AI systems which do work on behalf of people, acting independently for significant periods of time."

Action: Audit your most repetitive digital workflows this week and identify one that could be fully delegated to an AI agent.

Next week, we'll explore how the new AI governance frameworks are reshaping corporate compliance.

Forward this to colleagues who need to stay ahead of the AI revolution. Have insights to share? Reply – I read every message.