Perplexity’s Comet, OpenAI’s Atlas, and Google’s AI Mode are all fighting to own the moment when your browser stops waiting for clicks and starts acting on your behalf. Control is the issue. Whoever manages the agent that completes your purchase controls the data, the relationship, and the revenue.

Sam Altman announced that OpenAI expects to top $20B in annualized revenue this year, with plans to reach hundreds of billions by 2030. The company has committed to $1.4T in infrastructure spending over the next eight years. How exactly are they planning to pay for all this?

Apple has finally admitted what the rest of us have known for years: Siri is terrible. After more than a decade of living in last place among voice assistants, Siri is about to borrow someone else’s brain.

This is the first major public clash between a top retailer and an AI company over agentic transactions. If you’re thinking about building AI systems that execute actions in third-party environments, expect a wave of new access controls, licensing rules, and compliance audits.

Studio Ghibli, Bandai Namco, Square Enix, and other Japanese rights holders have asked OpenAI to stop using their content to train Sora 2.

Once teams master generative tools and see their productivity improve, many begin to ask, “If AI can do this much of my work, what remains that’s uniquely mine?” Researchers call this Identity Threat, and it’s becoming one of the biggest hidden obstacles to effective AI integration.

OpenAI just updated Sora 2, its AI video app, with tools that let users build reusable AI characters, stitch clips together into longer scenes, and remix content inside a social feed.

According to reports from Engadget and Tom’s Hardware, a livestream from last week featuring a deepfake version of NVIDIA CEO Jensen Huang drew more viewers than the company’s real event. It's time for countermeasures. Read more below.

The trend toward a highly agentic future is clear. I’m excited to see where this goes. Are you?

This is not a reason to panic, but it is a reason to plan. Governance, agent permissions, and human-in-the-loop safeguards have to be adapted to the new threat. The future of productivity is agentic, but so is the future of cybersecurity.

AI isn’t hard to understand. It’s hard to talk about clearly. The technology moves faster than the language used to describe it, which leads to misalignment and wasted time. Every effective AI meeting starts with a shared vocabulary. Use this glossary to make sure your team is on the same page.

Microsoft just announced Copilot Mode for its Edge browser, turning its traditional browser into an active AI-powered assistant.

Amazon’s latest delivery innovation, AI-powered smart glasses for its drivers, combines computer vision and augmented reality to deliver turn-by-turn navigation, package scanning, and proof-of-delivery capture without drivers having to look at their phones.

The flow of consumer intent is the most valuable signal in marketing. Google won’t wait long to ship a Gemini-powered Chrome with similar capabilities. To quote Yoda, “Begun, the agentic browser wars have.”

Adobe announced AI Foundry, a new managed service where its engineers work with clients to create brand-specific models trained on their own data and creative assets. Adobe is turning trust into a service. This may be what it takes to bring AI out of the lab and into the workflow.

A group of major players just introduced AdCP (Ad Context Protocol), which could change how media gets bought, sold, and measured. Instead of humans operating dashboards, AI agents will negotiate directly across platforms. If this works, ad operations will never look the same. Will it work?

What needs to happen before Agents start buying for us? • How do we define “shopping” vs “buying”? • What data access is required — and who do we trust to hold it? • What incentive would make us hand over control? When Are We Going to Let Bots Shop for Us? shellypalmer.com/2025/10/when...

Just when you thought enterprise AI had peaked, Google introduces Gemini Enterprise. Query your docs. Automate workflows. Build agents. The fight for the AI‑powered enterprise layer is fully underway. Read → shellypalmer.com/2025/10/google-launches-gemini-enterprise

The last time the MPA and the RIAA got their knickers in a knot, they were the 800-pound gorillas. This time… not so much.

OpenAI used DevDay to reposition ChatGPT as a place where work gets done. Here’s a quick overview.

My Sunday essay is about Tilly Norwood: the AI-generated “actor” at the center of Hollywood’s latest controversy. It has people debating whether machines can perform, create, or even feel, but maybe the real question isn’t about acting at all. Maybe it’s about first principles.

As of today, OpenAI is the most valuable startup in the world, and the signal to employees, rivals, and investors is unmistakable: the market believes.

This has been such a busy AI news week that I did not get a chance to properly talk about Anthropic’s new Claude Sonnet 4.5 upgrade. After a day of using it, it’s clear that 4.5 is a meaningful upgrade.

OpenAI just released Sora 2, a new video and audio generation model that feels like the jump from GPT-1 to GPT-3.5, but for moving images. You need to see the demo.

OpenAI just launched Instant Checkout in ChatGPT, powered by a new open standard called the Agentic Commerce Protocol. For the first time, U.S. users can buy directly from Etsy sellers inside ChatGPT, with Shopify merchants like Glossier, SKIMS, and Spanx coming soon.

Answer engines look for facts and context. They decide which brands, products, and sources to present when people ask questions in natural language. The industry expert consensus is that structured data that declares your entities, attributes, and relationships is key. This is a how-to guide.

Imagine if everyone had a world-class personal assistant who briefed you every morning, anticipated your questions, reminded you of your daily goals, and organized your day across apps. Heaven for some, hell for others, but an awesome use case nonetheless.

Those of us wishing Copilot didn’t suck just got a gift from our friends in Redmond. Starting today, Anthropic’s Claude Sonnet 4 and Claude Opus 4.1 models will be available to Copilot users in the Researcher tool and through Copilot Studio.

A new study published in Harvard Business Review by Stanford University and BetterUp Labs introduces a term that should worry every executive: “workslop.”

Compute and electricity are no longer background resources; they are board-level concerns. If you depend on AI to drive growth, you need to treat power supply and GPU hours as tier-one inputs. Prices will rise, scarcity will matter, and companies that secure access to both will have an advantage.

Reports have surfaced about SpamGPT, an underground “spam-as-a-service” platform -- think Mailchimp for criminals. My Sunday essay explores whether this tool is merely hype or a sign of what’s to come – and why leaders should prepare for the inevitability of commoditized cybercrime.

Yesterday, Google rolled out 10 new AI-powered features, calling it “the biggest upgrade to Chrome in its history.” Google’s strategy is clear: integrate Gemini into Chrome to evolve Google into a transaction-oriented answer engine.

One of my favorite quotes about the future comes from a paper written in 1965 by I.J. Good: “Thus the first ultraintelligent machine is the last invention that man need ever make.”

OpenAI is building a system to estimate whether a ChatGPT user is under 18. If it thinks you are, it will change how the product works. This is a design shift with serious implications. Age is just the start.

OpenAI just dropped a comprehensive study of consumer AI usage and the findings fundamentally challenge what we thought we knew about AI adoption.

The creative industry is facing an uncomfortable reality: AI-generated content is increasingly indistinguishable from human work, and audiences don't seem to care. My new essay examines why the distinction between "required" and "inspired" content changes everything about AI's role in business.

If you haven’t used the memory feature in ChatGPT or Gemini, you don’t know what you’re missing.

The web desperately needs a scalable solution for compensating creators whose work trains AI models. RSL provides the framework for that solution. Will it achieve the scale it needs for success? That’s up to the foundational model builders.

Apple was right about one thing: I was in awe of how little there was to care about.

For those keeping score, “This s#%t just got real.” We’re witnessing the emergence of “vibe-movie-making,” a process where you describe (prompt) what you want and AI generates it for you.

A relatively new AI model got a lot of attention because Sapient Intelligence (a Singapore-based company) open-sourced their Hierarchical Reasoning Model (HRM). The headlines screamed breakthrough… but is it? My Sunday essay will help you decide if HRM is a parlor trick or a paradigm shift. -s

Remember when Photoshop added the drop shadow feature? History's about to repeat itself.

AI-powered Pizza? Yes. Ooni (the makers of my favorite outdoor pizza oven) has announced an AI-powered indoor version called the Volt 2. This is actually more exciting than the AI-powered GE Profile Smart Indoor Smoker I got after seeing it at CES. Read more: shellypalmer.com/2025/09/ooni...

The smart move is to prepare to continuously adapt to the speed of technological change and to seriously concentrate on AEO (answer engine optimization). It’s the new, new thing.

We’ve built AI systems that mirror human psychology so closely that they inherit our social vulnerabilities. The more human-like we make AI communication, the more human-like its vulnerabilities become.

Yesterday, Anthropic quietly dropped a bombshell. Unless users explicitly opt out by September 28, it will use consumer chat data to train future AI models. This is a stunning reversal from Anthropic’s previous position as the privacy-first alternative to ChatGPT.

Google launched a free version of its AI-powered video editor Vids, while keeping premium AI features behind subscriptions. This is not a parlor trick – it’s a paradigm shift. The AI may not be a complete solution today, but by the time you’ve adapted your production workflows, it will be.

There's a meme that's been circulating around the interweb this week. I like the simplicity of this meme, but there are far more than three probable futures. Let's explore.

Search and web browsing behaviors have measurably shifted. People are asking questions and expecting synthesized, citation-rich answers. The trend is clear. The business must optimize for answer engines. Will the new paradigm remunerate original content creators? Let’s see how this goes.

White collar roles will evolve. Today, humans are in the loop, teaching systems how to do their work. Those humans will shift to writing and training agents, then to orchestrating fleets of agents. Then, the agents will operate and report anomalies and adjustments. This is the end state. Or is it?