Entertainment value (most to least). video games/interactive video > video > audio > photo > text. Friction involved to create each (least to most): . text < photo < audio < video < video games/interactive video. i.e., tweeting this is easier than creating a video game.

Engagement ≠ monetization, at least not right away. We’ve seen this story before: search engines, social media, messaging apps, connected TV… Tons of users, lots of excitement, but the money took time to really show up. AI feels different because subscriptions are baked in.

Join millions who have switched to Grok.

🤷‍♂️

What are some good examples of each?

Be honest: when was the last time you truly browsed the internet?. In 10 years, browsing as we know it will be nearly extinct. Because we perceive AI as browsing on our behalf, we stop doing it ourselves. Instead of digging through results, we get immediate, tailored answers.

RT @LorenzoARK: Ethereum’s inflation is running at just 0.74% annually, with sub 10 gwei gas. Since April 2nd, ETH DATs have bought 3.57M….

When AI crosses the uncanny valley, do we celebrate the breakthrough — or begin to devalue the medium, now that it no longer feels uniquely human?.

We waited 2 years for GPT-5 and it’s just a wrapper for all the other GPTs?.

A new monetary system is emerging in the digital world. Chips manufacture the raw currency: compute. Compute becomes the medium of exchange in AI. Cloud platforms act like central banks—allocating, pricing, and controlling access. Large language models spend compute to generate.

The greatest innovations in history share a common trait:. They give us time back. The car collapsed distances. The washing machine eliminated hours of manual labor. The computer automated complex calculations. The internet reduced the cost of knowledge transfer to near-zero,.

Also, here is Part I, which walks you through the history of commerce:.

By 2030, AI agents could drive 72% of online commerce, reshaping the entire value chain. The battleground is shifting—from branding and eyeballs to data access and operational execution. This is Agentic Commerce. 📖 Read the full piece:

To power this shift, Model Context Protocol (MCP)—pioneered by Anthropic—could serve as the “pipes” of agentic commerce. MCP enables AI agents to pull real-time merchant data (price, inventory, fulfillment) directly into their context window. Goodbye SEO. Hello Agent-Oriented

Instead of browsing listings, a consumer tells an AI agent what they want. The agent already knows their preferences, budget, and context. It bypasses search, filters marketplaces, and transacts—all in one step. This is “one-query purchasing.”

Part II of our work on Agentic Commerce is out. AI agents are becoming an abstraction layer between consumers and marketplaces. They compress the funnel—from search and discovery to purchase—by bringing knowledge to the transaction itself. 🧵

In past innovation cycles, new technology enabled startups to disrupt the incumbents. But what if AI is different?. What if the scale of capital and self-reinforcing moats around compute, data, and distribution make this a game only the tech giants can play?.

We’ve spun the AI flywheel enough times that only a handful of companies still have the capital—or access to it—to stay in the game. Even fewer have the distribution and cash flow from existing businesses to tip the scales. At scale, they can slash prices, bundle into existing.

The AI flywheel today:. Compute →.Talent →.State-of-the-art models →.Better products →.More usage →.🔁Back to compute. Each turn of the wheel demands more computing power, and each unit of compute demands more cash. Training state-of-the-art models is not just a technical.

Ask your favorite LLM to generate an image of an analog clock showing any time, except 10:10 😉. Please reply with results.

RT @JozefARK: Had a great conversation with @GrousARK and @skorusARK about how AI models like Grok 4 are making progress towards practical….

It’s rare to see a company so in tune with both its users and its market. Robinhood gets it — simple incentives, real value, and user-first design.