The four stories on this desk share one thread: AI capability is being shipped faster than the business models around it can absorb. Google is growing search queries briskly while its search revenue decelerates; Uber fulfilled four rides from AI agents in a quarter; AT&T has wired Gemini into its home security system; and ChatGPT now makes animated GIFs. Products and features are arriving on schedule. The money and the usage are lagging behind them.
The monetisation gap at Google
Alphabet's earnings report last week, as The Verge reported, contained a genuinely awkward pair of numbers. Search queries are growing at their fastest rate in five years, but Search revenue growth slowed to 12 percent year over year in the third quarter, down from roughly 17 percent in the previous quarter. Growth in paid clicks was also part of the picture the outlet flagged.
The plain reading is that Google has figured out how to make AI drive engagement and has not figured out how to make it pay like the search engine it is layered onto. More queries producing slower revenue growth implies a lower yield per query, or a mix shifting toward interactions that do not carry the same advertising load. That is the central commercial question for the largest advertising business on the web, and it is unresolved. Google's AI features are doing their job at the top of the funnel while the bottom of the funnel narrows.
The demand that isn't there yet at Uber
Uber's disclosure, also via The Verge, is the starkest data point in the set. Only four rides were arranged by third-party AI agents during the third quarter of 2026. Uber Eats orders fulfilled by AI agents came in at around 500.
Four rides is not a rounding error in a meaningful sense; it is a rounding error in the literal sense. The agentic commerce thesis, in which software agents transact on behalf of consumers, has attracted enormous attention and investment. Uber's numbers show that on its platform the thesis is, so far, a demonstration rather than a market. The company disclosed the figure on an earnings call, which means it considers the metric worth reporting and small enough to report honestly. For US consumers, the practical consequence is that the agent-mediated economy remains a novelty rather than a channel.
Integration as the current form of progress
AT&T's Connected Life system now integrates with Gemini for Home, which The Verge describes as the first home security product to integrate with Google Home Premium. The Gemini-powered features summarise home events and search video history.
This is a real product with real utility, and it is also the shape that most AI commercialisation is taking in late 2026: a capability embedded into an existing subscription or hardware relationship rather than sold as a standalone thing. The revenue shows up as retention, differentiation or an upsell, not as a line item called AI. That is a legitimate strategy, and it is much harder to measure than a new product with its own price. It also means the returns are diffuse and slow, which is exactly the profile that sits awkwardly against the pace of feature shipping.
