Published on LinkedIn and amitabhapte.com | 1 November 2026
Key developments and opinions which caught my attention this week.
This Is What I Read This Week
The biggest story this week was where the ad money goes when AI answers the question. Meta and Google showed two sides of it. Meta’s ad revenue grew 26% to $63.1bn, helped by AI targeting tools. Google’s search revenue grew 12%, down from 19% at the start of the year, and Google admitted that its AI Mode earns less per query than a classic results page. Alphabet lost about $475bn of market value in a day, its worst session since it listed in 2004. The pain spread. Reddit fell 29%, its worst day ever, after US daily users dropped 5% in the quarter. Its CEO pointed to Google’s AI answers, which are sending fewer people to the site. Pinterest slid too, on slower revenue growth and lower ad prices.
Investors are tired of the bill. The Nasdaq 100 was on course to lose more than $2 trillion in October, as the largest tech companies raised spending plans again and leaned harder on debt to fund data centres. JP Morgan expects global AI capex of $900bn this year.
Agents hit another wall. JPMorgan now restricts outside AI agents that log in with customers’ bank details, allowing read-only access such as checking a balance. Schwab is blocking unauthorised agents from moving money, and Amazon has kept Meta’s Muse off its store since September. The common worry: who is responsible when an agent gets it wrong.
Security had a heavy week. The New York Times reported that AI is finding bugs faster than Silicon Valley can fix them. Microsoft issued more than 1,400 patches in October alone, more than in all of 2025, and Chrome has fixed more bugs since June than in the previous two years combined. And Binance lost about $6.3bn in a hack that reportedly began with one compromised engineer’s laptop. Stablecoin issuers froze around $3bn, and Binance says it has set aside its own funds to cover the losses.
So What? My PoV and Takeaway This Weekend
The web ran on a simple deal for twenty years. Search sent people to your site, and you earned from the visit. AI answers are quietly breaking that deal. The answer arrives, the click never happens, and the money flows to whoever owns the conversation and the data behind it. A community site losing a million search referrals a week tells you the shift is already here.
The same pattern shows up in agents and security. Banks are blocking agents because nobody has agreed who carries the risk. AI finds flaws faster than teams can fix them. One laptop opened the door to a multi-billion-dollar theft. In each case the technology moved first, and the rules and the people behind it are catching up. Markets noticed. A $2 trillion month is investors asking the biggest spenders when the returns arrive.
My takeaways this week. Map how your customers find you when an AI gives the answer instead of a list of links, and treat being cited in that answer as a channel worth measuring. Fund fixing as well as finding: scanning tools are cheap, while the engineers who patch are scarce. And give every AI investment a payback story your board can test, because the market just asked the largest companies in the world for exactly that.
