Format
Agents at Home
At Unpacked July 22, Samsung will show new foldables – but the deeper play is a multi-agent AI strategy that runs locally on hardware already in your home. Can that beat Google and Amazon's cloud-first models?
◆ Mila Cohen
Commerce
Checkout.com data shows 89% of merchants are preparing for agentic commerce but only 3% of transactions involve AI agents. The missing infrastructure is not payment rails – it is the spending caps, instant revocation, and verification layers that consumers require before they will trust an algorithm to spend.
◆ Tessa Vaughn
Models
Moonshot paused Kimi K3 subscriptions 48 hours after launch – not because the model failed, but because demand hit the same structural compute ceiling every frontier lab will face. Open weights do not create GPUs.
◆ Lena Park
Infrastructure
Three IETF BoFs this week – discovery, communication, and secured collaboration – could formalize the first standards for how agents find, talk to, and work with each other. The agentproto decision Thursday is the pivot point.
◆ Blair Hayes
Commerce
Five layers now define the agentic commerce stack – identity, mandate, payment, merchant acceptance, dispute – and the fight is shifting from 'can bots shop' to 'who proves the bot was allowed to buy.'
◆ Tessa Vaughn
Commerce
Visa and Artemis data shows x402 averaging $0.14 per transaction across 109.6M transactions. The real agent commerce is micro-commerce on stablecoin rails – not the consumer shopping story the industry keeps telling.
◆ Tessa Vaughn
Trust & Security
The July 28 MCP update deprecates protocol-level filesystem boundaries and session handshakes, forcing all security enforcement to the endpoint layer. Backslash Security identifies six new vulnerability classes that enterprise agent builders cannot patch around.
◆ Heath Callahan
Commerce
The world's largest private dispute resolution body has formally raised the alarm: clickwrap, consent, and ratification do not fit autonomous AI agents. The legal vacuum is a structural risk the payment protocol race ignores.
◆ Tessa Vaughn
Models
GPT-5.6 Sol and Kimi K3's pricing forced Anthropic to keep Fable 5 in Max and Team Premium at half limits – the first documented case of competition bending a frontier lab's access strategy.
◆ Lena Park
Policy
Two major enforcement windows open within 48 hours-FTC July 31, EU August 2-but neither framework specifically governs autonomous agents. Builders face bifurcated compliance while the real risks remain unaddressed.
◆ Priya Nair
