Amazon bars Meta's Muse agent from shopping its site

Share
Amazon bars Meta's Muse agent from shopping its site

Two agentic-commerce stories landed within hours of each other: Amazon enforced its own store rules against a rival's shopping agent, and Huawei put a coding model in front of every HarmonyOS developer.

Amazon cut Meta's Muse shopping agent off from Amazon.com over the weekend, and told users why in a popup: continued access by an unauthorized AI agent violates the retailer's Conditions of Use. Amazon's objection is less about competition than about consent and disclosure — per GeekWire, Meta never told Amazon that Muse would shop its store, the agent does not identify itself when it browses, and Amazon says it appeared to capture customer credentials. "We think it's fairly straightforward that third-party applications that offer to make purchases on behalf of customers from other businesses should operate openly and respect service provider decisions about whether or not to participate," an Amazon spokesperson said. Techmeme's summary adds merchant consent to the list of cited reasons.

That framing is the story. Amazon has spent a year arguing that an agent acting for a user is still the user's agent, and that sites get to decide who gets in — the company sued Perplexity over Comet shopping last November, lost the injunction in August, and has since stripped item names and product images out of its order-confirmation emails to limit what outside agents can mine. Muse is a harder case for Amazon than Perplexity's browser, because the two companies are partners: Meta signed a multiyear deal in April to run agentic workloads on AWS Graviton chips, and Amazon products have been purchasable inside Facebook and Instagram since 2023. Meta's own launch materials said Muse cannot see secure login details or card payment information; Amazon's block says it isn't confident that holds. We went through Amazon's contradictory position on agents in Amazon blocked AI agents for a year. Now it sells ads inside one.


Huawei Cloud opened a coding model tuned specifically for HarmonyOS to the ecosystem's developers, with the largest error reduction any vendor has put on a number: 80 percent fewer errors per thousand lines of code, plus more than 20 percent fewer tokens for the same task. The upgrade to its CodeArts coding agent bundles three pieces — the HarmonyOS coding model, a CodeArts HarmonyOS agent, and a developer practice center — and Huawei calls it the first AI coding agent built for the HarmonyOS ecosystem alone. The model trains on HarmonyOS corpora, covers the high-frequency ArkUI components and the ArkTS syntax constraints, and the agent can drive an on-device emulator directly, which Huawei says is an industry first. HarmonyOS now has more than 11 million registered developers.

The numbers deserve the usual discount — they are Huawei's own, measured on Huawei's own corpus — but the direction is plain. Huawei is not competing for general coding-agent share; it is building a moat around a platform whose persistent weakness has been the cost of writing apps in a language almost nobody else writes. The same summit brought a date for its next Ascend 950DT cluster service and a sandbox runtime it says starts in a hundred milliseconds, all aimed at the same enterprise buyers.

What to watch: whether Meta renegotiates Muse's access with Amazon or routes around it, and whether Huawei's 80 percent error figure survives testing outside Huawei.

If an agent buys on your behalf, should the merchant get a veto over which agent walks in? Tell us in the comments.

Read more

Mistral's Le Chonk puts Europe's sovereignty bet on a download date

Mistral's Le Chonk puts Europe's sovereignty bet on a download date

Mistral's biggest model ever is real, benchmarked and for sale today — but the thing that makes it matter to Europe's sovereignty argument, the weights, is still three weeks out. The preview settles who built it; the release will settle whether it counts. What Mistral actually shipped Mistral opened a public preview of Mistral Large 4 — unofficially ML4, officially le Chonk — a 1 trillion-parameter mixture-of-experts model with 49 billion active parameters and native multimodal input. The p

Mistral unveils Le Chonk: a 1T-parameter open-weights model

Mistral unveils Le Chonk: a 1T-parameter open-weights model

The biggest open-weight release outside China lands in public preview today, and the country that spent the week promising its own frontier model just put a price on the ambition. Mistral has opened a public preview of Mistral Large 4 — codenamed "le Chonk" — a 1 trillion-parameter mixture-of-experts model with 49 billion active parameters, natively multimodal, which the company calls its largest and most capable model to date. The preview API is live today on Mistral Studio at $1.36 per milli

Google signs 3.6 GW power deal, a quarter of it new nuclear

Google signs 3.6 GW power deal, a quarter of it new nuclear

Grid capacity, not chips, is becoming the binding constraint on the AI buildout — and on the same day, the labs told an Australian inquiry they can live with mandatory incident reporting. Google has contracted 3.6 GW of power from Constellation Energy across the PJM grid — the largest electricity deal in the region's history, and the biggest single power commitment any AI company has made. The agreement covers 3,590 megawatts over 13 states, with 890 megawatts of new nuclear capacity coming fr

Korea probes AI agents in bank hacks as president cites 'signs'

Korea probes AI agents in bank hacks as president cites 'signs'

South Korea opened a formal investigation into whether AI agents drove a wave of bank breaches — and it isn't the only AI story moving money today. President Lee Jae Myung said "signs" point at AI models, and the probe is now at the highest level a national banking sector has seen. Speaking at a cabinet meeting, Lee said that "in some hacking incidents, signs have emerged of AI being used, causing considerable public concern and anxiety," and police have since opened a full-scale investigation