gm.
three stories worth knowing since monday.
the 3
on sunday night, amazon blocked meta's muse from completing any purchases on amazon.com. the day after, shopify announced it would open every store on its platform to the same agent. two very different bets on how AI agents fit into shopping.
context: muse is meta's personal AI agent, launched september 8. it's the "assistant that actually does the work" pitch that people have been waiting on. muse can send emails, book appointments and travel, buy movie tickets, fill out forms, monitor your home cameras, and shop online for you. users can name their muse, give it a custom avatar, and tune how it talks. that's why the personalized muse pfps may have been all over your feed. it runs on iOS, android, whatsapp, and muse.ai. free for most tasks, paid tiers at $20 and $100 per month for heavier use. downloads have been surging in september.
why amazon blocked it: muse didn't identify itself as an agent inside amazon's checkout flow, which violates amazon's terms of service. amazon also said it was concerned muse could access customer account data, purchase history, and payment info as an "undisclosed third party moving through customer accounts." amazon asked meta to keep amazon out of muse's scope. meta didn't. amazon pulled the plug.
why shopify went the other way: shopify's business is helping merchants sell more stuff. if muse brings shoppers to their stores, that's growth. amazon's business is different. amazon makes nearly $70 billion a year in ads, plus it needs you to keep coming back to amazon.com so it can market prime and its own products to you. if muse buys for you from wherever the best deal lives, you never see the ads and you might never visit amazon at all. amazon has a lot more to lose from agent shopping than shopify does.
2. AMD just hit a $1 trillion market cap for the first time.
on monday, AMD (advanced micro devices, the chipmaker that competes with intel in CPUs and nvidia in AI GPUs) crossed $1 trillion for the first time. shares are up nearly 180% year to date. the stock jumped roughly 9% on the day it crossed and had ripped about 24% over the prior week.
context: AMD is now roughly 5x smaller than nvidia (~$5.4 trillion) but has jumped past intel by more than $350 billion. the rally is riding the same wave that lifted every major AI chip name this year, but AMD specifically has been closing gaps in the AI training and inference market where nvidia has dominated. investors are pricing in a real second player.
what this signals: chip stocks are now some of the biggest weights in the S&P 500. AMD, nvidia, broadcom, TSMC, and micron are all trillion-dollar names or close. if AI infrastructure spending slows, the whole index feels it. if it doesn't, chip valuations have more room to run.
3. NYT: draftkings uses AI to target the gamblers likeliest to lose.
a new york times investigation published friday says draftkings has been using a machine learning model since 2023 to identify which of its customers are most likely to keep gambling and losing money. the model scores each customer on what workers internally call an "elasticity" score: playing frequency, daily account balance, and the ratio of losses to wagers. higher score means the customer is worth targeting with more promos.
draftkings used this system to distribute roughly $400 million in AI-automated promotions in 2025 alone. the NYT also reported that draftkings shelved an internal effort to build a separate AI model that would identify problem-gambling behavior, while continuing to build out the targeting model. draftkings denied the framing and said its promos go to customers with "sustained, engaged use of our platform," not customers based on their losses.
why it matters: US sports betting handled $166 billion in 2025. this is one of the first inside looks at how a top sports-betting platform uses ML to segment and target customers. worth watching how regulators respond.
see you monday.
want the Higher Watchlist?
refer 3 friends to unlock our monthly rundown of the coins and narratives on our radar.
not financial advice. aggregated community trends and commentary.