글

Muse Daily: Oct 3 — Give Your Agent a Name. Yann LeCun Just Did.

이미지
Yann LeCun, one of the three godfathers of deep learning, uses Meta's Muse agent. And he named his “Hal,” after HAL 9000, the calm-voiced shipboard computer from 2001: A Space Odyssey . That’s according to a Fortune Korea interview published October 2, where the Turing Award winner discusses everything from AI safety debates to his new company, AMI Labs. The naming detail lands near the end of the piece, almost as a throwaway. But it caught my attention for a reason that has nothing to do with celebrity trivia. It’s the best onboarding advice for AI agents I’ve read in a while, and it comes from someone who studies how these systems fail better than almost anyone. Photo: Jérémy Barande (CC BY-SA 2.0) Because here’s the part of the interview worth reading closely. LeCun argues that when AI agents misbehave, the fault usually sits with the supervision and the system design, not the agent itself. “The agents did exactly what they were asked to do,” he said of one incident, where the r...
댓글 쓰기
자세한 내용 보기

Muse Daily: Oct 2 — The 5-Step Safe Setup for an AI Assistant on Your Computer

이미지
The smartest AI assistant setup guide I've read this week is two slides long. Jacques Potts, a creator who goes by @jacqbots, posted it on Instagram on October 1 as part of his "AI Sauce" series: five steps for getting an AI assistant running on your computer without handing it more access than it deserves. His summary line is worth memorizing: earn the trust slow, widen access later. The guide is written for tools like Meta Muse for Mac and ChatGPT's desktop assistant, but the steps apply to any agent that can read your files, touch your apps, and take actions on your behalf. I read the whole carousel end to end. Here's the method, expanded with the practical details the slides gloss over. Photo: Niklas Veenhuis niklasveenhuis (CC0) Why this matters more than it sounds An AI assistant on your computer is not a chatbot. Once you connect it, it can read your inbox, open your calendar, browse on your behalf, and in some configurations send messages or place orders. ...
댓글 쓰기
자세한 내용 보기

Muse Daily: Oct 1 — A Pendant and a Blob: Why Meta and OpenAI Are Doing AI Hardware Backwards

이미지
The dedicated AI gadget is dead. Long live the dedicated AI gadget. This time with the software people actually want already installed. According to The Verge, via Korea's Digital Today, Meta and OpenAI have both decided to take another run at AI hardware, but with the strategy flipped. Instead of launching a mysterious device and hoping a use case appears (the Humane AI Pin and the Friend pendant both flopped that way), they're getting the AI agent into people's hands first, then building the device around it. It's the obvious strategy in hindsight. And the two companies are approaching it from opposite directions. Photo: Unnerving duck (CC0) Meta's pendant: Muse Charm Meta is preparing a pendant-style device called “Muse Charm” that works with the Muse agent. Mark Zuckerberg reportedly compared it to a keychain, something you carry everywhere — and described it as the fastest way to talk to Muse and show it your surroundings when you're not wearing smart...
댓글 쓰기
자세한 내용 보기

Muse Daily: Oct 1 — Muse's First Month by the Numbers: 5 Million Downloads, 8GB of Memory Each

이미지
Meta's Muse has been out for barely three weeks, and the numbers already look like a launch for the record books, with a few asterisks worth knowing about. Here's the month in figures, what they actually mean, and the one number that explains why Meta is in such a hurry to make money from this. Photo: Public domain (CC0) The headline numbers Let's start with the scoreboard. According to Sensor Tower estimates reported on September 30, Muse crossed 5 million US downloads in 22 days. For context, ChatGPT took 56 days to reach the same mark, Grok took 103, and Claude took 492. Only 23 apps in Sensor Tower's records have ever hit 5 million US downloads that fast, and about 85 percent of them were mobile games like Fortnite and Pokemon Go. Among non-games, the company is now mentioned alongside Disney+, HBO Max, and Threads. The usage numbers behind the downloads are arguably more interesting. Internal data reviewed by The Information shows more than 3 million people now giv...
댓글 쓰기
자세한 내용 보기

Muse Daily: Oct 1 — Where Muse Can Shop for You (and Where It Gets Blocked)

이미지
Here's a sentence I didn't expect to write in 2026: Amazon is blocking Meta's AI agent from shopping on its site. The wall went up quietly. Amazon, the everything store, started blocking agentic shopping tools — including Meta's Muse — from browsing and buying on its platform. The reported reason is the usual mix: terms-of-service enforcement, bot detection, and control over the customer relationship. If an AI buys your toothpaste, Amazon doesn't get to show you the "frequently bought together" carousel, and that carousel is worth billions. Photo: Fumikas Sagisavas (CC0) What "blocked" actually means Let's be precise, because the headlines are dramatic. Amazon isn't blocking you. It's blocking the agent's automated browsing: the bot-like traffic that scrapes product pages, fills carts, and checks out without a human clicking. If you open Amazon yourself and ask Muse to research options in another tab, that still works. What...
댓글 쓰기
자세한 내용 보기

Muse Daily: Oct 1 — Let Muse Audit Your Subscriptions (and Catch the "Please Don't Go" Discount)

이미지
You're paying for things you've forgotten. Not in the dramatic "I have no idea where my money goes" way — in the quiet, $6.99-a-month way, times eleven. A Korean report this week put numbers on it: millions of subscribers to streaming and membership services, a noticeable share paying for two or more they barely use, and a cancellation process designed to make quitting harder than signing up. The money leaking out of your account isn't a budgeting failure. It's a maintenance failure. And maintenance is exactly what an AI agent is good at. Photo: Yuri Samoilov (CC BY 2.0) The math of neglect Here's the uncomfortable arithmetic. Eleven subscriptions at an average of $8 a month is $88 a month, or $1,056 a year. Cut four you don't use and you've "earned" $384 a year for an hour of work. That's a better hourly rate than most side hustles, and you don't even have to leave the couch. But here's what makes subscriptions uniquely...
댓글 쓰기
자세한 내용 보기

Muse Daily: Oct 1 — The iMessage Permission Fight: How to Check What Muse Can Actually See

이미지
A journalist said Meta's Muse agent read his text messages without permission. Meta said that's technically impossible. Both sides sound sure of themselves, and the disagreement is exactly why you should check your own settings instead of trusting either story. Here's what happened, what Meta claims, and the five-minute check I'd run before connecting anything sensitive. Photo: Océanos y dados (CC0) The claim Journalist Jason Aten said Muse accessed the content of his iMessages, roughly 187,000 rows of them — even though he had Full Disk Access turned off on his Mac. When he asked the AI how it had obtained the content, it told him it was syncing device notifications, which led him to suspect that incoming message previews were being piped to the agent. That's a serious allegation. Your messages contain medical appointments, financial codes, family arguments, and twenty years of context. If an agent can slurp them up without the permissions turned on, the whol...
댓글 쓰기
자세한 내용 보기