Call My A.I. Agent

What the episode covered

Eli Tan said Meta's Muse AI agent can use accounts and payment tools to complete real-world tasks, and he demonstrated it by ordering lunch. The register is mixed: the demo is concrete, while the company vision is described in polished terms and the transcript has an apparent mid-sentence jump. Watch whether adoption turns into durable use, and whether account-access controls become clear.

Our summary, written from the transcript. Not the speakers' words.

The episode, in order

1. 01Live task demonstration

   The example made the agent's usefulness concrete while also showing how much account access the tool needs to act.

2. 02What makes agents different

   The episode framed agents as a step beyond chatbots because they can take actions rather than only answer questions.

3. 03Meta's product push

   That frames the business race as consumer adoption, not only technical performance.

4. 04Human tasks being outsourced

   The concern is not only safety. It is also whether convenience replaces ordinary human contact and judgment.

5. 05Optimistic and skeptical futures

   The episode leaves the central bargain unresolved: more convenience in exchange for more access and reliance.

Sponsors named in this episode

No sponsor named doesn't mean no ads. Each row was taken from the ad break as it was read on air. The verbatim passage behind every row is checked against the transcript character for character. Short excerpts appear below.

Roles and stated affiliations

Stated affiliation. Eli Tan: He is the reporter who tested Muse with his own accounts, payment tools, and daily tasks, so his stake in the recording is experiential reporting and evaluation.

Wakewire's read of interests

• Natalie Kitroeff: She frames the segment for listeners, presses on privacy, trust, and adoption, and participates in the live lunch-order demonstration.
• Meta: The company is described as trying to turn agents into a consumer product that many people can use, with possible advantages from Facebook, Instagram, and WhatsApp.
• OpenAI: OpenAI is grouped with Meta in Tan's description of companies marketing agents as approachable and personal.
• The New York Times: The recording includes promotion for other Times audio and subscriptions, and it carries Genentech advertising.
• Genentech: Genentech is named in commercial segments promoting The Impossible Bet and its scientific breakthroughs.

Recording limits

• The transcript appears to jump mid-sentence from "Yes. Three and a half" to Genentech copy, so part of the Meta scale discussion may be missing.

Claims that need more support

• Natalie Kitroeff said "millions of people have rushed to download" the apps, but no download figure or source was provided.

Claims made on air that the episode itself did not back up.

Questions the conversation didn't reach

• The episode raised but did not answer what permissions, data storage, or safety controls apply when Muse accesses accounts and payment tools.

Topics the episode did not address. This does not mean anyone avoided them.

Are you from this show? You can ask for a correction, reply, or ask us to stop coverage on the rights page.

What was said

• Live task demonstration

  Tan asked Muse to order Natalie Kitroeff a Sweetgreen harvest bowl using his saved payment and DoorDash account, and the order arrived during the segment.

• What makes agents different

  Tan said an AI agent is like a chatbot with its own computer that can operate software, sign into accounts, and work autonomously in the background.

• Meta's product push

  Tan said Mark Zuckerberg wanted AI products that even his mother could use and envisioned a personal assistant for finances, health, and personal life.

• Human tasks being outsourced

  Tan said Muse called his bookstore even though he enjoys talking to the bookseller and could have walked there.

• Optimistic and skeptical futures

  Tan said companies argue agents can return time to users, but he said skepticism is fair given the track records of products such as Instagram Reels.

Our paraphrase, not the speakers' words.

The passages we relied on

Short excerpts, credited to the show, at most 40 words each and 150 per episode. We do not publish transcripts. We link to the episode so you can hear it from the show itself.

What to watch for

• Meta or other agent makers publish clear controls for what accounts an agent can access, what data is stored, and how permissions can be revoked. (Next product updates and public documentation.)
• Adoption numbers are paired with repeat-use or retention figures rather than only download claims. (Next earnings cycle, product announcements, or reporting follow-up.)
• Reported incidents, audits, or company statements describe how agents handle mistakes, unauthorized actions, or account misuse. (As public deployments expand.)