Loading page…
Dan Shipper and the team will demonstrate two projects and explain when they choose a faster model. Attendance requires an Every subscription.
Listen to this story
Every’s one-hour camp will pair two Claude Fable 5.2 project demos with a practical look at overnight agent work, including how to define a finish condition and decide whether the model’s token costs and slower speed are justified. The session offers a view into the company’s current workflow, not an independent benchmark of the model. Attendance requires an Every subscription, so access costs $30 monthly or $288 annually.
The virtual session runs November 13, 2026, from noon to 1 p.m. ET; attendees reserve access through Every’s event page.
Examples include Everycut, an AI-agent video editor built by Kieran Klaassen, and a racing game Jack Cheng made from a playroom photo.
Every says Fable 5.2 is slow and token-intensive, and plans to explain when it chooses the model over faster options.
Every will hold its Claude Fable 5.2 virtual camp on November 13, 2026, from 12:00 to 1:00 p.m. ET. Dan Shipper and the team will demonstrate projects and explain overnight workflows. To attend, reserve a place through Every’s official event page; access is included with an Every subscription.
The session is titled “How We’re Working Right Now: Claude Fable 5.2.” Every lists the full noon-to-1 p.m. slot as the virtual camp, with no separate keynote or doors-opening time. The registration page is the official starting point for access; a direct public stream link is not provided.
The reservation form says attendees will use their email address to access the event. It offers monthly and annual Every subscriptions rather than listing a separate admission charge. The page also invites participants to bring questions, making audience questions part of the planned session.
Every’s confirmed agenda centers on work its team says it has already built with Fable 5.2. Shipper and colleagues plan to walk through two examples, covering both a video-editing tool and a game made from a photograph:
Beyond the project demonstrations, the team promises to show how to set up a job that Fable 5.2 can work on overnight. The agenda specifically includes telling the model how to recognize that a job is finished—not just giving it an assignment and leaving it running.
Cost is also an explicit part of the session. Every describes Fable 5.2 as slow and heavy on token use, and says the team will explain when it chooses this model and when it uses something faster. That is the organizer’s assessment of its workflow, not an independent performance comparison.
Loading discussion...
Join the conversation
Explain what would make you comfortable handing over that judgment.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.