Urgent.News

What's breaking now, across thousands of outlets.

AI

Google Tabletop AI: Build Custom Collaboration Apps by Describing Them

Google Tabletop AI: Describe the App, Skip the Spreadsheet Google just admitted something the productivity industry has been dancing around for years: nobody actually wants to configure a collaboration tool. They want to describe what they need and have it exist. Tabletop, the new Google Labs experiment, is exactly that — a shared space where you type what you want and an AI builds the tool. The…

Google has unveiled Tabletop, an experimental AI tool that allows users to describe their desired collaboration app and have it created by the AI. The service, currently available only to adults in one country via a waitlist, functions like a shared space where users can drop in links, files and ideas and prompt "@tabletop" to generate a custom interactive tool.

This could range from a trip itinerary or dinner poll to an expense splitter or tournament bracket. Users simply describe the tool they need, and the AI constructs a working artefact that all members can use simultaneously, similar to a Google Doc. The aim is to simplify the coordination of small tasks, such as family trips, rec leagues, or potlucks, rather than enterprise workflows.

While the examples are deliberately small, the idea is to shift AI from answering questions to building tools. The article suggests that Tabletop could be a compelling alternative to traditional workplace software, offering a more informal and user-friendly approach for groups that wouldn't typically use software like Asana or Monday for informal coordination.

However, the article also notes that Tabletop is currently an experiment, not a product, with no pricing, launch date, or expansion timeline disclosed. The age limit for access is 18+, which may be lifted after testing. The takeaway is that Tabletop represents a significant shift in how AI tools are developed and used, moving them out of chat windows and into shared spaces where groups actually accomplish their tasks.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

How to review an AI-written pull request: a 10-point checklist

AI-written pull requests have a particular look. The code is tidy, the names are sensible, the description is confident, and the tests are green.

  • Read the task description to understand original intent
  • Compare diff against task to avoid extra changes
  • Examine tests to validate new functionality

Is that really GPTBot? How to verify AI crawlers in your logs

Open any access log today and you'll see GPTBot, ClaudeBot, PerplexityBot and friends. The problem: a user agent is just a string.

  • Compare bot's IP address against OpenAI's published CIDR ranges
  • Use DNS reverse and forward lookups for Googlebot and Bingbot
  • Consult official bot documentation for accurate verification methods

Query fan-out: why AI search never answers the question you typed

When you ask an AI assistant with web search something like "best CRM for a 5-person agency", it usually doesn't run that exact search.

  • AI search breaks down queries into smaller parts
  • Generates answer from amalgamation of results
  • Focus on concise sections answering sub-questions

More from Friday 13 November →