Which AI writes the better take? You decide — blind.

Two top models go head-to-head on today's AI news. Pick the sharper summary without seeing the names — the crowd's verdict builds the leaderboard.

Agents & InferenceSimon Willison

datasette-agent-edit 0.1a0

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Simon Willison is developing plugins for Datasette Agent to enable collaborative text editing, including features like Markdown, SQL queries, and SVG files. He introduced the base plugin *datasette-agent-edit* to streamline core editing tools for reuse across other plugins. The project was announced on June 7, 2026, with sponsorship options available for updates.

Summary B

A new base plugin called datasette-agent-edit has been released to provide core text-editing tools for Datasette Agent. It aims to support upcoming plugins for tasks like collaborative Markdown editing, updating large SQL queries, and editing SVG files, drawing on the design of Claude's text editor tools. The approach centralizes these editing patterns so they can be adapted across multiple plugins rather than rebuilt each time.

0 picks

Permalink Embed Leaderboard →

Browse editions · 59 days

Newer Older

Latest 12 days

07-23 07-22 07-21 07-20 07-19 07-18 07-17 07-16 07-15 07-14 07-13 07-12

More stories

Agents & InferenceHugging Face

Designing the hf CLI as an agent-optimized way to work with the Hub

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

The Hugging Face hf CLI tool has been redesigned to better serve both human users and AI coding agents, optimizing outputs differently for each. AI agents receive compact, structured responses to minimize token usage, while humans get more visually rich terminal outputs. Early data shows significant adoption by coding agents like Claude Code and Codex, with the CLI proving up to 6 times more efficient than manual approaches for complex tasks.

Summary B

Hugging Face has redesigned its official hf command-line tool to serve both human users and AI coding agents like Claude Code, Codex, and Cursor, automatically detecting when an agent is driving it and adjusting output accordingly. The CLI now tags and tracks agent traffic, which has grown significantly since tracking began in April 2026, with Claude Code alone accounting for roughly 40,000 users and nearly 49 million requests. Benchmarking showed that on complex, multi-step tasks, agents using non-CLI methods like curl or the Python SDK consumed up to six times as many tokens as those using the optimized hf CLI.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceTechCrunch

Google will pay SpaceX $920M per month for compute

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Google has agreed to pay SpaceX $920 million per month from October 2026 through June 2029 for access to roughly 110,000 NVIDIA GPUs and related compute, citing higher-than-expected demand for its Gemini Enterprise agent platform. The deal, disclosed in a regulatory filing, resembles a similar arrangement SpaceX struck with Anthropic and includes a cancellation clause allowing either party to exit with 90 days' notice after December 31, 2026. The agreement comes just a week before SpaceX is expected to begin trading on the Nasdaq, with the company aiming to raise around $75 billion at a valuation near $1.75 trillion, which would be the largest IPO in history.

Summary B

Google will pay SpaceX $920 million per month from October 2026 through June 2029 for access to around 110,000 NVIDIA GPUs and related compute resources. The deal, similar to SpaceX's earlier agreement with Anthropic, allows Google to expand its AI capacity amid surging demand for products like Gemini Enterprise. Both parties can terminate the agreement with 90 days' notice after December 2026, and Google's access will ramp up gradually by September 2026.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceSimon Willison

datasette-agent-micropython 0.1a0

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

An early alpha release, datasette-agent-micropython 0.1a0, aims to let Datasette Agent safely generate and execute Python code within a secure sandbox. Initial testing appears promising, with GPT-5.5 so far unable to break out of the sandbox.

Summary B

Simon Willison released an alpha version of Datasette Agent, designed to safely generate and execute Python code. Early tests show promise, with GPT-5.5 unable to break the sandbox. The project aims to enhance secure code execution capabilities.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceHugging Face

Adding MCP Tools to Reachy Mini

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Reachy Mini's conversation app now supports tools hosted on Hugging Face Spaces via MCP, allowing users to add new capabilities like weather checks without modifying the app directly. These tools run remotely in the Space, eliminating the need for local code downloads while enabling users to publish and share their own tools. The update includes built-in tools for robot functions while expanding possibilities for external integrations like web searches or data lookups.

Summary B

Reachy Mini's conversation app can now use external tools hosted in public Hugging Face Spaces and accessed via the MCP protocol, allowing users to expand the robot's abilities—such as checking weather or searching the web—without editing the app's code. The tools run remotely in their Spaces rather than downloading to a user's machine, and developers can publish their own tools for others to use. This adds a new remote category alongside the robot's existing built-in and custom local tools, which are controlled through profiles that determine what capabilities are active.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceTechCrunch

The token bill comes due: Inside the industry scramble to manage AI’s runaway costs

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Companies across the tech industry are grappling with skyrocketing costs tied to AI token consumption, with budgets often exceeded months before planned. Firms like Uber, Microsoft, and Priceline face significant financial strain due to increased AI adoption and the push for more autonomous tools, despite falling per-token prices. To address this, industry players are forming standards bodies and tools to help businesses better manage and audit AI-related expenses.

Summary B

Companies across the tech industry are reeling from soaring AI costs, with token consumption surging despite falling per-token prices as firms adopt increasingly autonomous agents. Uber exhausted its 2026 AI coding budget by April, Microsoft revoked developers' Claude Code licenses, and some companies reportedly racked up bills in the hundreds of millions after failing to set usage limits. In response, a market for cost-control tools is emerging, and the Linux Foundation has unveiled plans for a new Tokenomics Foundation to bring spending discipline to AI usage, much as FinOps did for cloud costs.

0 picks

Permalink Embed Leaderboard →

See who's winning the model face-off

Tomorrow's blind matchup and the running leaderboard — one email a day.