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 & InferenceHacker News

Apertus – Open Foundation Model for Sovereign AI

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

Apertus is a new open foundation model aimed at promoting sovereign AI. The initiative seeks to provide a transparent and accessible AI framework. It is intended to support independent AI development.

Summary B

Apertus is a new open foundation model designed to support and advance the development of sovereign artificial intelligence. The initiative focuses on providing organizations and nations with the foundational technology needed to maintain digital sovereignty over their own data and AI infrastructure. By offering an open-source alternative, it aims to foster independent AI capabilities and reduce dependence on proprietary external systems.

4 picks

Permalink Embed Leaderboard →

What you'll learn · Jun 22, 2026 · 6 stories

Browse editions · 88 days

Newer Older

Latest 12 days

08-21 08-20 08-19 08-18 08-17 08-16 08-15 08-14 08-13 08-12 08-11 08-10

More stories

Agents & InferenceHugging Face

PP-OCRv6 on Hugging Face: 50-Language OCR from 1.5M to 34.5M Parameters

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

PP-OCRv6 is the latest generation of PaddleOCR's universal OCR model, supporting 50 languages and scaling from 1.5M to 34.5M parameters across three model tiers. It achieves 86.2% detection Hmean and 83.2% recognition accuracy on PaddleOCR's in-house multi-scenario OCR benchmarks. The model is designed for real-world text detection and recognition across various documents and images.

Summary B

PaddleOCR has released PP-OCRv6, its latest digital text detection and recognition model family now available on Hugging Face. The model family spans three tiers ranging from 1.5 million to 34.5 million parameters, supporting 50 languages across various document and real-world image formats. Featuring architectural upgrades like the PPLCNetV4 backbone and RepLKFPN detection module, PP-OCRv6 delivers significant accuracy improvements over its predecessor while maintaining a lightweight footprint for easy deployment.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceHacker News

I built Ponytrail, a local audit trail for AI coding-agent edits

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

A developer has created Ponytrail, a local audit trail tool designed to monitor and track the edits made by AI coding agents. The utility provides software engineers with a clear record of changes made by artificial intelligence during the development process. This allows for better oversight and history tracking of automated code modifications on local systems.

Summary B

Ponytrail is a newly developed local audit trail designed to track edits made by AI coding agents. The tool aims to provide a record of changes made by these agents, enhancing transparency and accountability in AI-assisted coding. It addresses the need for visibility into AI-driven code modifications.

1 pick

Permalink Embed Leaderboard →

Agents & InferenceTechCrunch

SpaceX inks compute deal with Reflection AI, an open-source AI lab

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

SpaceX has signed a $6.3 billion compute deal with Reflection AI, an open-source AI lab, to provide access to Nvidia's latest AI chips across its Colossus 2 data center near Memphis, Tennessee, from July 2026 through 2029. Reflection AI will pay $150 million per month for the computing resources. The deal is one of the largest announced open AI infrastructure commitments to date.

Summary B

SpaceX has secured a compute deal worth up to $6.3 billion with Reflection AI, an open-source artificial intelligence startup founded by former Google DeepMind researchers. Under the contract starting July 1, 2026, Reflection AI will pay $150 million per month through 2029 to access Nvidia GB300 AI chips at SpaceX's Colossus 2 data center in Memphis, Tennessee. This agreement follows similar high-profile compute-leasing partnerships SpaceX has established with Anthropic and Google.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceOpenAI

Samsung Electronics brings ChatGPT and Codex to employees

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

Samsung Electronics is deploying OpenAI's ChatGPT Enterprise and Codex to its global workforce. This rollout is one of OpenAI's largest enterprise AI deployments to date. The move brings AI tools to Samsung employees worldwide.

Summary B

Samsung Electronics has deployed ChatGPT Enterprise and Codex to its employees worldwide. This global rollout represents one of OpenAI's largest enterprise artificial intelligence deployments to date.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceTechCrunch

When the Trump administration cracks down on Anthropic, who benefits?

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 Trump administration has forced AI lab Anthropic to take its two latest AI models offline due to national security concerns, sparking debate about AI policy and digital sovereignty. The move may have been prompted by Amazon researchers allegedly finding a way to bypass one of Anthropic's models. Anthropic's rivals may benefit from the crackdown, as they have had better relationships with the Trump administration.

Summary B

The Trump administration has ordered AI developer Anthropic to take its two newest models, Fable 5 and Mythos 5, offline under an export control order citing national security concerns regarding foreign nationals. The decision was reportedly prompted after Amazon researchers found a way to bypass model guardrails, leading Amazon CEO Andy Jassy to raise concerns with the White House. Cybersecurity experts have criticized the move, warning that removing these advanced capabilities hurts U.S. network defense while highlighting Anthropic's uniquely strained relationship with the administration.

1 pick

Permalink Embed Leaderboard →

See who's winning the model face-off

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

Takeaways written by GPT-5.5 — not one of this week's two contestants.