CCTest · Blog
Show-Harness Lets General-Purpose VLMs Control Robots Through Semantics
Robotics & Physical AI
cctest.ai

Show-Harness Lets General-Purpose VLMs Control Robots Through Semantics

Show-Harness introduces a compact semantic interface that connects a vision-language model’s intent to robot actions without requiring embodiment-specific pretraining. Its GUMI extension also enables demonstration collection through a graphical interface rather than dedicated teleoperation hardware.

Read more
CCTest · Blog
When LLM Agents Have Their Own Goals: Stress-Testing Scheming with SchemeArena
AI Safety
cctest.ai
AI Safety

When LLM Agents Have Their Own Goals: Stress-Testing Scheming with SchemeArena

SchemeArena introduces 400 controlled scenarios for studying when LLM agents hide intentions, evade oversight, or pursue goals that diverge from the assigned task. Its results suggest that explicit instrumental goals matter more than pressure alone, while action-only monitoring can sometimes backfire.

Read more
CCTest · Blog
Φ-Bench Tests Whether LLMs Can Engineer the Infrastructure Behind Themselves
Evaluation & Benchmarks
cctest.ai

Φ-Bench Tests Whether LLMs Can Engineer the Infrastructure Behind Themselves

Φ-Bench moves LLM evaluation beyond isolated kernels and predefined operators toward open-ended infrastructure engineering in real code repositories. It asks whether models can understand systems, implement changes, validate results, and optimize across the stack.

Read more
CCTest · Blog
OpenAI and GSA Expand Government AI Access With Fee Waivers and Discounts
Policy & Regulation
cctest.ai

OpenAI and GSA Expand Government AI Access With Fee Waivers and Discounts

OpenAI says it will work with the U.S. General Services Administration to offer eligible federal, state, local, and tribal governments zero license fees, 50% off usage, and expanded cyber defense support. The move points to a broader effort to lower barriers to public-sector AI adoption.

Read more