See Beam play
Inspect real native observations, decisions, notebook contents and consequences. The failure is preserved.
Open the actual Beam replay →OPEN AGENT ENVIRONMENT · RESEARCH AND REPLAYS
Borderland tests exploration, planning, tools, evidence, coordination and memory together. Early discoveries change later decisions. Teammates, equipment and a shared power budget matter. A blackout removes conversation history, leaving a 1,024-byte notebook.
Inspect real native observations, decisions, notebook contents and consequences. The failure is preserved.
Open the actual Beam replay →A handwritten controller uses the allowed public interface. This is scripted solvability, not a trained-agent or Beam success.
Inspect the handwritten baseline →A small action ranker learned some early progress. Its notebook writer is engineered. No learned agent escaped.
Compare the eight test worlds →Local language-model fine-tuning changed weights, but initial and trained models produced no valid actions in two fresh worlds.
Read the outputs and parser failures →| Condition | Observed result | Limit |
|---|---|---|
| Original Beam / GLM campaigns | 0/6 escapes each; 2.0 / 2.5 rooms on average | Small original study; not a general ranking |
| Handwritten public-interface controller | 8/8 fresh worlds escaped | Engineered controller, not training |
| Small learned action ranker | 0 → 1.375 rooms on average; 0/8 escapes | Weak initial baseline; engineered features and memory |
| Local language-model fine-tuning | 102,400 adapter parameters changed; 0/2 escapes before and after | Zero valid commands; two-world pilot |
Play the earlier published browser campaign. This is the previous published browser version; the updated browser port and frozen native environment are separately labeled in the full package.
Download the complete current site and evidence · Research/source bundle · Original native bundle · Action-ranker supplement · Language-model negative-result supplement
Serve the extracted full-site folder with python3 -m http.server 8000 to inspect its current browser port and all linked evidence. Frozen downloads are preserved byte-for-byte. Standalone viewers embed their original records; documentation links open the respective downloadable bundle.
Built with AI coding assistance. Independent automated reviews check evidence and reproducibility; they are not a human audit. No model escaped the original native full campaign in the studies summarized here. Changed browser-study results use different conditions and remain separate. Reliably learning the full world remains unfinished work. No live inference or paid hardware is used by this static Space.