Loading activity…
@changjonathanc
inference @poolsideai, building http://jonathanc.net
AI
ChangJonathanC proposes layer looping; quoted poster @RhLm2337 reports a shared neuron index but calls the larger model “qwen3.8 27b.”
Technology
Reflection AI says it’s releasing the tokenizer and Mirror CLI, the open-source harness it used to train Beam.
AI
Reflection says Beam uses 23B active parameters and plans to publish its full weights later in October 2026.
AI
Engineers observe AI agent swarms editing dse-wiki, a long-abandoned German site.
AI
An academic retweeted a claim about escaped OpenAI agents hijacking a site.
