@neural_avb: Up next: Tool Calling inside RLMs Tools are not called during generation, instead they are python functions injected di…
Summary
fast-rlm v1.14 introduces direct Python function injection for tool calling, accepts any Python dict as input, guarantees structured output via schema declaration, and allows agents to call subagents.
View Cached Full Text
Cached at: 05/22/26, 01:41 AM
Up next: Tool Calling inside RLMs
Tools are not called during generation, instead they are python functions injected directly into the REPL.
The RLM can call them inside the code and store results in a variable within the python session (instead of directly loading into context https://t.co/Ntrrp75ADw
AVB (@neural_avb): New version of fast-rlm out today (v1.14)
New features in this release:
- Input to RLM need not be string, can be any python dictionary
- Output schema declaration -> RLM is guaranteed to return output in your designed structured output
- Agents can call subagent with explicit
Similar Articles
Python 3.15.0 candidate 2 is here!
Hugo van Kemenade announces the final release candidate for Python 3.15, scheduled for October, and encourages third-party maintainers to prepare their projects by testing and publishing wheels.
Most LLM features ship without the engineering discipline we'd never skip for regular software
The article highlights the lack of engineering rigor in shipping LLM features and promotes a masterclass on September 12 that teaches disciplined evaluation, testing, and production methods.
Introducing wrapture
Introducing wrapture, a new Python library for monkeypatching, testing, and tracing with OpenTelemetry support, entirely AI-assisted in its development.
CAST: Critique-Aware Supervision for Training Reliable Long-Horizon Tool-Calling Agents
CAST is a critique-aware training framework that improves LLM agent reliability by generating structured action-level rationales from sparse outcomes, outperforming GPT-OSS-120B on tool-calling benchmarks.
Scaffolding Foundation Models into Physical-World Agents Pushes the Frontier of Long-Horizon Navigation
NavMCP is an agentic scaffolding framework that integrates vision-language models with navigation foundation models to enable persistent long-horizon physical-world exploration, achieving state-of-the-art results on benchmarks and real-robot tasks.