Using AI for Force Multiplication | Toronto .

Members-Only

Recent Talks & Demos are for members only

Exclusive feed

You must be an AI Tinkerers active member to view these talks and demos.

December 03, 2025 · Toronto

Control+S: AI GRC Sandbox

See a technical deep dive into Control+S, an AI agent sandbox tool for evidence-first GRC analysis against frameworks like ISO 27001 and NIST CSF.

Overview
Links
Tech stack
  • GPT-4
    GPT-4 is OpenAI’s large multimodal model: it processes both text and image inputs, delivering human-level performance on complex professional and academic benchmarks.
    This is OpenAI’s latest milestone in scaling deep learning: a large multimodal model accepting both text and image inputs. It demonstrates a significant capability leap over its predecessor, scoring in the top 10% on a simulated bar exam (GPT-3.5 scored in the bottom 10%). The model handles nuanced instructions and long-form content, supporting context windows up to 32,768 tokens (32K model). This capacity allows processing up to 25,000 words in a single, complex prompt. GPT-4 is engineered for enhanced reliability, steerability, and advanced reasoning across diverse tasks.
  • Claude Agent SDK
    The Claude Agent SDK (Python, TypeScript) transforms Claude into an autonomous, production-ready agent by granting it controlled access to a computer: file operations, Bash execution, and external tool integration via the Model Context Protocol (MCP).
    This SDK provides the core agent harness that powers Claude Code, enabling developers to build sophisticated, general-purpose agents. It offers official TypeScript and Python libraries for seamless integration. Agents gain powerful, safe capabilities: reading and writing files, executing sandboxed Bash commands, and leveraging a rich tool ecosystem. The framework includes essential production features (e.g., automatic context management, granular permission modes like `acceptEdits`, and built-in cost tracking). Use the open Model Context Protocol (MCP) to connect custom tools, databases, and APIs, extending Claude’s utility far beyond simple chat to complex, multi-step workflows.
  • GLM
    An open-source bilingual language model using an autoregressive blank-filling objective to unify NLU and text generation.
    GLM (General Language Model) bridges the gap between BERT-style understanding and GPT-style generation. Developed by the Knowledge Engineering Group at Tsinghua University, this architecture uses a unique blank-filling objective to master diverse NLP tasks. The GLM-130B variant features 130 billion parameters and supports 2D positional encoding, delivering high-performance results on both English and Chinese benchmarks. By optimizing for zero-shot transfer and efficient fine-tuning, GLM provides a versatile alternative to proprietary LLMs for developers and researchers.
  • Convex
    Full-stack TypeScript platform: your reactive database and backend, always in sync.
    Convex is the full-stack development platform that eliminates backend complexity. It replaces traditional stacks (database, servers, WebSockets) with three core building blocks: Mutations (atomic writes), Queries (live, reactive subscriptions), and Actions (serverless functions for third-party APIs like OpenAI). You define your entire backend, from database schemas to authentication, in pure TypeScript. This 'everything-is-code' approach ensures end-to-end type safety and guarantees real-time synchronization: when data changes in the backend, your frontend updates automatically, simplifying state management significantly.
  • E2B Sandboxes
    E2B (Execute to Build) delivers a secure, scalable cloud environment for running untrusted AI code: We use Firecracker microVMs to provide hardware-level isolation, ensuring every sandbox is fully contained
    E2B (Execute to Build) delivers a secure, scalable cloud environment for running untrusted AI code: We use Firecracker microVMs to provide hardware-level isolation, ensuring every sandbox is fully contained. This architecture achieves lightning-fast startup, typically under 150 milliseconds, which is critical for real-time AI agents and data analysis workflows. The platform supports multi-language execution (Python, JavaScript, etc.) and is leveraged by major players like Perplexity for advanced data analysis and Groq for Compound AI Systems. E2B abstracts complex infrastructure, allowing developers to focus solely on agentic logic via simple Python or JavaScript SDKs.

Related projects