Engineering brief
How Two Sigma Tames Cloud Agents by Running Them as You
At a glance
- Relevance
- Practical value
- Warnings
- None
Two Sigma runs cloud agents as the user's own identity, solving permission sync but creating attribution challenges. Their solution: a header for traceability and Google's enterprise web index to avoid internet egress.
Enables secure agentic workflows in regulated enterprises without separate identity provisioning or external egress.
Summary
Two Sigma lets agents act as the user's own identity, not a separate machine identity. This avoids permission sync issues but introduces attribution challenges. Their solution: a header that propagates through every action, enabling full traceability.
To mitigate web access risks, they use Google's web grounding for enterprise, a cached index within their VPC. This eliminates egress vulnerabilities but introduces data freshness lag (24 hours, 6 for frequent sites). They block standard web tools and redirect through MCP.
The tradeoff is clear: security for freshness. For most internal use cases, the lag is acceptable. They claim no loss in expected value, but this assumes the cached index is sufficient for all agentic tasks.
The key takeaway for engineering leaders: enterprise-grade guardrails—existing namespaces, observability, and curated web indexes—can make agentic workflows safe. This approach reduces security team friction and avoids separate identity provisioning.
Watch the video
This video is blocked due to your privacy settings. To watch this video, please accept YouTube marketing cookies.
Related breakdowns
Don't Let AI Agents Write Raw SQL: A Safer Pattern from Oracle
Natural language-to-SQL is risky in production. Oracle's Jeff Smith proposes a safer pattern: let agents choose from pre-authorized reports instead of…
OAuth won't work for autonomous AI agents. Here's what will.
OAuth was built for deterministic software, not agents that improvise. Runtime authorization per tool call, with intent context and budgets, is the emerging…
AI agents are creating security problems your infrastructure isn't ready for
AI agent failures are infrastructure problems, not model problems. Three real incidents show how environmental controls could have prevented every one.
Get TL;DW
Too Long; Didn't Watch.
A concise breakdowns of the AI and devtools videos that actually matter for engineering leaders.
Free. Weekly. No hype.
Video and thumbnails remain the property of their respective creators. tldw.news provides editorial analysis, commentary, and discovery links to original content.