Denoise · Twitter

The conversation is shifting from traditional software security to agent-specific attack vectors, as shown by @AnthropicAI and @GoogleDeepMind.

Today's focus is on red-teaming and securing autonomous agents, with major labs disclosing vulnerabilities and new frameworks.

• Anthropic@AnthropicAIrising

  Responsible disclosure on a Claude jailbreak chain we patched last week. Full write-up including our red team timeline.

• Google DeepMind@GoogleDeepMindrising

  New red team framework for prompt injection in autonomous agents. Covers cross-tool leakage, scanner evasion, and sandbox escape patterns.

• MalwareTech@MalwareTechBlogrepeated

  Autonomous agent running pentest flows against a real SaaS. First real-world run: fewer false positives than I expected on the vulnerability surface.

@AnthropicAI's Claude Code 1.5 directly competes with IDE plugins, and commentary from @karpathy suggests a convergence on the terminal as the primary AI developer interface.

Major releases center on terminal-native coding agents, with benchmarks and early adoption reports validating the shift away from IDEs.

• Anthropic@AnthropicAIrising

  Claude Code 1.5 is live. Terminal-native coding agent with full Claude Opus reasoning, file-ops sandbox, and session replay.

• Andrej Karpathy@karpathyrising

  The developer-experience shift from IDE to terminal agent is underrated. Coding workflows are about to look nothing like 2024.

• swyx@swyxrising

  Codex vs Claude Code terminal agent benchmarks. Pass@1 diverges more than I expected on the long-context editor tasks.

• DSPy@dspy_airising

  DSPy 3.0: prompt optimization via compile-time search over system prompt variations. Benchmarks inside.

• @levelsio@levelsiorising

  Switched my whole editor setup to Claude Code this week. Shipping faster than when I used Cursor + Copilot.

A pattern of convergence is visible, with @OpenAI, @vercel, and @replit all shipping agent deployment primitives, while @LangChainAI provides an interoperability layer.

The infrastructure race for hosting and orchestrating agents is heating up, with new SDKs and managed runtimes being released.

• OpenAI@OpenAIrising

  New agent SDK: protocol-level tool calling, deployment harness, and multi-worker orchestration primitives. Docs live.

• LangChain@LangChainAIrising

  MCP protocol integration thread. How to wire existing LangGraph agents into the Anthropic Model Context Protocol server spec.

• Vercel@vercelrising

  Edge runtime for agent workers is live. Spawn durable background agents from any serverless deployment.

• Alex Albert@AlexAlbert__rising

  When your security scanner finds nothing scary on an agent deploy, check the orchestration layer again. That's usually where the jailbreak sneaks through.

• Replit@replitrising

  New agent deployment harness. One command to go from local orchestration to hosted agent worker.

@MistralAI's dataset release indicates a continued investment in foundational data for multimodal models, even as the application layer focuses elsewhere.

A large-scale OCR dataset was released by a major European AI lab to support multimodal model training.

• Mistral AI@MistralAIrising

  Open dataset release: 100M-row web OCR dataset. Cleaned, licensed, ready to train.

Thinkers like @GregKamradt are moving beyond simple RAG, while startups like @mem0ai are building specialized memory components, fragmenting the simple vector search pattern.

The conversation around RAG is evolving into a more sophisticated discussion of 'context engineering' and multi-layered agent memory systems.

• Vaibhav Srivastav@reach_vbrising

  Tested the new 10M context memory window end to end. Surprising failure modes around rag retrieval cache invalidation, thread below.

• Greg Kamradt@GregKamradtrising

  RAG is dead, long live context engineering. My framework for when to cache, when to retrieve, and when to just dump memory into the prompt.

• mem0@mem0airising

  Memory layer for agents: differentiating working memory from the subconscious store. Vector index isn't enough anymore.

• LlamaIndex@llamaindexrepeated

  Knowledge graph retrieval walkthrough: when semantic vector search misses, graph hop beats it every time.

Tools like @NotionHQ and @linear are integrating agent-like automation, suggesting a broader trend of 'workspace agents' beyond developer-focused tools.

General workspace automation tools are adding AI-powered features for auto-triaging and data entry.

• Notion@NotionHQrising

  Notion workspace automation is out of beta. Auto-fill tables, chained updates across databases, and a new audit log surface.

• Linear@linearrising

  Linear now auto-triages incoming issues. Quiet launch, but already our favorite workspace feature of the year.

• Temporal@temporaliorepeated

  Orchestrating agents with durable workflows: replayable, resumable, and multi-worker by default. Walkthrough from our infra team.

• James Clear@jamesclearrepeated

  The best habit tracker is the one you actually open. Three open-source alternatives worth trying.

The practice of prompt engineering is maturing from anecdotal 'tricks' (@dotey) to industrial-scale, data-driven optimization, as exemplified by @weights_biases.

The focus is on systematic, large-scale benchmarking of prompts to discover non-obvious performance patterns.

• dotey@doteyrising

  Five prompt tricks learned this week from reviewing 200 production prompts. Short thread.

• Weights & Biases@weights_biasesrising

  System prompt benchmarking at scale: we ran 40k variants across 6 frontier models. The efficient frontier is not where you think.

@jerryjliu0's tweet points to data curation as a key, unsolved problem for scaling reliable agent behavior, a layer below the application discussions.

A specific challenge in agent training is highlighted: filtering synthetic data to avoid performance degradation.

• Jerry Liu@jerryjliu0repeated

  Dataset curation for agent training: how we filter synthetic data that looks good but poisons generalization.