Denoise · Twitter

The conversation is maturing from prompt injection to orchestration-level exploits, with @AnthropicAI and @GoogleDeepMind defining this new security frontier.

The focus is on new attack surfaces created by autonomous agents, with major players publishing red-teaming 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.

The primary competition is now between terminal agents (Anthropic's Claude Code) and IDE extensions (Copilot), a convergence pattern validated by @karpathy and @swyx.

Major releases from Anthropic and DSPy signal a shift towards terminal-native, autonomous coding agents with programmatic optimization.

• 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.

OpenAI, Vercel, and Replit are converging on a similar stack for agent deployment, while @LangChainAI adapts to integrate with emerging protocols.

A wave of new SDKs and platforms from OpenAI, Vercel, and Replit are being released to support agent deployment and orchestration.

• 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 contribution of a large-scale OCR dataset indicates continued investment in foundational data for multimodal models, even as agent narratives dominate.

A single, significant data release from Mistral AI provides a large-scale, cleaned dataset for web OCR model training.

• Mistral AI@MistralAIrising

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

Thought leaders like @GregKamradt and projects like @mem0ai are converging on the idea that vector search alone is insufficient for robust agent memory.

The discourse is evolving from simple RAG to more sophisticated 'context engineering' and structured memory architectures for agents.

• 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.

The auto-triaging in @linear and database automation in @NotionHQ reveal a trend of embedding agent-like capabilities directly into established SaaS workflows.

Workspace automation tools from Notion and Linear are launching features that mirror the autonomous capabilities of AI agents.

• 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.

Efforts by @weights_biases to benchmark prompts at scale signal a move towards industrializing prompt engineering as a systematic optimization problem.

The focus is shifting from anecdotal prompt 'tricks' to scalable, systematic benchmarking of system prompts for frontier models.

• 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 highlights a critical bottleneck: filtering synthetic data to avoid 'poisoning' agent generalization, which is crucial for building more capable models.

The conversation centers on the nuanced challenges of curating high-quality synthetic data for training effective agents.

• Jerry Liu@jerryjliu0repeated

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