Denoise · Twitter

The frontier of security research is moving up the stack from the LLM itself to the agent's orchestration layer, as shown by red-teaming frameworks from Google DeepMind and disclosures from Anthropic.

Security discourse is focused on the new attack surface created by autonomous agents, including jailbreaks, tool leakage, and sandbox escapes.

• 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.

A clear schism is forming between incumbent IDE-based tools like Copilot and the new wave of terminal-native agents from actors like Anthropic, with developers like @levelsio reporting faster shipping on the new platform.

Anthropic's release of Claude Code 1.5, a terminal-native agent, dominates the conversation, with benchmarks and early user feedback emerging.

• 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.

The race is on to provide the 'Heroku for Agents', with Vercel and Replit offering serverless runtimes that compete directly for deploying agents built with frameworks from OpenAI and LangChain.

Major infrastructure providers including OpenAI, Vercel, and Replit are shipping primitives for building, orchestrating, and deploying agents.

• 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.

While agents dominate the application layer, foundational dataset work continues to be a key differentiator, with MistralAI continuing its strategy of releasing high-quality open data.

MistralAI released a large-scale, cleaned web OCR dataset, providing a significant public resource for training multimodal models.

• Mistral AI@MistralAIrising

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

The limits of vector search are pushing developers like @GregKamradt and @mem0ai towards new memory architectures that differentiate between working and long-term storage, a critical component for agent autonomy.

Discussions are shifting from basic RAG towards more complex 'context engineering' to handle large memory windows and stateful agent operation.

• 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 agentic paradigm is not confined to developer tools; business-facing SaaS platforms like Notion and Linear are converging on the same pattern of autonomous task completion.

SaaS workspace tools like Notion and Linear are shipping agent-like automation features that triage and execute workflows for users.

• 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.

Prompt optimization is becoming a data science problem, with organizations like Weights & Biases running massive experiments that treat prompt variants as a hyperparameter to be tuned.

The focus in prompt engineering is shifting from anecdotal tricks to systematic, large-scale benchmarking to find optimal system prompts.

• 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.

As agent training becomes more accessible, the bottleneck shifts from compute to high-quality data, with practitioners like @jerryjliu0 highlighting the need for sophisticated filtering to avoid generalization failure.

The conversation centers on the practical challenges of dataset curation for training agents, particularly filtering harmful synthetic data.

• Jerry Liu@jerryjliu0repeated

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