Denoise · Twitter

The discourse is shifting from model-level jailbreaks to system-level exploits in the orchestration layer, a pattern highlighted by both @AnthropicAI and @GoogleDeepMind.

Major labs and independent researchers are focusing on the security of autonomous agents, releasing red-teaming frameworks and public vulnerability disclosures.

• 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 release, contextualized by @karpathy, signals a convergence on the terminal as the new primary interface for AI-driven development.

The release of Anthropic's terminal-native coding agent triggered widespread discussion about a fundamental shift away from traditional 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 battle to define the standard agent orchestration protocol is underway, with @OpenAI's SDK release directly competing with primitives being adopted by @LangChainAI and others.

A wave of new infrastructure for agents has been released, including competing SDKs and specialized deployment runtimes.

• 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 data release underscores that alongside the development of agentic systems, the curation of foundational datasets remains a critical path for model improvement.

A large, open web OCR dataset was released, providing a new resource for training multimodal models.

• Mistral AI@MistralAIrising

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

The ideas proposed by @GregKamradt and tools like @mem0ai suggest a move away from reliance on vector search towards multi-layered memory systems for agents.

The conversation around model memory is maturing from simple RAG to 'context engineering', addressing complex caching and retrieval issues in large context windows.

• 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.

Agent-like automation from @NotionHQ and @linear shows the agent paradigm moving from developer-facing tools to mainstream productivity applications.

SaaS workspace tools are increasingly embedding autonomous features, such as automated issue triage and database updates.

• 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 work from @weights_biases and @dspy_ai indicates a clear trend towards treating prompt engineering as a formal optimization problem rather than an intuitive art.

Efforts are focused on creating systematic, scalable methods for benchmarking and optimizing 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 highlighted by @jerryjliu0, filtering out synthetic data that harms generalization is a critical, unsolved problem in the MLOps for agents stack.

The challenge of curating high-quality synthetic data remains a key bottleneck for training effective agents.

• Jerry Liu@jerryjliu0repeated

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