Denoise · Twitter

The security focus is shifting from the model itself to the orchestration layer where agents interact with tools, as noted by MalwareTechBlog and AlexAlbert__.

Major labs like Anthropic and Google DeepMind are releasing formal red-teaming frameworks and disclosures for AI agent vulnerabilities.

• 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 competition in AI coding is moving from IDE autocompletion (Copilot) to stateful, terminal-based agents (Claude Code), with DSPy suggesting agent behavior may soon be 'compiled'.

Anthropic launched Claude Code 1.5, a terminal-native agent, prompting immediate benchmarks and discussion about a workflow 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 consensus is forming around a dedicated agent orchestration layer, with OpenAI's SDK and LangChain's protocol work signaling a move toward standardized agent communication.

OpenAI, Vercel, and Replit released new infrastructure for deploying and orchestrating agents, including SDKs, edge runtimes, and deployment harnesses.

• 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.

The release of a foundational dataset by a major player like MistralAI indicates a strategic push to accelerate progress in visual and text understanding models.

MistralAI released a large-scale, cleaned web OCR dataset to facilitate training for multimodal models.

• Mistral AI@MistralAIrising

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

The limits of basic vector search are pushing frameworks like Mem0ai and LlamaIndex towards structured memory (working vs. subconscious) and knowledge graph retrieval.

Discussions are moving beyond simple RAG to more complex "context engineering," exploring cache invalidation in large contexts and new agent memory architectures.

• 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 'agentification' of SaaS is happening in parallel with developer agents, as tools like Notion and Linear embed autonomous capabilities directly into user workflows.

Workspace automation is a key theme, with Notion and Linear launching features for auto-filling data and auto-triaging issues.

• 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 engineering is becoming more data-driven, with platforms like Weights & Biases enabling systematic optimization over thousands of variants, much like hyperparameter tuning.

Practitioners are sharing empirical results from large-scale system prompt benchmarking and reviews of production 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 capabilities grow, the bottleneck is shifting from model architecture to high-quality, poison-free data curation—a classic ML problem applied to a new domain.

The focus in agent training data is on sophisticated filtering techniques for synthetic data to prevent generalization failures.

• Jerry Liu@jerryjliu0repeated

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