… AI could help create a dramatically better future, but that outcome is not guaranteed. The world’s leading AI companies believe they could be close to automating AI research. It is hard to predict exactly how much this will accelerate AI progress, but there is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.
To realize AI’s potential, industry, government, and society at large may need the option to buy time to address emerging risks, develop security measures, and strengthen oversight. But each company—and country—is under intense competitive pressure not to unilaterally slow that acceleration. And today, the world lacks the technical and governance tools to deliberately pace frontier-wide progress. — Read More
Monthly Archives: July 2026
Comparing Open-Source AI Code Security Harnesses
There’s a huge new wave of open-source AI security projects that point a large language model at a codebase to find vulnerabilities. We looked at some popular (or less popular, but from interesting companies) OSS tools and compared how they work, and when you’d reach for each.
We find three complementary categories in the new world of harnesses: LLM-led exploitgen, LLM-skill-boosting, and SAST+LLM hybrids. — Read More
The Arguments Against Open Source AI are Very Bad
The release of Kimi K3 has opened a fresh round of angst and confused discourse. There’s a loud cohort of journalists, business leaders, and politicians arguing that open source AI is a dangerous threat.
Freely available AI for anyone? The horror!
Frontier labs’ case against open source AI is essentially: Open source models1 are dangerous (and un-American!). We should open the AI Pandora’s Box, but only with responsible gatekeepers (toll collectors, preferably us!). Only trusted users (our most profitable customers) should be able to use it. — Read More
How ChatGPT Optimizes its Agent Loop: Harness, API, and Inference
AI labs are moving faster than ever and releasing the most capable models we have ever seen. … But capability is only half of the picture. The other half is how much it costs these models to complete tasks, that is, the cost per successful task. Lower cost makes the model more affordable for users and less costly for the provider. A huge amount of effort inside the labs goes into making every component and layer of their AI applications optimized and more efficient, to reduce the overall cost. For example, GPT 5.6 Sol with max reasoning scores higher than Fable 5 on the Artificial Analysis Coding Agent Index while costing less than half as much.
To understand what techniques are adopted in frontier labs to make AI applications more efficient, we met with the OpenAI engineers who developed and shipped various efficiency techniques into the systems behind Codex and ChatGPT Work. — Read More
Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
A companion technical writeup to our incident disclosure. This post walks through how the intrusion actually worked: the two initial-access vectors, how the agent pivoted and moved laterally, representative examples of the commands that were run and how we investigated with GLM 5.2 (an open-source model). Live credentials, internal hostnames, and specific indicators have been redacted or genericized, while the techniques are described exactly as observed by Hugging Face.
We are publishing this level of detail because the technique matters more than the incident, as it reveals the emerging attack capabilities of the frontier agents, how they could be used by rogue actors, and how everyone should be prepared as defenders. — Read More
Agent Governance Toolkit
Your AI agents call tools, browse the web, query databases, and delegate to other agents. Once deployed, they make decisions autonomously.
… AGT does not try to win that fight inside the prompt. Every tool call, message send, and delegation is intercepted in deterministic application code before the model’s intent reaches the wire. Actions the AGT kernel denies are not “unlikely.” They are structurally impossible. That is the difference between asking an agent to behave and making it incapable of misbehaving. — Read More
The Orchestrator’s Tax
Subagents get justified by time saved and parallel execution, but that’s not what matters most in long-running multi-agent work. Every token in the orchestrator’s context is competing for its attention, and the real value of a subagent is what it keeps out of that context, not how fast it runs. I argue that subagents should be treated as a tool for protecting the orchestrator’s working memory, offloading reasoning it doesn’t need to hold onto. Doing this well means giving the orchestrator explicit ground rules for when and how to delegate. This is exploratory work, built from one real incident, and it ends with more open questions than settled answers. — Read More
#architectureData Integration vs Workflow Orchestration: Connecting Systems Is Not Coordinating the Work
Data integration connects systems and moves data between them. Workflow orchestration coordinates what runs, in what order, and what happens when a step fails. The data integration vs workflow orchestration split is simple to state: connecting systems is not the same as coordinating the work that runs across them.
… The connectors serve different jobs. Integration uses them to move and reshape data. Orchestration uses them to trigger and coordinate tasks, then to recover when one fails. Same plugin to one system, two different reasons to call it.
This post draws the line between data integration and workflow orchestration, shows where the two overlap, and explains why one control plane across categories is becoming the way enterprises manage both. — Read More
Where Should Your Company’s AI Brain Live?
I have been working with a lot of companies that are trying to become AI-first.
At first, that usually means helping individuals get much better at using AI.
Soon the AI knows that when someone says MRR, they do not just mean the generic finance definition.
… It is no longer just automation. It is institutional memory – the company’s knowledge, history, processes, workflows, and way it uniquely does work.
And it likely will be the new backbone and most significant software infrastructure in the AI era. … AI company brains will become the next version of indispensable software. — Read More
The Data Pyramid in Robotics
Frontier LLMs were pre-trained on trillions of tokens of text, most of it at least initially taken from the web. The internet was, in effect, a free pre-training corpus that humanity had collectively assembled.
In robotics, there’s no such equivalent. That generally means all attempts to solve and get to physical AGI in some form involve working through this idea of the data bottleneck. In this piece, I’ll go deeper into the bottleneck, and introduce a version of the robot data pyramid highlighting the various kinds of data used in robot learning that can be combined to help address it. — Read More
Rick's Cafe AI 