The Big Story
Cisco AI Introduces FAPO: Pipeline-Aware Prompt Optimization With Step-Level Failure Attribution and Claude Code Orchestration.
According to MarkTechPost, Cisco Foundation AI has open-sourced FAPO (Fully Automated Prompt Optimization), a Claude Code-driven system that autonomously optimizes multi-step LLM pipelines from baseline prompts to target accuracy.
The new toolset enables researchers and developers to focus on the creative aspects of their work, rather than manually tweaking prompts. FAPO's pipeline-aware approach ensures that each step in the optimization process is carefully evaluated and adjusted to minimize errors and improve overall performance.
This innovative solution has far-reaching implications for the development of AI systems, particularly in areas such as language processing and generation. By streamlining the prompt optimization process, FAPO paves the way for more accurate and efficient AI models that can be applied to a wide range of applications.
As AI continues to transform industries and revolutionize the way we live and work, tools like FAPO will play a critical role in driving innovation and progress. By making it easier to develop and train AI models, FAPO opens up new possibilities for researchers, developers, and businesses alike.
What Shipped
Crawlee for Python: Build a Web Crawling Pipeline with Robots Handling, Link Graphs, and RAG Chunk Export
A new open-source release from Crawlee allows developers to build web crawling pipelines using Robots Handling, Link Graphs, and RAG Chunk Export. This comprehensive tool enables users to create a complete workflow from setup to AI-ready output.
Nous Research Updates Hermes Agent with Blank Slate Mode
Nous Research has updated its open-source Hermes Agent with a new Blank Slate mode, which starts an agent with everything off except provider, model, File Operations, and Terminal. This feature allows users to opt in to the rest of the toolset.
In the Weights is your new AI-centric vanity search
A new AI-powered vanity search called In the Weights has been released. Users can now check their scores using this innovative tool.
Project Fetch: Phase Two
Anthropic has announced Project Fetch, a research initiative that aims to improve language models through novel prompts and fine-tuning strategies.
Cognitive Debt is the New Technical Debt
A new report suggests that cognitive debt is becoming a significant issue in software development. According to CTOs, cognitive debt refers to the mental strain caused by constantly updating knowledge and skills to keep up with rapidly changing technologies.
From the Labs
Nous Research Updates Hermes Agent with Blank Slate Mode
Nous Research has added a Blank Slate setup mode to its open-source Hermes Agent. It starts an agent with everything off except provider, model, File Operations, and Terminal. You opt in to the rest.
Building reliable agentic AI systems
A new whitepaper from Martin Fowler discusses building reliable agentic AI systems. According to Martin Fowler's article, developing reliable AI models requires careful consideration of various factors, including model interpretability, bias mitigation, and performance evaluation.
Project Fetch: Phase Two
Anthropic has announced Project Fetch, a research initiative that aims to improve language models through novel prompts and fine-tuning strategies. According to Anthropic's website, the project focuses on enhancing the robustness and coherence of AI-generated text.
Other Notable News
Nous Research Updates Hermes Agent With a Blank Slate Mode That Pins Toolsets via platform_toolsets.cli and disabled_toolsets.
A new update to the open-source Hermes Agent from Nous Research has added a Blank Slate setup mode that starts an agent with everything off except provider, model, File Operations, and Terminal. You opt in to the rest.
Signal's Meredith Whittaker wants you to remember that AI chatbots 'are not your friends'.
"These are not your friends. These are not conscious beings. These are not sentient interlocutors."
In the Weights is your new AI-centric vanity search.
So ... what's your In the Weights score?
Nobel laureate John Jumper is leaving DeepMind for rival Anthropic.
Jumper isn't the only big name leaving Google DeepMind.
A 3D voxel game engine written in APL.
Comments...
Google Hits 50% IPv6.
Comments...
Developers don't understand CORS (2019).
Comments...
Loupe – A iOS app that raises awareness about what native apps can see.
Comments...
Renting a sewing machine from the library.
Comments...
Windows UI evolution: Clicking an unassociated file.
Comments...
Epoll vs. io_uring in Linux.
Comments...
Comments...
CTOs Agree: Cognitive Debt Is the New Technical Debt.
Comments...
SMPTE Makes Its Standards Freely Accessible.
Comments...
DOS Game "F-15 Strike Eagle II" reversing project needs DOS test pilots.
Comments...
Unauthorized alert sent to cell phones across Brazil.
Comments...
UHF X11: X11 Built for VisionOS and Apple Vision Pro.
Comments...
The Take
In the realm of AI, this past week has seen significant strides in innovation and technology. The release of FAPO (Fully Automated Prompt Optimization) by Cisco Foundation AI marks a major milestone in the field of multi-step LLM pipeline optimization. This Claude Code-driven system autonomously optimizes pipelines from baseline prompts to target accuracy, paving the way for more efficient and effective AI applications.
Meanwhile, Nous Research has updated its open-source Hermes Agent with a Blank Slate setup mode, allowing users to start an agent with everything off except provider, model, File Operations, and Terminal. This new feature offers greater flexibility and customization options for developers working with the Hermes Agent.
However, as Signal's Meredith Whittaker astutely reminds us, AI chatbots are not conscious beings or sentient interlocutors. They are mere tools designed to assist and augment human capabilities, not replace them. It is essential that we remain aware of this distinction and use AI responsibly.
In other news, a 3D voxel game engine written in APL has been released on GitHub, offering a new platform for developers to create immersive gaming experiences. Additionally, Google has reached the milestone of 50% IPv6 adoption, marking significant progress towards a more decentralized and secure internet infrastructure.
As we move forward in this rapidly evolving landscape, it is crucial that we prioritize reliability, transparency, and accountability in our AI systems. Building reliable agentic AI systems, as Martin Fowler suggests, requires a deep understanding of the complexities involved and a commitment to continuous improvement.
Crawlee for Python Cisco AI introduces FAPO Nous Research updates Hermes Agent