First draft of workflow post

content/blog/workflow-vision-neuromorphic/index.md

@@ -0,0 +1,164 @@
+---
+title: "From Idea to Implementation: Reimagining the Neuromorphic Workflow"
+date: 2025-07-06
+description: "Comparing neuromorphic and conventional ML workflows reveals a massive infrastructure gap—and big opportunity to accelerate the field."
+image: typewriter.jpg
+image_attribution:
+  text: "Photo by"
+  author: "Glenn Carstens-Peters"
+  url: "https://unsplash.com/photos/macbook-beside-typewriter-machine-6rkJD0Uxois"
+draft: false
+author: ["Jens E. Pedersen", "Danny Rosen", "Justin Riddiough"]
+showTableOfContents: true
+show_author_bios: true
+tags: ["strategic-vision", "workflow", "infrastructure", "community"]
+---
+
+**TL;DR:** Neuromorphic computing has incredible potential, but our development workflow is far behind conventional machine learning. While ML researchers can go from idea to deployment in days, neuromorphic researchers often spend months just getting their algorithms running on hardware. This infrastructure gap isn't just frustrating—it's fundamentally limiting the field's progress. We need to build the tools, standards, and ecosystem that will unleash neuromorphic computing's true potential.
+
+The post is the second in a series of three posts that lay out the **Open Neuromorphic Strategic Initiative**. [Last post presented the **Open source vision**](/blog/strategic-vision-open-neuromorphic/) and the subsequent post will announce **new initiatives** that will be kickstarted by the [newly elected Executive Committee](/neuromorphic-computing/initiatives/executive-committee/).
+
+Join the discussion [on Discord](https://discord.gg/hUygPUdD8E), star us [on GitHub](https://github.com/open-neuromorphic/), follow us [on LinkedIn](https://www.linkedin.com/company/98345683/), and give us a watch [on YouTube](https://www.youtube.com/@openneuromorphic).
+
+---
+
+## The Two-Week Reality Check
+
+Imagine you have a breakthrough idea for a neuromorphic vision algorithm. Your brain is buzzing with possibilities—you've identified a novel approach to temporal processing that could revolutionize edge computing.
+
+In conventional machine learning, your journey would look like similar to this:
+- **Day 1:** Prototype in PyTorch, leverage existing datasets
+- **Day 3:** Train on standardized benchmarks, compare with baselines
+- **Day 5:** Deploy to cloud infrastructure, scale testing
+- **Day 7:** Share reproducible results with the community
+
+In neuromorphic computing, your journey looks more like this:
+- **Week 1:** Hunt for hardware-specific documentation
+- **Week 2:** Struggle with vendor-specific toolchains
+- **Week 4:** Implement basic algorithm from scratch
+- **Week 5:** Debug toolchain-specific edge cases
+- **Week 8:** Debug hardware-specific edge cases
+- **Week 16-52:** Finish and hope someone else can reproduce your results
+
+This isn't just frustrating—it's [a fundamental barrier to scientific progress](/blog/strategic-vision-open-neuromorphic/). While ML researchers iterate rapidly and build on each other's work, neuromorphic researchers often reinvent the wheel, working in isolation with incompatible tools.
+
+## The Tale of Two Workflows
+
+The contrast between these development paths reveals the massive infrastructure gap holding back neuromorphic computing:
+
+### Conventional ML: The Mature Ecosystem
+- **Frameworks:** PyTorch, TensorFlow, JAX—choose your flavor
+- **Datasets:** CIFAR, ImageNet, GLUE—standardized benchmarks everywhere
+- **Hardware:** CUDA, TPUs, cloud instances—unified APIs across platforms
+- **Deployment:** Docker, Kubernetes, serverless—production-ready pipelines
+- **Community:** Kaggle, Papers with Code, Hugging Face—sharing is seamless
+
+### Neuromorphic Computing: The Fragmented Landscape
+- **Frameworks:** Intel's Lava, BrainScaleS PyNN, custom implementations
+- **Datasets:** DVS128, N-MNIST, custom conversions—limited and inconsistent
+- **Hardware:** Loihi, SpiNNaker, BrainScaleS—each with unique interfaces
+- **Deployment:** Research prototypes, manual processes, vendor lock-in
+- **Community:** Scattered papers, non-reproducible results, isolated toolchains
+
+The numbers tell the story: A CNN implementation that takes 10 minutes in PyTorch can take 2 weeks to port to neuromorphic hardware. A benchmark that runs across dozens of ML frameworks might work on exactly one neuromorphic platform.
+
+## Root Causes: Why We're Stuck
+
+![Comparison table showing the workflow differences between conventional ML and neuromorphic computing](comparison_table.png)
+
+This infrastructure gap isn't accidental—it's the result of several interconnected challenges visible in the comparison above:
+
+### 1. Dataset Fragmentation
+ML researchers enjoy rich, standardized datasets with mature tooling—from CIFAR to ImageNet to Hugging Face's ecosystem. Neuromorphic computing relies on a single primary dataset source (Tonic) with limited standardization. This creates a chicken-and-egg problem: without diverse, accessible datasets, it's hard to develop robust algorithms, and without robust algorithms, there's less incentive to create comprehensive datasets.
+
+### 2. Tool Ecosystem Sprawl
+The neuromorphic design landscape is a patchwork of specialized tools: BindsNET, Brian, Lava, Nengo, Nest, Norse, SNNTorch, and SiMLabs—each with different APIs, philosophies, and capabilities. Unlike ML's convergence around PyTorch and TensorFlow, neuromorphic computing lacks dominant frameworks that can absorb the ecosystem's energy and create network effects.
+
+### 3. Deployment Complexity
+ML deployment has matured into standardized workflows using common platforms and APIs. Neuromorphic deployment relies on emerging standards like PyNN and NIR, alongside specialized tools like NeuroML. While promising, these tools haven't yet achieved the seamless integration that makes ML deployment trivial.
+
+### 4. Analysis Tool Shortage
+ML analysis benefits from mature, interconnected tools—MLflow for experiment tracking, Weights & Biases for visualization, TensorBoard for monitoring. Neuromorphic computing has fewer specialized analysis tools, with researchers often relying on general-purpose solutions that don't capture the unique characteristics of spiking neural networks.
+
+### 5. Community Fragmentation
+Each tool in the neuromorphic ecosystem has its own community, documentation style, and learning curve. Unlike ML's shared knowledge base, neuromorphic researchers must master multiple distinct toolchains, making collaboration difficult and slowing knowledge transfer across the field.
+
+## The Vision: A Unified Neuromorphic Workflow
+
+But here's the exciting part: we can fix this. The neuromorphic field is still young enough that we can architect the infrastructure we need. Imagine this workflow:
+
+### 1. Hardware-Agnostic Development
+```python
+# Write once, run anywhere
+network = neuromorphic.Sequential([
+    neuromorphic.Linear(784, 128),
+    neuromorphic.LIF(tau_mem=20.0),
+    neuromorphic.Linear(128, 10)
+])
+
+# Deploy to any platform
+network.deploy(target="loihi")  # or "spinnaker", "brainscales", "akida"
+```
+
+### 2. Standardized Benchmarks
+A comprehensive suite of neuromorphic benchmarks that work across all platforms, with automatic performance comparisons and reproducibility guarantees.
+
+### 3. Seamless Prototyping Pipeline
+From idea to hardware testing in hours, not weeks. Cloud-based development environments with instant access to neuromorphic hardware.
+
+### 4. Community-Driven Tooling
+Open-source tools that improve with every contribution. Debugging, profiling, and optimization frameworks built by the community, for the community.
+
+### 5. Reproducible Research
+Every paper comes with runnable code, standardized datasets, and hardware-agnostic implementations. No more "contact authors for code" or "results may vary by platform."
+
+## The Path Forward: What We're Building
+
+This isn't just a dream—it's a roadmap. Here's how we're making it happen:
+
+### NIR: The Foundation
+The [Neuromorphic Intermediate Representation](https://github.com/neuromorphs/NIR) is already enabling hardware-agnostic development. By providing a common format for neuromorphic networks, NIR is the first step toward our unified workflow.
+
+### Standardized Datasets
+We're curating and standardizing neuromorphic datasets, creating the benchmarks that will drive reproducible research forward.
+
+### Open Infrastructure
+Building cloud-based development environments and shared compute resources that democratize access to neuromorphic hardware.
+
+### Community Tooling
+Developing the debugging, profiling, and optimization tools that will make neuromorphic development as smooth as conventional ML.
+
+## Your Role in This Transformation
+
+This vision only works if we build it together. Here's how you can help:
+
+### For Researchers
+- **Adopt NIR:** Use standardized formats in your work
+- **Share openly:** Make your code and datasets available
+- **Contribute benchmarks:** Help establish community standards
+
+### For Industry
+- **Support standards:** Implement NIR in your hardware tools
+- **Open interfaces:** Provide APIs that work with community tools
+- **Invest in ecosystem:** Fund open-source infrastructure development
+
+### For the Community
+- **Join the conversation:** [Discord](https://discord.gg/neuromorphic), [GitHub](https://github.com/neuromorphs), [LinkedIn](https://www.linkedin.com/company/open-neuromorphic)
+- **Contribute tools:** Every utility helps build the ecosystem
+- **Spread the word:** Help others discover neuromorphic computing
+
+## The Moment of Opportunity
+
+We're at a unique moment in neuromorphic computing's history. The field is mature enough to recognize these infrastructure challenges but young enough that we can still architect elegant solutions. The hardware is becoming practical, the algorithms are proving their worth, and the community is ready to collaborate.
+
+The question isn't whether we'll build this infrastructure—it's whether we'll build it together, creating an open ecosystem that accelerates everyone's research, or whether we'll remain fragmented, each group struggling with the same fundamental problems.
+
+The conventional ML community didn't achieve its current momentum by accident. They built the tools, standards, and ecosystem that made rapid innovation possible. Now it's our turn.
+
+The future of neuromorphic computing isn't just about better algorithms or faster hardware—it's about creating the infrastructure that will let a thousand innovations bloom. That infrastructure starts with us, today.
+
+*This is the second post in our strategic vision series. Read the [first post](https://open-neuromorphic.org/blog/strategic-vision-open-neuromorphic/) on our foundational vision for open neuromorphic computing.*
+
+---
+
+Ready to help build the future of neuromorphic computing? Join us on [Discord](https://discord.gg/neuromorphic), contribute on [GitHub](https://github.com/neuromorphs), or connect with us on [LinkedIn](https://www.linkedin.com/company/open-neuromorphic). Together, we can transform how neuromorphic research gets done.

content/contributors/jens-e-pedersen/index.md

@@ -1,6 +1,6 @@
 ---
 title: "Jens E. Pedersen"
-description: "Doctoral student at KTH, modeling neuromorphic control systems to solve real-world ambiguity. Maintainer of Norse, AEStream, and co-author of NIR."
+description: "Doctoral student at KTH, modeling neuromorphic systems to solve real-world challenges. Maintainer of Norse, AEStream, Faery, and co-author of NIR."
 image: "jens-egholm.png"
 social:
   - icon: "fa fa-house"
@@ -11,7 +11,7 @@ social:
     title: "github"
 draft: false
 ---
-Jens Egholm Pedersen is a doctoral student at the [Royal Institute of Technology (KTH)](https://www.kth.se/profile/jeped/) working to model and construct [neuromorphic](https://en.wikipedia.org/wiki/Neuromorphic_engineering) control systems.
+Jens Egholm Pedersen is a doctoral student at the [Royal Institute of Technology (KTH)](https://www.kth.se/profile/jeped/) working to model and construct [neuromorphic](https://en.wikipedia.org/wiki/Neuromorphic_engineering) systems.
 
 When faced with the complexity and ambiguity in the real world, contemporary algorithms fail spectacularly. There is a strong need for self-correcting, closed-loop systems to help solve our everyday physical problems.