BrickSimple LLC

Every software ecosystem tells a story.

At BrickSimple, we define, design, build, deploy, and operate resilient software that is strategically important for our clients. We are 100% US-based and much more efficient and cost-effective than off-shoring. We have the numbers and named references to prove it. With 20+ years experience in several industries, we know the intricacies of your marketplace deeply, but never pretend to know more about your business than you.

In the age of AI, is "coding" as a viable career dead? Our take: "coding" was never a viable career in the first place, and we've never hired "coders" in 20+ years in business. That said, AI is a useful tool for speeding up the "coding" process and is a valid assistant to a knowledgeable software developer. However, AI-generated code is like stock photography - quick, easy, and "good enough" when you need something right now. But it's not complete or usable without someone to give it context and to decide if it is correct for purpose. For example, here's some number-sorting Python Code written by an AI: def main(): ???user_input = input("Enter numbers separated by spaces: ") ???numbers = [int(number) for number in user_input.split()] ???numbers.sort(reverse=True) ???print("Sorted numbers from highest to lowest:") ???print(numbers) if __name__ == "__main__": ???main() But....will the numbers always need to be sorted this way? What if the input is not a number? Will this code properly work with negative numbers? Do we even need this code at all? Is there a better way to accomplish this function? Are those numbers already sorted somewhere else? Good software development requires clearly understanding what the whole system needs to do, evaluating and managing the interactions between various parts of the whole, and thinking about the best way to create the intended outcome. That's why we've never hired "coders" - only Software Developers who think about much more than the code.

Oh look, another Apple Vision Pro thought piece. However, unlike the many others, we have actual practical experience for over 15 years in "spatial computing" and we have some questions about if this device is a Newton or a Lisa. https://lnkd.in/ejaMmUit

What makes a great structure? It's just not about the materials. It's about knowing what needs to be built and how to build it. It's the idea, the architecture, the context of the structure, and the care and skill that goes into the build process. What makes great software ecosystems? It's not about the code. It's about knowing what needs to be built and how to build it. It's the idea, the architecture, the context of the code, and the care and skill that goes into the build. Image AI-generated by DALL-e

Imagine if you had to check if your phone was compatible with Google. It’s an absurd thought, but for inventors of medical devices, when it comes to data flows and interoperability, that is the case today. Each device inventor has to make costly decisions about what data it produces, how the data are packaged, and more. A medical device can’t just be turned on and connected to any sort of back-end data analysis system; there is no common way for a device to connect and communicate with a larger data collection system without a lot of technological drama. Trying to easily get data out of devices and into contexts where they can be useful is a big challenge because there is no working regulatory model for creating an interoperable data ecosystem - a normalized way for data to flow between devices and systems. The work done to date isn’t promising because it is lagging behind other highly digitized industries such as commerce. Your credit card works with any card reader device in the world because of ecosystem-oriented standards like PCI. These and other functional interoperability standards have made global commerce work for decades. Consider that the FDA’s Center for Devices and Radiological Health (CDRH) specifically says, “CDRH does not require the use of a specific format for clinical trial data” (Source: https://lnkd.in/gmFjS2Cy) On the same page where it says there is no specific format, there’s a list of various data standards that could be used, including DUNS numbers. Despite the fact that on April 4, 2022, the federal government stopped using the DUNS Number to uniquely identify entities. https://lnkd.in/gjywtnbC? This does not bode well for interoperability of data flows. Sure there are data structural standards - there’s CDASH and STDM and ADaM and dozens of others, but these are like describing the dimensions of a room without defining what the room is for. In our view, there’s too much data and drama in the middle - the place between the device and the systems that make the data from the device valuable. We see the development of well-trained LLM AIs, and even simple machine learning algorithms as a viable solution to this challenge; data ETL (Extract, Transform, Load) in real-time is not an unreasonable requirement for systems that want to ingest and utilize medical device data. In the same way you don’t worry about what phone you use to connect to Google, no medical device inventor should need to worry about what data platform their device connects to…that’s not the job of the device, that’s the job of the software ecosystem around the device.

We build software ecosystems for new medical wearable devices. If you are an inventor of or investor in new medical wearable technologies, we can help you bring your product out of the lab and into the marketplace. We cover the Design, Architecture, Development and Management of: - Wearable Medical Device Connectivity - On-Device Embedded Software - Companion Mobile & Web Apps - Cloud Platforms & Services (inclusive of AI, Data i/o) - 3rd party Integrations (including all EHR's) We have over 20 years experience and a long list of satisfied customers. Contact us: [email protected]