Why NASA, IBM, And Hugging Face’s Open Source Model Is A Big Deal
What do you get when you combine an open source platform, a massive and critically useful dataset, and an ability to open-source an AI foundation model?
If you’re NASA, IBM, and Hugging Face, you get a massive opportunity to make geospatial data available to all through an open source geospatial AI foundation model. We like this open source geospatial intelligence resource and commitment for three reasons:
- By making US geospatial data and intelligence available under an open source model, it creates a level playing field for all people and organizations to bring geospatial intelligence into their own applications. We’ve written extensively on the importance of location data and open mapping platforms in real-world applications. This can be a key resource.
- The conversational interfaces on a geospatial foundation model unlock the value of that data and intelligence — this puts knowledge in play. For geospatial data, the value of all that intelligence creates opportunities to improve applications in resource planning, business optimization, crop yields, policy foundations, and more.
- It’s a signal to other owners of large data and knowledge libraries that they, too, can make that intelligence available through a domain-specific foundation model. I predict that we’ll see first dozens, then hundreds, and, ultimately, thousands of domain-specific generative AI (genAI) intelligences made available. Most will be paid services, though, not open source.
This is also a great reminder to technology executives that you will incorporate many intelligences into your genAI-fueled applications. Do not expect or plan to rely solely on a large language model from Microsoft or Google. Most of the specific value will lie in these domain-specific genAI intelligences.
You will create real applications by orchestrating the intelligences you need (including your own knowledge foundation models and your machine-learning models and software). We are thinking of this as a genAI application runtime based on layers of intelligence (yours and theirs), input and output gates, prompt (and cost) engineering, and orchestrating of the best intelligences to generate and verify the output.
Let’s Connect
If you’re a Forrester client and you would like to learn more, please set up a time for us to talk. I can help direct you to the best analyst to speak with if I’m not the best person. You can also follow or connect with me on LinkedIn if you’d like. If your company has expertise to share on this topic, feel free to submit a briefing request.