DevelopersDevOpsFeaturedOpen SourceState of EnergyVideo

Michael Dolan Talks About Open Source Foundations, OpenAI And More

0

Guest: Michael Dolan (LinkedIn)
Organization: Linux Foundation (Twitter)
Show: State of Energy 

In this episode of State of Energy recorded at the 2023 LF Energy Summit in Paris, Michael Dolan, SVP and GM of Projects at the Linux Foundation, talks about the various aspects of open source and the foundations that support them.

LF Energy

  • LF Energy is about taking an ecosystem that is very entrenched in its ways and the processes and the inertia required to make any changes extremely difficult.
  • Dolan shared the story of how LF Energy came to be. He was at the Linux Foundation office in San Francisco having a meeting with his colleagues when Shuli Goodman happened to show up. After a brief introduction, she said she had this big idea about how open source and open collaboration can change things that are going on in the energy industry. It takes bringing people together and she didn’t know exactly how to do that, but she heard the Linux Foundation does. The initial cohort started off very small. Today, LF Energy has over 70 members, 24 projects, and over 1,500 active contributors.
  • To meet some of the Climate Accord objectives and all the different climate initiatives that are out there, doing what we have today and using the current toolbox is not good enough.
  • What is going to be the future is going to depend on collaboration and on software. If you look at any code base, roughly 70%-90% of that code base is going to be based on open source.
  • The energy sector struggles with the fact that there’s a very scarce resource in terms of engineers who know how to program and develop systems for energy. When you’re starting at that point, collaboration makes even more sense.
  • LF Energy can be a diffusion point for disruptive or change technology.  It’s not tethered to some of the constraints of the past. For microgrids, renewable energy sources, batteries, electrification of vehicles, and the whole supply chain of energy, these need to be connected through software. And when you talk about software, you’re generally talking about building it with a base of open source that can be collaboratively built, and then reused to help everybody in the world.
  • A common element of open source is reusable components. LF Energy’s FledgePOWER is a spinoff of LF Edge’s Fledge project i.e., it is based on software that came out of the telecommunications industry but with added capabilities. They’re learning about software elements that are outside of the energy industry and incorporating those best elements into the energy system.
  • What better demonstration of the power of open source is there when you can literally just pull components off-the-shelf from other open-source projects and get 80% of where you need to be, add an additional 20% and off you go.
  • LF Energy provides a forum for all of them to come together and have conversations about their challenges and who wants to work together to address them. It is also educating leaders about some of the business, legal, and engineering elements of structuring plans for their internal software stack.
  • In Europe, the early participants have been visionaries in terms of what needs to be done. They’re also under a lot of pressure, which has pushed them to make these investments to be out there in front because in order to meet the decarbonization goals and the other commitments that they’ve made to their constituent audiences, they need to move fast.
  • Regulators and governments are now paying attention to open source. They’re responsible to citizens for delivering certain types of infrastructure and innovation and open source is a great accelerant to many of those initiatives. Challenge: Governments get involved at a high level, and they don’t understand the details of how open source actually works. There may be a mismatch between the regulations and the practices in the industry.
  • Some of the current conversations revolve around who should be responsible for making sure that the products that are delivered to constituents have, at least, a baseline level of security: is it the distributor of the product or is it the open source project community upstream?
  • Being able to keep that organic, community-driven innovation is really important to LF Energy.

Linux Foundation Europe

  • Every region around the world is trying to grapple with how to innovate, improve their systems, and stay competitive. Every region is going through digital transformation and trying to figure out how to do it. A lot of local collaboration needs to happen because there are support systems around these companies that need to come along with them.
  • Linux Foundation Europe was born out of the need to collaborate amongst different countries on a set of framework initiatives under which they can work. It provides a structure and a forum for European organizations to collaborate on European problems. The European Union is funding a number of these initiatives, and they want to see it land in an entity that’s in Europe.
  • Europe is a bit unique, so the Linux Foundation does not necessarily need to set up a similar type of structure in other parts of the world yet. But if there is an opportunity where there’s this cross-border collaboration, LF would definitely look into it.

Linux Foundation AI & Data

  • Generative AI has taken the world by storm not just by its capabilities, but how fast these tools have improved. In the very near term, there will be a lot more powerful tools than we even have today. LF AI & Data has been working on a full stack and suite of tools that organizations need in order to put artificial intelligence and machine learning into practice in their organization.
  • Dolan envisions having open datasets like what LF has been working for years in terms of community data license agreements and getting some of the core legal frameworks down, so that people can be comfortable sharing data openly, especially for training AI models.
  • When people talk about AI and the risk, they’re generally talking about accountability, traceability, and issues around provenance. LF AI & Data has developed tools for those.
  • Current trends: Having a pipeline, from training sets to tools, operational elements, and the large language models that are all open. Many companies do not have enough AI engineers and experts to do this themselves so collaboration makes a lot of sense for them.
  • What is “open” in terms of AI is a question that needs to be answered. There is a need to work on a path of what is open. Communities are moving fast, and they’re moving with the speed that whatever definition you come up with, maybe out of sync or out of date with what’s been going on in the community. Dolan thinks that at every element of the stack, whether it’s the model, the data, the tools used to build it, are those openly available? Are they reusable? Are they modifiable? Can they be redistributed?

Open Source Licensing

  • There is a shift towards more permissive licensing.
  • Project communities have solved the challenge with some sort of conformance or testing or compatibility suite program in order to solve this from a trademark point of view, not a licensing perspective. So, some of the policing is driven through trademark programs.
  • When an organization is fearful of a SaaS provider or cloud provider taking their code and making it available at scale, Dolan questions whether they should have been using an open-source license at all.
  • A lot of the relicensing are often those single-company projects, where one company already has control and makes all the engineering investment. Ultimately, there’s going to be another open source alternative out there to which the people may switch.

Role of organizations like Linux Foundation, Apache Software Foundation, Eclipse Foundation

  • Provide a neutral, non-competitive open space for people to collaborate, knowing that no one company can then suddenly take it over or change the terms.
  • Focus on writing down the governance so people know what the rules of decision-making are, writing down the intellectual property terms to make sure everybody’s operating on the same playing field from an intellectual property perspective.
  • Democratize the use of the project trademark. If a project wants to put together a trademark usage policy, they do it as a community, not as one company dictating to everybody else what it needs to be.
  • Facilitate that collaboration.

How Linux Foundation projects work

  • Companies who are investing in LF open-source projects are generally investing engineering resources.
  • Decision-making on code bases is done by the technical community under a typical democracy governance model, meaning, the people doing the work are making the decisions.
  • Nobody pays any money to the Linux Foundation or any of its sub-projects to make decisions about the technical code bases or the technical artifacts that are being produced. That’s all based on showing up, making a contribution, having the merit of that contribution evaluated, reviewed, and accepted by the community.
  • LF does not give some companies the ability to veto or decide something, but companies understand that they have to apply a lot of their engineering resources. For example, if they have 10 engineers working on Kubernetes at $250,000 or more a year in annual fully loaded cost, that’s a $2.5 million-dollar investment. They might give the Linux Foundation $100,000 to help support them with  a build release engineer, or a documentation writer, or marketing support, or an event to get these 10 developers together with 1,000 other developers from other companies to decide how the next generation of some API architecture works.
  • Companies do invest a lot in the Linux Foundation projects. They benefit from the productivity of their engineers and the success of the project that they want to see adopted more and more worldwide.

This summary was written by Camille Gregory.