Open-source technology has been one of the greatest contributors to tech innovation over the past 30 years. The ability to share R&D costs, reuse common code blocks, and accelerate proprietary applications fostered a technological boom that still shapes the digital ecosystem we occupy today.
Our modern web servers, operating systems, and everyday software are built on a foundation of open-source infrastructure amounting to trillions of dollars in economic value. The Linux Kernel, Wikipedia, Firefox, and GitHub are just a few of its countless offshoots. It’s the underacknowledged foundation beneath our modern tech boom. Despite the success, its story was defined by a perpetual struggle with Microsoft and Big Software.
Open source’s influence was resisted from the outset, and not until the 2000s did the movement and broader community get a permanent place in the tech world. Now, just as AI once looked on track to become the new crown jewel of the open-source resurgence, its future is being shaped by the exact same factors. But it may not have the same ending.
The Shift: From Collaboration to Competition
You may have heard that the AI movement was at a crossroads, and for a while it was. After ChatGPT’s release in 2022, OpenAI was flirting with a hybrid half-and-half model, but since their decision to go fully for-profit, the gates are closed and sealed shut. With all the incredible progress made by open-source AI, one could reasonably assume we’d be doubling down on this same approach although the reality is not so simple. Cooperation used to help AI researchers and enthusiasts advance the field and share their knowledge.
(SomYuZu/Shutterstock)
Now, even the slightest advancement could be worth billions in potential profits and IP value. AI models also now require astronomical quantities of data to train, much of which includes increasingly private and sensitive data. That’s not something the AI power players are willing to give up. The movement once built on cooperation, shared knowledge, and benevolent ideals has now become an arms race.
Building Your Own AI Strategy
The differentiator, therefore, comes from being able to train these open models to supplement your organization’s needs. With high quality data, and a strategic knowledge of how best to utilize it, businesses can create their own personalized AI toolkit. You don’t have to match OpenAI’s computing power, nor its billion-dollar coffers. You just need to properly utilize the one commodity you have that they don’t—your data.
Too often, companies have viewed data as a reactive asset. Amid the AI explosion, there’s an over fixation on the tools and models themselves, and not what powers them in the first place. Now, data is squarely at the forefront. With curated, agile datasets, companies can create bespoke AI tools for their needs in real time. When generating valuable insights, streamlining workflows, and automating key processes, this can make all the difference, especially as competitors struggle to keep up.
The Historical Roots and Lasting Impact of Open Source
The story of the open-source movement is inextricably tied to the broader history of programming, coding, and computer science. As personal computers first became widely available, a community of likeminded enthusiasts and experts came together to solve problems, advance the industry, and pool their collective knowledge. This became a deeply prosperous and productive movement, with a community that’s still active today. The fruits of its labor lead to the creation of open-source tools we still use, such as MySQL, Linux, and countless others. According to a recent HBS study, open-source infrastructure amounts to over $8 trillion in economic value. Without it, the study estimated that the average software purchase would be four times as expensive.
Open Source’s Role in AI’s Early Development
The open-source movement was instrumental, too, in the development of AI. Just as in academia, the challenges and complexity of AI was far too much for a single person or team to tackle alone. By democratizing the project, an open-source approach allowed people to share findings and see what worked. Soon, progress accelerated alongside the advancements of processing power and big data. With the technological leap, AI suddenly went from a thought experiment to a genuine plausibility.
Sam Altman at OpenAI DevDay on November 6, 2023
This came to a crescendo when OpenAI entered the public sphere in 2015, claiming to be “advancing digital intelligence in the way that is most likely to benefit humanity as a whole.” Quietly, though, this was still the incubation phase, and progress continued to the point where more data, more power, and more chips were required. The non-profit model was too limiting. While the open-source stage was central to its development, it had grown far beyond this narrow scope.
In this new era, companies must think differently, if not fully reconsidering what they want their data to do/be. An occasionally helpful resource? Or your company’s greatest asset—the one-two punch of a transformative AI strategy? Doing the latter means building out your data team: data architects, data engineers, data scientists, and data analysts, equipped with all the tools they need.
For too long, these specialists have been undervalued in the IT ecosystem, seen as data librarians or storage experts rather than actual strategists. With the onset of AI, this is the data super bowl, and their responsibility (and budget) must grow accordingly. Data scientists should be involved in the big decisions; data architects should enjoy the freedom to build new internal systems. At every level, the data team should have a hand in the decision-making process. That’s the work that will elevate your AI strategy from standard-practice to best-in-class.
A Closed Future
It’s difficult to speculate on the future of AI, but it seems unlikely that open source will play a role in it. Just like with Big Software before, the power players are now light years beyond everyone else – and gaining. This time though, the datasets, algorithms, and training methodologies have progressed so rapidly that they now border on outright trade secrets, encapsulating years of insights and research. Despite the enthusiasm of the open-source AI community, these may prove too valuable to wrest from private hands. If you once thought you might be able to make a comparable AI tool through the collaborative power of open source, that ship has sailed. Luckily, the actual AI models themselves are the most valuable tools for everyday businesses, and because these can be shared without compromising the proprietary inner workings, they will continue to be accessible.
Conclusion
No matter how big the AI titans become, and or how advanced their models get, they will never know your company like you do. A great deal has changed over the past few years: good data is now among the most valuable commodities a company can have, and they should utilize it accordingly. With the right team behind you, and the right approach, you can build an AI strategy that not even money can buy.
Chris Stephenson
About the authors: Chris Stephenson is the Managing Director of Intelligent Automation, AI & Digital Services at alliant. Chris has delivered on multiple internal and client-facing AI products and boasts over 25 years of entrepreneurial and consultative experience in various sectors, advising companies like Amazon, Microsoft, Oracle and more.
Dhaval Jadav
Dhaval Jadav is the co-founder and CEO of alliant. Under Jadav’s leadership, the firm has empowered 24,000+ companies in 70+ industries with actionable business solutions. Prior to founding alliant, Jadav developed his expertise in high-tech business deals as a member of a Mergers & Acquisitions/Private Equity/Strategic Buyer Services Group in San Francisco and honed his business operations acumen with Deloitte & Touche in its Washington National Office.
The post AI’s Open Origins and How it Affects us Now appeared first on BigDATAwire.
Leave a Reply