Rethinking the cost of open-source software
“If I have seen further than others, it is by standing on the shoulders of giants.”
Isaac Newton’s phrase is often used in conjunction with Open Source Software (OSS) – software that developers make available for others to use, modify, and distribute. Since it gained momentum in 1983, it’s become the foundation of our digital economy, comprising 70-90% of any modern application. Not only is it pervasive in our digital lives and at the core of many of the companies that provide our digital services, but it still has huge relevance. Some of the big generative AI models have been built using open-source models.
However, prolific and critical open-source vulnerabilities and attacks, including the “catastrophic” Log4J and the Equifax breach, which resulted in the records of over 147 million people being exposed, have fueled questions about its security. In parallel, OSS’s viability has become increasingly called into question. Developers are commonly uncompensated for their contributions to open source, while large corporations have profited from their work, creating an inequality at the heart of the movement.
As one of the founders of the Open Source movement and the creator of the Open Source Definition, the set of legal requirements for Open Source licensing which still stand today, Bruce Perens has unique insight into this landscape.
One of his proudest moments, he tells TFD, was getting Linux on the Space Shuttle – proof that OSS could do mission-critical work. But despite the success of the movement he helped create, he has questions about whether it has fulfilled the potential and the promise the early OSS pioneers hoped for.
In conversation with TFD, Perens shares what he sees coming in the next decade for software. He discusses whether the price paid for sharing was too great and offers a new future vision of how we can use open source more fairly in the next decade. And he looks at how AI will disrupt the software industry, the new issues it will bring, but also the hope it offers in helping solve some of the biggest challenges the human race faces.
I feel we have actually failed to help people because our software is used to monitor them, surveil them, and even exploit them. We could provide them with software that respects their freedom, and we’re not doing so.
When the World Wide Web was merely a twinkle in Sir Tim Berners-Lee’s eyes, there was already a philosophy that would guide it. And that philosophy was based on the academic principles of shared knowledge.
Software was not seen as a commodity and early developers worked in collaboration, making their source code available for others to learn from and improve in the great tradition of scientific discovery.
Fast forward 40 years and software is now the engine that powers our digital lives. And open source software has played a powerful role in that.
“Every company uses open source. All the devices around you use open source. It’s in your phone; it’s probably controlling the traffic lights when you walk down the street. It’s used in air traffic control when you take a flight,” said Perens.
Recent research from GitHub, the Linux Foundation, and the Laboratory for Innovation Science at Harvard suggests that organisations now invest around $7.7 billion in the ecosystem each year.
Enhancing democracy
But that doesn’t necessarily equal success, he argues.
“It’s everywhere, it is used for everything, and yet still today we have 66% of developers that are not compensated in any way for its creation,” he told TFD.
And it is not just the monetary problem that Perens worries about.
“The creators of the internet, and all the open source developers, thought what we were doing would enhance democracy. But we didn’t have this flow of intellectual capital that we expected to happen. I think part of the problem was that we were all scientists and technologists, and we thought that if you connected all of us, great things would happen. And we did not really consider who the real users would be, and what would happen with them.
“I feel we have actually failed to help people because our software is used to monitor them, surveil them, and even exploit them. We could provide them with software that respects their freedom, and we’re not doing so.”
There was, he said, a disconnect between what the open source community wanted and how society responded.
“We started sharing, and most people took it as a gift. But that isn’t sharing. It requires some give and take.”
He now preaches a new philosophy and a solution called Post Open, which attempts to shore up the gaps the current licensing of software leaves open. It is based on the idea that large companies pay fairly for the software they use.
AI’s binary dilemma
But, in the age of AI, will it even be down to humans to code our future?
Google’s CEO, Sundar Pichai, has said that more than a quarter of new code for Google’s products is generated by AI and such tools are becoming more popular and ubiquitous. For any youngster considering a career in programming, it poses a problem: why bother if AI will replace you?
When it comes to how AI will play out, Perens has a developer’s binary approach to how the future may unfold.
“Software is the creation of a system that performs a job with a very large number of logical decisions that a human being currently makes. Software at scale, something like a web browser, has so many lines of code that no human being can actually understand all of it. If AI becomes capable of that, then AI is capable of performing all jobs; writing novels, being the CEO. If that actually happens, I think we have been replaced,” he said.
But getting there will require AI to make some bigger leaps than the ones seen since ChatGPT burst onto the scene in November 2022. He questions if we are about to enter a new AI winter, a period of time when no great steps forward are made in the development of the tech.
“Will we hit a boundary? Right now, the large language models have tremendous weaknesses and are prone to hallucinations. We have had 50 years of thinking we would have strong AI, something that thinks the way we do, and it didn’t come. And now we’ve seen this tremendous jump, but will it go further?”
Even if AI becomes smart enough to take on coding jobs, history tells us that humans are eminently adaptable in the face of new technology.
“When automation came to factories, there was a lot of anxiety about it and the creation of the Luddite movement, but it became part of everyone’s life and did not obsolete us,” said Perens.
And he sees huge potential for AI-driven software in one key area that the early web pioneers are also keen to influence.
“At the beginning of the web, we talked about it being a tremendous boost for education, but we’ve also seen these negative impacts on society. I think we will see a revolution in education in the era of AI, where education becomes more personal and more individual.”
I actually see security getting better. One of the reasons for that is that we are abandoning computer languages like C++ and Java in favour of things like Rust, which very deliberately shuts out an entire class of bugs that exist in older languages.
Genetic code
When it comes to how software may develop in the next decade, he believes it has a huge role to play in helping to solve some of the biggest challenges we face.
“If we look at genetic code, it’s very much like a computer program and thus something that we can understand in a computer science way. Ultimately, cancer is an error in the program, a bug in your genetic code and computational biology can help solve these problems.”
Businesses may be looking for more prosaic bugs to be solved in the short term, as software has become synonymous with hacking.
Software bugs have led to thousands of data breaches over the last few decades and wiped billions off our economies.
Post-truth era
But in the future, Perens sees improvements.
“I actually see security getting better. One of the reasons for that is that we are abandoning computer languages like C++ and Java in favour of things like Rust, which very deliberately shuts out an entire class of bugs that exist in older languages. We will continue to see the evolution of computer languages that will help programmers create fewer bugs.”
Whatever the next decade holds, Perens remains hopeful that those tenets of shared scientific knowledge that began the software movement will propel it forward too, but there will be big barriers to overcome first, though.
The internet may have enabled shared knowledge on scales never seen before, but that has brought with it a new philosophical problem – the daily bombardment of information we can all access so readily has led to misinformation.
“The biggest problem that we have in society today is that we are living in the post-truth era, the era in which simple lies are more compelling than complex truths. Guiding society past that is the biggest challenge we now face, and I don’t have the first idea how to solve it.”