What's in a name -- why you should care what open source is
"We really should stop getting hung up about what open source means," a statement repeatedly made by those for whom a lack of understanding of open source is advantageous. Generally those who don’t want to meet the standards that are set out in the Open Source Definition (OSD) which all Open Source Initiative approved license must meet, including the requirement that open source software must be usable by anyone for any purpose. And usually they are in denial of the requirement for open source licenses to be usable for any purpose, which includes commercialization. This really sits at the heart of open source.
In a commercial context open source means enabling your competitors with your own innovation. In a world where companies are driven by shareholder value this simply would not happen if it did not come with clear and measurable advantages, like collaboration, creating a defacto standard or building an ecosystem around it, and where equivalent or greater value than the proprietary royalty model is generated.
The journey to open source
The evidence of this can be found in companies like Microsoft -- once public enemy number one to open source software -- today the single biggest contributor to it. Microsoft’s journey to open source is explained by them on the basis that: all developers who have learned to code in recent years have been taught open source software methodologies and will only work for companies that allow them to work in this way, recycling and reusing code. In a digitalized world where competition is high to employ the best developers, this is an important point.
Open source as a disruptive force
Microsoft’s journey was a slow one. It took a decade. We live in a different time, today, from that near past. Generally sectors can be seen to transition to open source in around 3-5 years. As a sector begins to adopt open source, inherently it will be subject to disruption. A shift from the closed, proprietary model of the last 30 years. What that means in reality is that companies which have become reliant on the revenue generated from patents, paid for standards, software or any other form of royalty bearing IP, will find their revenue models undermined by the arrival of open source software, with its free flow and open sharing. Unsurprisingly that is not generally well received by those companies.
FUD and open washing
The response is known amongst the open source community as FUD -- fear, uncertainty and doubt. The creation of a cloud of confusion -- a trick of smoke and mirrors that is simple to achieve when many in that sector are new to open source and don’t understand how it works. It is particularly nuanced. Key in these activities is the act of "open washing" -- declaring something to be "open source" implying that software is distributed under an approved license to make it open source.
Of course this implication is disingenuous. Occasionally this is done to add an ethical type of restriction to the license. More likely it occurs when adding commercial restrictions such as a requirement to pay at a level of use or where use is commercial. That’s of course entirely fine, but it is not open source and should not be described as such.
Use of the term open source in other areas
As a consequence of the recent Hamas attacks on Israel, we have seen the EU Commissioner Thierry Breton post on X that Elon Musk needs to tackle the spread of disinformation on X. Musk’s response… "everything is open source and transparent."
Of course Elon does not mean it is open source software. Instead he’s referring to the term open source journalism, meaning it is community generated.
Unfortunately Elon also referred to open sourcing Twitter earlier this year -- when indeed he did mean open source software.
Think Apple Records and Apple computing -- two very distinct things, both abbreviated to Apple. In the same way, open source software and open source journalism are also frequently abbreviated to open source.
And to add insult to injury, many use open source to mean the generic set of 'opens' sometimes also referred to in a clearer way as open technology.
Open Source and AI
As AI permeates every conversation, and the viral scaremongering of exponential risk rings in our ears, many decry open source. They wish to push a 30 year old genie back in its bottle. It must be regulated they say -- and there was plenty of this at this year’s UK Labour Party Conference Fringe. Ask any of those naysayers what open source means and the likelihood is that they cannot explain or will mis-explain. This thing they call open source, and claim to not want, in actual fact is not open source.
The company OpenAI is a fabulous example of this. Its product ChatGPT is in no way, shape or form open source and despite its name it has nothing to do with open source. Confused?
Equally so Meta’s Large Language Models (LLMs). LLama, distributed on a research license and leaked, therefore unlicensed, not open source. LLama 2 distributed on the Llama Community License -- which has commercial restrictions and is not Open Source Institute (OSI) approved -- along with an Acceptable use policy, and is not open source. The same is true of Alibaba’s LLMs Qwen-7B and Qwen-7B-Chat which are distributed on non approved licenses with commercial restrictions. However the UAE’s Falcon LLM is open source as it has an OSI approved Apache 2.0 license without modification or any commercial restriction.
IBM’s platform for AI watsonx includes open and open source components. IBM, a long term user and significant contributor to open source, carefully distinguished each from the other. A clear and honest approach.
Ultimately we wouldn’t regulate a car with rules applicable to a bicycle or vice versa. And if one was to say a car was a bicycle, then oneself easily becomes the subject of ridicule… a case of the Emperor’s new clothes.
Not everything has to be open source software. But, if we are to create policy that protects our society and enables innovation and enterprise to flourish, surely we must be accurate in this?
To be accurate we must demonstrably know what we are talking about and accurately use terminology. Whether the argument is for or against open source, it will certainly have far more credibility if what is being discussed by tech leaders and policymakers actually is open source and the appropriate risk is being assessed and acted on.