What's next with open source AI?

What are the trends and counter trends of artificial intelligence in the time of OpenAI, ChatGPT, GPT-4 and stable diffusion?

Photo by Pavel Danilyuk from Pexels: https://www.pexels.com/photo/a-person-playing-chess-8438944/

Computer science is relatively democratized compared to other science fields

Daniel Lenton - Ivy

On the last hype of AI that we faced recently with StableDiffusuion and ChatGPT, we can see how AI became more popular, and people started to talk about it outside the engineering circles! This hype can show you the value of AI-related topics and widen the horizon for those who have never heard about AI use cases. 

Here we will talk about the trend and counter-trend of AI, based on our conversation with Daniel Lenton, the founder of Ivy, a company on a mission to unify AI. 

The Trend: building moats.

After reaching millions of users in three days for ChatGPT and being acquired by Microsoft OpenAi turned their work towards business! By doing, they started to build moats around their work and strategy. The moats aim to obscure their models and hide their information even though their main goal was an Open AI! Right now, they have an advantage in using Azure Cloud from Microsoft! Monetizing artificial intelligence made them release obscure papers about their work by hiding the mathematical model and the code information, which is a shift towards the closed source to keep themselves at the top of the market. 

The counter-trend: AI for all! 

On the other hand, many startups use open source to their advantage, which can help them market their company or get themselves known. Besides that, there is a massive move in other tech companies towards open-source AI, like HuggingFace, which opened its version of ChatGPT. Such movements can help improve the whole ecosystem and have a competitive advantage in AI to reach the less privileged people and make AI more inclusive. 

---

Why Open Source Artificial Intelligence (AI): 

When we discuss the advantages of open source in any topic, we will have broad topics to cover, like values, empowerment and ecosystem, inclusivity, and even cutting-edge development. 

Here we will focus on a few cases related to those topics about artificial intelligence and how this can help our world! 

Values: how altruism can be a driving force!

In terms of values, those who work on open source have different motivation values that usually bring them to open their work without any expectations. They value altruism and helping others. And see how doing such work can create an excellent place for everyone. Usually, they might be able to understand those who suffered. Another group of people can be motivated because they want to market themselves and are using open source as a tool. But the main idea here is that you will be able to understand the genuine people who use it as a step to reach their goals. A great indicator can be the work they do outside open source, like; volunteering in different NGOs or going towards a cause beyond themselves. 

Here I would like to quote Prof. James McClelland, the Director of the Center for Mind, Brain, Computation and Technology at Stanford, in his article "The Ai of our dreams":

Indeed, being able to pursue a goal that is greater than oneself is an important source of inspiration. Often throughout human history, truly innovative thinkers have appealed to someone or something greater than themselves, producing profoundly influential innovations.

Empowerment: Beyond well-developed environments

Thinking about empowerment, we can see how an open-source artificial intelligence model, mathematic explanation, and code can help more people to get into the field and help others to transfer knowledge from this field to another. Open source would primarily benefit those who are outside the network and impact circle of the USA/Europe tech industry, those who do not have the privilege to use a one thousand dollar service in a month, or even those who do not have the wealth and environment to get into top schools. 

Having an open-source ecosystem can help those people with the following:

• Adjusting the code to their needs and having faster development opens the door for them to have a higher quality of life by showing off their work or ability to understand detailed papers. 

• Opening the door for further development with lower computing needs, this can only happen with an open mathematical model that will have someone with constrained computing power to think about a solution! Such solutions could be more valuable to the whole ecosystem since they would be more power and money efficient. 

Cutting-edge: The art of engineering!

Those who work on cutting-edge research do not work on it for the status quo; they do it for their joy and like what they do. For sure, having stable life can help them to grow faster. But they would never stop what they were doing; even in the most challenging time, they would keep moving forward. For the same reason, you can find these traits in open-source lovers who do software and programming because they love and enjoy it. That kind of people work without thinking about hype or media and marketing, and usually, they are philosophers who work towards the development and gettings things better. They are more like an artist rather than engineers, and they are making things better without having an exhibition. That kind of people usually find the knowledge and learn it by themselves, through various resources, and share it with others; mostly, you can find them becoming educators or impacting their community.

Inclusive leads to creativity! 

As for inclusivity, open source development and, more importantly, open artificial intelligence will include more people in projects from different disciplines. Being inclusive can spot the blind spots of privileged people. Besides that, it will provide more diversity in the field. This diversity happens when people from NGOs or even from other areas can use and modify open-source models for their needs. The low barrier in training and getting educational resources will give them an advantage that might not be possible to consider otherwise. It could help more people get vaccines, clean water, and other food resources, or even plan the needs for a refugee camp based on the models we have trained before. 

In the end, artificial intelligence will take an exceptional part in our lives in the future, yet having it open-sourced will give us a significant step ahead and provide better quality code and diversity in the long term.