Unless you’ve been out of office, not reading email or somehow unplugged, you’ve probably heard about and played with one of the many generative AI tools out there: ChatGPT, JasperAI, Merlin and new entrants, Bard, from Google, and Microsoft’s AI-powered version of Bing. These new AI-backed chatbots promise and deliver ready-to-use “content,” or text to use in README files, blogs, emails, training modules, songs and more. New use cases pop up every day. Real estate listings are the latest darling for chatbots.
The sentiment at the onset of ChatGPT quickly went from snickers to shivers and back again. It’s clearly a powerful tool, but the application of such power is terrifying at first. Stories of AI bots passing law exams, writing college essays and penning trial judgements abound. As quickly as those stories erupted, so too did the tales and announcements of tools that can detect and govern AI-generated content. Intense debate over ChatGPT surfaced a range of positive outcomes, such as revolutionizing digital marketing, as well as warnings over its failings, namely the more writing AI does for us, the fewer of us will learn and practice the skill. The technology’s early days will continue to feature turbulence and swings of opinion tipping back and forth like, well, a swing. It’s evident that the next phase of skill development will be chatbot prompt creation (you can’t get good answers without good questions) and final editing.
While most chatbots are not open source, they are all based upon open source software. In the AI/ML space, open source plays a prominent role, whether it’s Python or PyTorch or TensorFlow, languages such as Python, or engines such as OpenML. A more comprehensive collection of open source projects can be found at the foundation dedicated to all things AI/ML, the LF Data & AI Foundation and of course the newly formed PyTorch Foundation. The VMware OSPO team is deeply engaged in open source projects across the board. Take a glance at adversarial AI, modeling, and ethical and security issues in Diana Atanasova and Teodora Sechkova’s blog “How to Build Trustworthy AI with Open Source.” Enrique Corro of VMware’s Office of the CTO explores the ethical principles of AI and ML in his blog “Why Your Organization Needs a Set of Ethical Principles for AI,” stating “biased data sets, careless misuse, and bad actors can easily turn AI into a weapon with dire consequences.” And VMware’s Research and Innovation team continues to advance ML with Project Trinidad and enabling enterprise-grade MLOps with Tanzu and cnvrg.io, an AI OS transforming the way enterprises manage, scale and accelerate AI and data science development from research to production, all with open source at its core.
What’s Old is New Again
But none of this is truly new. Consider TensorFlow and PyTorch: These two seminal AI/ML open source projects have been around for six years. And the term “artificial intelligence” was coined back in the 1950s by John McCarthy. Some sources assert that AI began even earlier, in the 1940s with Enigma and Alan Turing’s work. So, if AI/ML has been around for a while, why has the pace of innovation suddenly accelerated, with an explosion of new tools and ideas surfacing every day? What ignited this space, making it move so quickly? Well, aside from a deep and abiding case of FOMO (here’s looking at the big players Google, IBM, Alibaba …), I suggest it’s the open source community — the people — that got us here and will keep us moving forward.
Just for grins and giggles, I asked a generative AI chatbot to create a poem about open source, and our team volunteered some couplets to add to the ode. Here’s the result:
An Ode to Open Source Software
Open source software, a gift to us all,
A solution that answers the tech world’s call.
A collaboration of minds, a community so bright,
Bringing innovation, with code that’s just right.
No more vendor lock-ins, no more hefty fees,
A source of empowerment, for you and for me.
A playground for developers, to hone their skills,
A platform for creativity, where tech heartbeats thrill.
From Linux to Python, from Apache to Ruby,
A world of possibilities, just waiting to be.
Open source software, a friend to us all,
A promise of freedom that always stands tall.
But use and step forward with caution
For a repo isn’t the same as completion.
The code is but a starter — a dream in the making.
It’s yours to complete — it’s there for the taking.
But then you’ll put it into production
Grumbling, complaining, grousing about friction.
The install is hacky and the docs … Hey!
Why don’t these developers meet their SLA?
It’s always in there, deep in your code
And if you don’t care, bother’s down on the road.
So take good care and stick to the rules,
Or it’ll turn to nightmare, and joy turns to blues.
The shoulders we stand upon are strong and enduring
They inspire us to be always curious and learning.
To teach us inclusivity and how to pay respect,
How to tame our ego in the name of everyone’s best.
So here’s to the coders, who make it their quest,
To bring us the best, and put us to the test.
With every line of code, a better tomorrow,
Open source software, a joy to borrow.
Can you spot the bot? Can you find the human-created parts? Or perhaps it was all created by a human or a bot? (Pro tip: When you publish something that was created by or aided by a chatbot, always declare that upfront. Because you are NOT the only author and proper and complete attribution is key to integrity.)
The Hidden Superpower: You
A weakness in all these chatbots is the absence of a human touch. The answers, to me, as a self-ascribed “word nerd” are a bit empty, shallow, lacking warmth. Some of the responses and phrasing are quite puzzling. Sure, they’ll get better with practice, but for now, I prefer the “humanity.” The superpower of the written word, as crafted by a human, is that the person behind the pen provides the clever turn of phrase, the assignment of voice and tone, natural pacing, and tailors the piece to capture the moment and mood. Bots can’t do that — not yet. Nor should they be expected to.
And what’s the superpower of open source? It’s not the code. It’s the people who create the code. It’s the undeniable energy, the thrill of collaboration and joint discovery. Without the people, the notion of open source community ceases to exist. Even the bots know that. Take another look at these lines in the “Ode to Open Source Software:”
A collaboration of minds, a community so bright,
Bringing innovation, with code that’s just right.
So, let’s celebrate the advancement of AI/ML and the new chatbots, whether it’s for creating the written word, song lyrics or even basic code components. It can certainly help us get started, but it should be the just first step and not the only step. Use with care and eyes wide open. Let’s not lose sight of the people who got us here, keep us here and show the way forward. Power to the people!
Note: Thanks to ChatGPT, Tim Pepper, Suzanne Ambiel, Ivana Atanasova and Velichka Atanasova for the “Ode to Open Source Software” contributions.
Answer key:
Stanzas 1-3 ChatGPT
Stanzas 4 – 7 Team OSPO at VMware
Stanza 8 ChatGPT
Stay tuned to the Open Source Blog and follow us on Twitter for more deep dives into the world of open source contributing.
