Hear from a panel of developer marketing experts on AI’s impact on their hackathons.
Devpost hosted a panel discussion at the 2023 Developer Marketing Summit on how AI is affecting hackathons. From discussing the value of hosting hackathons in today’s rapidly transforming landscape, to hackathon best practices and using responsible AI, read below for a summary of the event. (Watch the full-length recording of their conversation here.)
Meet the speakers:
To kick off the conversation, the panelists shared how hackathons provide strategic value to their organizations. Some of the outcomes they highlighted included the projects submitted during hackathons, as well as the ripple effect they have in terms of reach and impact. It was also mentioned that hackathons provide internal customers, like product and engineering teams, with feedback from developers they otherwise wouldn’t get. Speeding up product testing is another key reason organizers host hackathons.
Karen: “When we are going to our executive stakeholders, we can say at the end of the event, ‘We predict that this hackathon is going to create this many new developers on our platform, this many new apps.’ So the metrics are very, very closely tied to our company's top-level goals.”
Chukwuemeka: “Hackathons are a great way to bring a lot of the developer relations and activities together in a steel thread, which cuts through so many facets from awareness, activation, and to adoption.
So if you orchestrate all these things properly, you're able to drive good metrics around each of these aspects. That's the way I've always liked to think about hackathons. It's not just about the actual building that happens. It's about all the stuff that is aligned with it.”
From a product perspective, Jayson also highlighted how hackathons help significantly accelerate product testing.
Jayson: “You want a hackathon as a validation strategy. In one of the first hackathons we sponsored when launching a developer platform for Dolby, we went to the Capitol Records Tower and took APIs.
All the APIs we took to that hackathon—none of them exist anymore. It was because we used that as a learning opportunity for how would people adopt. That was the return on that investment.
That was immediately a huge roadmap time saver. All those APIs got repackaged in a very different way and that became the product we launched.”
Brandon shared how at Devpost, we’ve seen that hackathons tend to have a bigger impact on developer marketing than traditional marketing tactics. Roughly 33% of people who register for hackathons go on to build a project. This is much higher than typical marketing metrics like conference attendance, ad clicks, tutorial attendees, etc.
The consensus from all panelists was that AI is definitely on every company’s mind right now. That being said, the way each company is approaching AI is quite different, which is a sign of how early things are in this space.
Most panelists see the obvious use cases of AI tools, like summarizing text, but also feel that we’re just at the beginning and are looking forward to seeing new innovations rise to the top.
Karen: “We've found that for over ten years, our developer community has been this huge source of innovation and creativity, and has helped push our products far beyond what we are building ourselves.
It’s really exciting to use the collective brainpower of all of these thousands of developers who are out there building apps and see what types of use cases are resonating and rising to the top.”
Chukwuemeka pointed out that AI is not a new concept, but that there’s a renewed focus on it now, especially in light of LLMs.
Chukwuemeka: “From a company perspective, there are a number of interesting [AI-related] things coming down the pipeline. The interesting thing is that some of those things came out of an internal hackathon.
(Read more about Okta’s internal hackathon here.)
Chukwuemeka also spoke about a current Okta public hackathon, which he hopes will give insight into how developers can apply AI in the B2B space.
Chukwuemeka: “For us as an identity company, one of the things we are focused on is figuring out how we can help developers who are building in the B2B space to be more enterprise-ready.
So, being more aware of building for the needs of the business to the end user in the business. We believe that there's a role for AI in that.”
While Jayson mentioned that his team might use AI tools that summarize texts or generate insights, he also spoke about Dolby’s caution with using generative AI, since the company works with rights holders.
Jayson: “We respect the creative intent. If you're using generative technology and it is generating something that is infringing on someone's rights—whether it's copyright like a creative work of art or if it's using code that was licensed under a prohibitive license, and using that to inject code into our code base—that becomes disastrous.”
The panelists who have organized AI hackathons acknowledged that these events are different from non-AI hackathons. With hackathons that don’t focus on AI, the goals tend to focus on encouraging developers to build enterprise-ready use cases.
However, with AI hackathons, the goals are largely focused on experimentation and using it as an opportunity to gain deeper insight into how developers are using these technologies.
Karen: “The prompts for our hackathon this year are around making developers' lives better through tools that incorporate AI.
The other thing that's driving this internally is a real desire for knowledge from our product team. As AI is making its way potentially into our developer platform, we want to understand what extension points developers are asking for and what are the access points in our APIs that developers want to have.”
Chukwuemeka also pointed out that the nature of hackathons is changing and that it feels like new technologies are emerging every day.
Chukwuemeka: “[In previous hackathons], you had to submit a model or something like that. Now, I don't even know what people are going to be submitting because there are so many new skill sets coming up. Prompt engineering is now a thing, for example.’
For Okta, he said it’s a balancing act to find new solutions that some larger enterprises might be hesitant to experiment with.
Chukwuemeka: “Because we are in the business of identity and access management for these companies–and we're trying to evangelize the concept of enterprise readiness amongst these developers–we are looking to figure out how can you take an AI solution that you created and make it enterprise-ready.
So we're putting that out there to see how developers react to that and if they are up to making that happen.”
While a lot of general hackathon best practices apply, the panel dove into specific tips for AI hackathons. A few highlights include responsible AI (which is addressed in more detail in the next question), providing participants with relevant resources, and keeping the topic open-ended.
Karen: “Build the guardrails that you want participants to keep in mind into the prompt as you're designing the hackathon.
Other best practices include providing resources, starting code, and example apps. With AI hackathons, there's also a need to point developers to a list of recommended services if those services are not provided by your company.”
Brandon: “Consider encouraging non-developers to participate. If it's gen AI-related, then they can create prompts and that can help both internally or externally.
“Keep it open-ended. As we've heard multiple times up here, given that we don't know exactly what those use cases are, it's one of the beauties of hackathons.
“Ensure that there is business value. That means bringing in other stakeholders to make sure that multiple things come out of this that will address the ROI.”
The panelists spoke about the importance of organizers ensuring that hackathons meet internal compliance requirements and providing participants with resources that would enable them to use AI ethically as well.
Chukwuemeka: “Keep your legal team close and if there's somebody in your organization who is in compliance, keep them in the loop. It's good to do it early because, in this whole AI thing, we're all learning.
So the sooner we engage them and get them in this learning mode together, it will result in a more favorable outcome.”
Karen referenced resources and guidelines around responsible AI that have been provided to participants in Atlassian’s most recent Codegeist hackathon.
Karen: “[It includes] things like disclosure to show where your app is using AI, read the fine print of the AI service you’re using, and make sure that you know what it's doing with user data.
We introduced a responsible AI bonus category in our hackathon. So we are, not just providing these guidelines, but incentivizing developers to go above and beyond and show how they're ethically building.”
Jayson: “Make sure to have that opportunity to educate. So if you're talking about self-driving cars, what is the liability of code that's being generated if that were to make a wrong decision? It's an education opportunity.”
We’re grateful to Karen, Chukwuemenka, and Jayson for participating in this panel! Contact our team if you’d like to know more about AI hackathons.