I've been fortunate to be a judge for numerous hackathons, and my recent trip to ETHGlobal Mumbai gave me time to reflect on where hackathons are heading and what we might see next as the year comes to an end.
Thinking back to my own experiences as a participant, hackathons usually meant three hectic days piecing together software to solve a problem. I would say around 80% of the time was spent coding, often with very little sleep.
me before coding agents
But now, with coding agents, the amount of time and effort needed to build a working prototype is shrinking rapidly.
Does this mean that hackathons, especially those focused entirely on software, are becoming less relevant?
Personally, I don't think so.
I think it gives builders an opportunity to spend their time differently. More time understanding the problem, experimenting with different approaches, and having conversations with people who can challenge their thinking.
A working prototype becomes easier to build but figuring out whether it solves a problem worth solving still takes thought.
So what makes a project worth building?
As more builders gain access to the same coding tools, I think understanding the problem becomes even more important.
Who needs your product? How are they dealing with the problem today? What would make them choose what you are building?
These questions have always mattered. But when so much of the weekend was spent getting the software working, it was easy to run out of time to think about them properly.
Coding agents gives us an opportunity to spend more time on that understanding.
A familiar idea can still have room for an original approach. Sometimes, knowing a particular problem well enough helps you see something that existing products have missed.
That is what I would encourage builders to spend more time exploring.
me figuring out what to do with the time my coding agent saved
More hardware is showing up
Now, this one caught my attention. Hardware.
It is a breath of fresh air to see something physical rather than just software.
Dynaura, built by @Dev_9007, @iamomsalunkhe, @pavvannn, brought a working solar setup, with meters tracking energy and smart contracts handling USDC payments based on what was delivered.
Vitaquack, built by @_krisgg, @Masashi_Ono0611 and @corotvoid, had a robot duck that could buy chai with USDC, within a budget set by its owner. Purchases above its limit required approval.
Mivo, built by @Harshith_Nair_ and @mein_hi_hoon, paired a Raspberry Pi and camera with an AI shopping assistant that could find products, negotiate prices and initiate USDC payments on Arc Testnet.
veris, built by @Umyal_dxt and @adp_alpha, prototyped a camera accessory that signs images at capture and connects them to an onchain record, giving photographers a way to establish where a photo came from.
There were also projects exploring how autonomous machines should be held accountable. Sky Proof, built by @itz_micky06 and @kira_906, used a drone simulation to demonstrate a black box for flight records, with USDC compensation paid from an operator's deposit when a breach was found.
It made me wonder whether coding agents are part of the reason. With software becoming easier to prototype, perhaps builders have more room to experiment with how it connects to the physical world.
Tokenisation isn't just putting things onchain
Another theme that stood out was what teams were doing with tokenised assets.
The part that interests me most is what builders can do once those assets are onchain.
we put it onchain. now what?
Earlyout, built by @Satyam_S1603, @Shreyas__Thakur and @dextroalien, demonstrated a way to sell a pending fund redemption for USDC before the payout date. Someone else takes on the wait and the associated risks.
Inphinity, built by @Ramitphi, used test versions of tokenised stocks to prototype a platform where people could create their own index funds, choose the assets and weights, and let others invest with USDC.
One of the reasons stablecoins are where they are today is programmability. Developers can build rules around how money moves and when payments happen.
I think that same potential could help explain why tokenised assets are gaining traction. With stocks and funds onchain, builders can define how those assets are combined, transferred and used across applications.
To me, I think we are still in the early innings of exploring what we can build with smart contracts.
Keep building after the hackathon
I believe builders participating in hackathons should come with a long-term mindset, or take the opportunity to make progress on a problem they already care about.

Learn about different technologies, understand how they can help solve your problem, and use this opportunity to connect with the teams behind them.
Having those teams in the same room is valuable. You can ask questions about your approach, get feedback on what you are building, and discuss the parts you are struggling with.
If you are already working on a project, use the weekend to test a different approach or make progress on a feature. If you are starting with a new idea, use it to find out whether the problem is worth continuing to work on.
Ideally, you leave knowing what to try next, and with people you can follow up with as you build.
So, what's next for hackathons?
I don't think AI makes hackathons less relevant. I think it has raised judges' expectations of what builders can do over a weekend.
If coding agents give you more time, I would love to see that time go into understanding the problem, trying different approaches and talking to people who might actually use what you are building.
Perhaps that means bringing hardware into the mix, finding new uses for tokenised assets, or experimenting with smart contracts in ways we haven't tried yet. There is still so much room to explore.
For those who have yet to try it, Arc Studio can help you keep experimenting and building on your ideas. Try installing the CLI and delegating your onchain and smart contract work to it from your coding agent.
Thank you again to all the builders who took the time to share their projects, ideas and questions with me in Mumbai, and to the team for making the week happen. Those conversations are a big part of why I enjoy attending these events.
These are really exciting times, and I'm looking forward to seeing what builders come up with next.