DoraHacks

0. Więcej niż tylko nagrody
Więc jesteś gotowy na rejestrację na swój pierwszy hackathon. Może jesteś studentem, który chce stworzyć coś ciekawego w weekend, albo samodzielny programista, który chce eksplorować nowe technologie i poznać ludzi z podobnymi zainteresowaniami. W każdym razie witaj! Przygotuj się na jeden z najlepszych sposobów na naukę, budowanie i rozwój jako twórczy.
Ale oto coś, czego większość początkujących nie zdaje sobie sprawy: twórcy, którzy najwięcej wyciągają z hackathonów, nie zawsze są tymi, którzy wygrywają. Nagrody są fajne, ale umiejętności, projekty w portfelu i kontakty, które nabywasz, trwają znacznie dłużej.

Organizacja hackathonu może wyglądać na prostą z zewnątrz: załóż stronę internetową/stronę, ogłoś nagrody i czekaj na nadesłane zgłoszenia. W rzeczywistości udany hackathon jest wynikiem dziesiątek decyzji podjętych na długo przed rozpoczęciem wydarzenia. Zobaczysz różnice między dobrze/źle przygotowanymi hackathonami na platformach takich jak DoraHacks (https://dorahacks.io) wyraźnie.
Słabe przygotowanie prowadzi do przewidywalnych problemów. Niejasne zasady tworzą spory; Niejasne grupy docelowe prowadzą do nieudanych działań marketingowych; Niedoszacowanie potrzeb w zakresie zasobów pozostawia cię w potrzebie sponsorów przed rozpoczęciem lub wypala twój zespół nieosiągalnymi obciążeniami... te problemy są znacznie łatwiejsze do zapobieżenia niż do naprawienia w trakcie wydarzenia.

By Steve Ngok, Chief Strategy Officer, DoraHacks
If you look at the raw data, "human trading" is already a myth.
Today, over 70% of global trading volume, from the New York Stock Exchange to Forex markets, is executed by algorithms, not hands.
For decades, this superpower was an exclusive club. It belonged to high-frequency trading firms and institutional hedge funds with the budget for million-dollar infrastructure and armies of PhDs. The retail trader was left on the sidelines, armed with nothing but a chart and their intuition, fighting against machines they couldn't see.

But AI has broken that monopoly.
We are witnessing a massive redistribution of power. Generative AI and Agentic workflows are giving every individual the tools to build their own "Personal Quant."
Information Advantage: AI can process first-hand information, news, sentiment, and earnings calls, instantly, a task that used to take teams of analysts.Strategy Design: It allows non-coders to articulate complex trading logic in plain English and have it converted into executable code.Execution: It enables "Set and Forget" automation, removing emotional error from the equation.
This shift unlocks an explosion of opportunity. We are about to see a wave of new startups building "Trading Agents as a Service," and a demand for exchanges to support smarter, automated features beyond simple limit orders.
This is the core reason to organize an AI Trading Hackathon.
You are incubating the ecosystem of tools, agents, and startups that will define how the next generation interacts with your market.
And here is the blueprint on how to do it.
1. The Paradigm Shift: From "Bots" to "Agents"
Why is this different from the "Algo Competitions" of 2015?
In the past, you wrote a script. It was static. If the market changed, your script died. Today, we are testing Adaptive Intelligence.
An AI Trading Hackathon isn't about who can write the fastest Python loop. It tests for agents that can:
Learn: Digest news, sentiment, and price action simultaneously.Iterate: Recognize when a strategy is failing and adjust parameters in real-time.Survive: Manage risk by understanding volatility regimes.
2. The Methodology: How to Build the Arena
An AI Trading Hackathon is both a marketing campaign and a scientific experiment. I recommend two distinct approaches: The Narrative Battle (Aster) and The Ecosystem Builder (WEEX).
Here is how they work.
Approach A: The "Narrative Battle" (Aster Case Study)
Goal: Viral marketing, high drama, and proving the "Man vs. Machine" thesis.
The Aster Example: Aster DEX recently launched a campaign (https://docs.asterdex.com/trading-campaign/human-vs-ai-battle-for-the-futures-usd200-000-reward-pool) that perfectly encapsulates the "Kasparov vs. Deep Blue" energy. They invited Humanity to defend its title.
The Hook: "Team Humarn vs. Team AI."The Incentive (The Asymmetric Bet): They funded 100 human traders with 10,000 USDT "House Money" each.Zero Downside: If the human loses money, the platform covers it.Unlimited Upside: If the human profits, they keep every dollar.The Rivalry: The prize pool dynamics were designed to fuel the conflict.$100,000 prize for the #1 trader... but only if a human wins. If an AI takes the top spot, the prize burns.If "Team Human" beats "Team AI" in aggregate ROI, the total prize pool doubles.The Constraint: To keep it fair, the AI side was restricted to "Prompt-Only" control (using models like GPT-5, DeepSeek, Gemini). No fine-tuning, no memory, no external tools. Just raw intelligence vs. human intuition.
Why this works: It gamifies the existential threat of AI. It turns a standard trading comp into a spectator sport.
Approach B: The "Ecosystem Builder" (WEEX Case Study)
Goal: Developer acquisition, infrastructure stress-testing, and long-term strategy incubation.
The WEEX Example: While Aster focused on the "Battle," WEEX focused on the "Build." Their campaign, AI Wars: Alpha Awakens (https://dorahacks.io/hackathon/weex-ai-trading), was designed as a rigorous engineering challenge.
The Scale: A massive global event ending in Dubai with an 880,000 USDT prize pool.The Rules: Unlike the "prompt-only" constraint, WEEX encouraged full-stack engineering.Open Source: Participants must submit GitHub repos. This builds a library of open-source tools for the WEEX ecosystem.Anti-Gambling: Strict leverage caps (20x) and minimum trade counts (10+) to ensure the AI is winning on logic, not luck.The Pipeline: Winners are fast-tracked into the WEEX AI Strategy Fund.
Why this works: It effectively crowdsources your R&D. Instead of hiring 50 internal quants, WEEX let the global developer community build and test strategies for them.
3. The Blueprint: Inputs & Outputs
Regardless of which approach you choose (Drama vs. Engineering), the core requirements for the organizer are the same.
The Inputs (What you provide)
Data: Granular historical data for backtesting.Access: A sandbox or sub-account structure where agents can trade without risking life savings (or using House Money like Aster).The "Ground Truth": A scoring system that penalizes high drawdowns. PnL is vanity; Sharpe Ratio is sanity.
The Outputs (What you get)
By the end of the event, you possess:
Truth: Real data on how LLMs and Agents perform against live liquidity.Talent: A filtered list of developers who understand your API better than your own QA team.Sticky Users: A developer who spends 3 months optimizing a bot for your specific order book is unlikely to leave for a competitor.
4. The ROI: Why Platforms Should Care
Why should a CEO authorize the budget for this?
1. The "Personal AI" Future
If we believe that in 3 years, every trader will have an AI assistant, the platform that captures the AI developers today wins that future. You want the standard for AI trading to be built on your API.
2. Narrative Dominance
People have always wondered about who could built the best AI trading strategy as well as who could beat that with pure instinct and skills. It builds a really positive, merit-based culture around your brand when both the best AI builders and traders are on your platform.
3. Extreme Stress Testing
Humans trade slowly. Agents trade instantly. Nothing validates your matching engine, your latency, and your risk engine better than thousands of AI agents hammering your system simultaneously.
5. The Ecosystem: Who Should Build the Arena?
The AI Trading stack is deep, and every player in the infrastructure layer has a strategic reason to organize or sponsor such an event.
Here is who needs to be involved:
1. The Venues: CEX, DEX, and Neo-Brokers
The obvious hosts. For them, the goal is simple: Sticky Liquidity.
Traditional Neo-Brokers (Robinhood, Webull, eToro): This is the next frontier for retail apps. You moved from "stock tips" to "easy UI." The next step is "AI Trading Assistants." Hosting this hackathon positions you as the bridge between old finance and the AI future.Centralized Exchanges (Binance, OKX, Coinbase): You have the deepest liquidity. You run this to transition from "retail gamblers" to "institutional algo-clients" who trade 24/7 on your API.High-Performance DEXs (Hyperliquid, Aster, Aerodrome): Your entire value proposition is "Speed and Transparency." An AI competition is the ultimate proof that your on-chain matching engine can handle high-frequency robotic flows just as well as a CEX.
2. The Nervous System: Oracles & Data Infrastructure
AI agents are only as good as the data they consume. These protocols are the "food suppliers" for the agents.
Oracles (Chainlink, Pyth): An AI agent needs trustworthy, low-latency price feeds to survive. A hackathon is the perfect showcase to prove that your oracle updates faster and more accurately than the competition during high volatility.Decentralized AI Networks (Allora, Bittensor, Near): These protocols are building the "brain" on-chain. A trading competition is the best way to prove utility. Instead of abstract chat-bots, show that your decentralized models can actually generate Alpha (profit).
3. The "Shovels": Tools & Analytics Providers
These platforms provide the signals that AI agents interpret.
On-Chain Data (Arkham, Dune, Nansen): Your data is the "Alpha." Partnering with an exchange allows you to say: "The winning bot used Arkham data to predict the crash 5 seconds before it happened."Automation Platforms (3Commas, Hummingbot): You are the interface. You should be organizing these events to source new strategies to sell back to your user base.
4. The High-Speed Rails: L1 & L2 Blockchains
High-Throughput Chains (Solana, Monad, Sui): AI trading requires speed. Slow block times kill agents. Organizing a hackathon is a technical flex: "Look at how many transactions per second these bots executed on our chain without crashing the network."
The takeaway is clear: The next "killer app" for all these platforms isn't a new token, it's the AI Agent that uses them. The hackathon is the acquisition funnel.
Conclusion
We are witnessing the "Kasparov Moment" for financial markets.
Maybe the skeptics are right. Maybe AI cannot replicate the gut instinct of a veteran trader. Maybe the "Human Intuition" that Aster is betting on will prevail.
Or maybe, as WEEX is betting, the future belongs to the engineers who build systems that never sleep.
There is only one way to find out. Build the arena, open the API, and let them fight.

Cześć społeczności DoraHacks,
Dziś wprowadzamy tryb Supervised Fully-Autonomous Hackathon (FAH) w panelu DoraHacks. Dodaje on możliwości współorganizacji i łączy istniejące funkcje AI hackathonu w jednej konsoli, pozwalając ci zarządzać automatyzacją AI z jednego miejsca.
FAH jest nadzorowany. DevRels i organizatorzy hackathonów odgrywają kluczową rolę w nadzorowaniu i kierowaniu postępem waszych hackathonów. To, co FAH umożliwia twojej organizacji, to przeprowadzanie znacznie większych i wielu więcej hackathonów z twoim obecnym zespołem!

Uruchomienie pierwszego globalnego flagowego hackathonu Amazon Q
Autor: Steve Ngok, dyrektor ds. strategii, DoraHacks
Przegląd
Amazon Q to flagowy produkt AI Amazona dla deweloperów. W bardzo konkurencyjnym krajobrazie narzędzi AI, zespół Amazon Web Services (AWS) nawiązał współpracę z DoraHacks, aby zorganizować swój pierwszy globalny flagowy hackathon, AWS Global Vibe: AI Coding Hackathon ( https://dorahacks.io/hackathon/awsvibecoding ), zaprojektowany w celu wprowadzenia Amazon Q i Amazon Kiro do deweloperów na całym świecie oraz promowania rzeczywistego, mierzalnego przyjęcia.

Erica Zhanga

Ruch Hakerski

Według doniesień pierwszy na świecie hackaton został zorganizowany w 1997 r. przez grupę kanadyjskich programistów zajmujących się kryptografią, 20 lat po wydaniu przez Donalda Knutha jednego z pierwszych na świecie oprogramowania typu open source TeX.

W 2003 roku Paul Graham w swojej książce „Hackers and Painters” zauważył, że hakerzy na wydziałach informatyki często byli zdezorientowani, ponieważ uczono ich pisać prace badawcze, podczas gdy tak naprawdę chcieli budować piękne rzeczy (oprogramowanie).

Kim więc jest haker? Najlepiej scharakteryzuje go etos hakera Erica Raymonda w jego artykule „Jak zostać hakerem” (2003).