We did it: Malaysia's largest AI Learn-a-thon

An eight-year-old walked onto a stage and demonstrated a game she had built that day.

Elsewhere in the room, Malaysians in their seventies were making apps for themselves and their communities. Hundreds of laptops were open. Volunteers moved between tables, helping people fix broken deployments, understand unfamiliar tools and get their first project onto the internet.

Across the Learn-a-thon, close to 700 people took part in one mass building session. Together, they produced nearly 90 working apps and AI agents. The Malaysia Book of Records gave the session its official title: Largest Artificial Intelligence (AI) Learn-a-thon.

We had a shorter way to put it.

We did it.

The numbers

More than 3,500 people came through AI Malaysia Takeover over two days.
More than 1,000 people registered for the Learn-a-thon.
Close to 700 people joined the record session.
They built nearly 90 working apps and agents.
Builders ranged in age from eight to over 80.
More than 50 experts and mentors helped people build.

The registration and record-session numbers describe different groups. More than 1,000 people had registered before the Learn-a-thon. Organisers later reported that close to 700 people joined the session counted for the record.

The certificate matters. It proves the scale. But it is not the best part of the story.

Organisers holding the framed Malaysia Book of Records certificate on the AI Malaysia Takeover stage — The official certificate for the Largest Artificial Intelligence (AI) Learn-a-thon.

The takeover was ours

“AI takeover” usually sounds like a warning. Machines take the jobs. Models make the decisions. People get pushed to the side.

That is not what happened at The Campus in Ampang.

For two days, Malaysians took hold of the technology themselves. Students, retirees, developers, founders, office workers, aunties, uncles and children sat in the same room and made things. Some arrived with experience. Many had never built an app.

The point was not to make everyone a software engineer in one afternoon. It was to break the assumption that building technology belongs to somebody else.

The eight-year-old with the game understood that. So did the nearly 80-year-old participant who bought a laptop the day before because he wanted to learn AI. So did the senior builder who went on stage and told the room, “If we can do it, so can you.”

That was the takeover. Malaysians stopped watching AI happen and started building with it.

An eight-year-old participant demonstrating the project she built during the Learn-a-thon — A young builder shows the room what she made.

The room existed before the record

The Learn-a-thon looked sudden because much of it came together in about four weeks. The community behind it had been gathering for much longer.

Malaysian AI began as a place for local builders to meet, work beside one another and show what they had made. The Malaysian AI Residency gave founders and developers a regular base. Weekly Show and Tells gave people a reason to keep shipping. Local groups brought in their own members, tools and ways of teaching.

By the time AIMTO began, the community had already run more than 50 Show and Tells and related events. The people needed to fill a mentor floor were already in the room. The next step was to open that room to the public.

AI Malaysia Takeover 2026 was organised by Ludic Asia and 500 Global, with support from the Ministry of Digital and MyDIGITAL Corporation. The Learn-a-thon brought together the Malaysian AI and Residency communities, KrackedDevs, YTL AI Labs, trainers, founders and volunteers from across the local ecosystem.

Malaysian AI did not make the record alone. Nobody did. Our role was to bring the builder community, help design the experience and put people on the floor who would sit beside a stranger until their idea worked.

How do hundreds of people build at once?

A mass build session can easily become a long lecture with laptops open. We wanted the opposite.

People came in with different skill levels, so there could not be one narrow technical path. Builders could bring an idea or pick a starting prompt. They could work alone, find a team or ask a mentor to help them take the first step.

The floor was organised around a simple journey:

Arrive, connect and get set up.
Choose a problem worth solving.
Start building with an AI tool.
Find a mentor when something breaks.
Debug, deploy and make the project work.
Show other people what you made.

Mentors did not take the keyboard and finish projects for people. They helped builders get unstuck. That distinction mattered. Everyone needed to leave knowing that the thing on screen was theirs.

A Malaysian AI volunteer kneeling beside a participant and helping him work through his project — The human help layer: one volunteer, one builder and one problem at a time.

YTL AI Labs gave participants access to ILMUchat, its Malaysian-developed AI platform. That gave builders a local AI tool to experiment with during the event and for a month afterwards. Other participants used the coding and automation tools they already knew. The tools varied. The task stayed the same: make something that works.

A child using ILMUchat on a laptop during the AI Learn-a-thon — Building with ILMUchat during the Learn-a-thon.

What Malaysians built

The resulting projects were wonderfully specific.

One team made a Duit Raya Planner for organising festive giving. Another made a Traffic Light Simulator to model traffic flow. Builders produced games, websites, personal agents and shared knowledge databases that could supply several AI agents with the same information.

The project list also included tools for travel planning, personal finance, studying, invoices, fitness and family care. These were not abstract demonstrations of what AI might do one day. They came from problems people already had.

Some projects were rough. Of course they were. They had been built in hours by people learning as they went. A finished first attempt is more useful than another year spent believing you are not technical enough to begin.

The people behind the laptops

Large events tend to produce photographs of stages. The work that makes them possible happens elsewhere.

Lizzie Tan and Ashvin Praveen worked through repeated revisions of the Learn-a-thon experience. Malaysian AI and 500 Southeast Asia floor volunteers included Maxine Lai, Xian Jun Chong, Andrew G., Gabriel Yong, Janelle Tan, Tharshen A/L Surian Balan, Mohtasham Murshid Madani, Lynnette Tee, Muhammad Fathy Rashad, Dylan Tan, Khalil Nooh, Andre Teow and Eileen Liew.

The YTL AI Labs team included Shannon Choo, Hong Bing Sim, Han Tham, Daniel Wu, Ian Low, Nicholas Chong Yee Yang, Cody Liew and Edwin Fong. The 500 Southeast Asia team included Amir Zakwan Anuar, Nathaniel Tham, Wafi Suhairi, Sim Tze Hin, Adelyn Chan, Suzi Khizari and Suraya Abu Bakar.

Rachel How stepped in to teach product design. Jing Hui Yeoh and the Ludic Asia team helped run the event. Zen Ben Teo, Charmaine Chee, Aditya Ansh, Gracie T., Bede Lau, Chong Wei Jie, Nur Aleya Natasha Yasir, Shaahid Marikar, Mun Yau Chin, Wasif Hussain, Ziad Ismail, Kristina Timofeeva, Akedil Yerkin, Tam Zhao Kai, Tareeqah Amber, Shaidan Shaari bin Abd Rasep and Wan Zulaikha Nasuha joined the wider volunteer floor.

There were more. People taught, registered participants, carried equipment, answered the same question for the fiftieth time, found spare chargers, fixed Wi-Fi problems and stayed after the room should have emptied.

The photographs on this page keep some of their faces in the story because the record belongs to them too.

AI Malaysia Takeover volunteers gathered around the Malaysia Book of Records certificate — Some of the volunteers who turned a very large room into a place where anyone could build.

A record, then what?

The public record says a lot of people built with AI in one place. It does not tell us whether they will build again.

That is the harder measure.

Malaysian AI continues to run events and weekly Show and Tells where people can bring unfinished work, demonstrate something small and meet others who are figuring it out. YTL AI Labs and KrackedDevs also announced a three-month programme for Learn-a-thon participants who want to continue working with ILMUchat and develop the projects they started at AIMTO.

We want to see the second app. We want the first-time builder to return as a mentor. We want more rooms like this outside Kuala Lumpur, taught by people who understand the problems in their own towns, schools and workplaces.

The record is a photograph of one day. The work is everything that comes after it.

In the media

We did it together

Government opened a national learning platform. Partners supplied space, tools and people. Experienced builders taught first-timers. Volunteers stayed on the floor. Thousands of Malaysians showed up ready to try.

On 12 August 2026, that combination produced Malaysia’s largest AI Learn-a-thon.

The certificate says we hold the record. The room showed why it mattered.

Join the next Malaysian AI event