Mansour Fall

[Carver @carverfomo]

A 13 year old kid from Shenzhen spent his entire winter break in his bedroom while every classmate was outside playing. He sat in front of a MacBook Air for 21 days straight watching Claude tutorials and trying to copy what he saw on screen.

By the end of the break he had built his own AI agent that solves Codeforces problems in 45 seconds each. He uploaded the whole thing to a public GitHub repo on the last day of vacation. README in broken English. Description: ai agent solves codeforces fast for me. Watched it sit at zero stars. Went back to school the next morning.

While the West argues about whether AI will replace coders, China is already letting 13 year olds publish working agents on GitHub between school terms. He thought he was building a homework helper. He just showed too much.

Six months later his computer science teacher opened his GitHub profile to check on a class assignment. The repo the kid had pushed at the end of winter break had 3,100 forks. The teacher clicked through to one of the forks and found a wallet sitting at $4,526,176.

432614799197. $4,526,176 profit. Joined January 2026.

→ https://t.co/Xv5TKvZUvu

A 27 year old developer in Singapore had cloned the kid's repo, swapped the Codeforces plugin for a sports betting one, pointed Claude at Asian bookmaker odds and walked away. Six months of green logs later he posted the profit curve to a small dev forum with one line of caption: I did not write a single line of this code. The original repo was a kid's homework project.

Dev Twitter exploded. 180K views in 24 hours. Everyone asking for the GitHub link. Everyone wanted to know who wrote the original.

The Singapore guy told them. That was the mistake.

Pause at 0:08 of his demo video. Look at the contributor name in the top right of the GitHub page.

That contributor is 13 years old.

The comment section turned into a detective board. Someone slowed the demo to 0.25x. Someone else opened the original repo and pulled the contributor history. A third guy traced the contributor's other commits and found a school programming contest entry from Shenzhen Number 4 Middle School with the same username.

The kid had won third place. The judges wrote that his project was creative but had no real world application.

The Singapore guy is still running the script. 4,548 trades since January. All sports. Across six leagues. NFL, Premier League, Champions League, La Liga, Ligue 1, NHL.

Biggest single win on the wallet: $1.5M. From one football match.

The repo is not the product. The repo is the noise.

The real engine is what the kid built and the judges missed. Asian bookmakers post lines 2 to 3 hours before Western platforms update. Shanghai moves before London. Beijing moves before New York. By the time American traders open their laptops the gap is already closed. The Claude agent catches every one of them while the West is still asleep.

The kid wrote the script over winter break. The Singapore guy cloned it for free. The market gave back $4.5M.

The Singapore guy deleted his post when he realized what he had revealed. Too late. The repo had already been forked 3,100 times. Someone in the kid's school recognized the username.

The kid still has not made a single dollar from the script. The terms of service on his GitHub account say he is not allowed to sell anything. He is too young.

The teacher who opened the profile to grade homework closed the laptop and called the principal instead.

Before the post got buried, one reply got pinned: the smartest builder in this story is the one who is not allowed to use what he built.

[ZAYVEN KNOX @ZayvenKnox]

ANOTHER QUIET POLYMARKET ACCOUNT JUST TURNED $50 INTO $625,000.

Claude rebuilt the strategy in 40 minutes, and the edge was simple: hit BTC contract lag before Polymarket catches up.

You only need Claude + laptop + 1 hour/day.

Giving This Free for 24 hours. To get it:

1. Comment the word 'CLAUDE'
2. Like and Retweet this post
3. Follow me
@ZayvenKnox
(so i can DM you)

[Mansour Fall @moumaf]

@ZayvenKnox @tomreiai Claude

[Alexandr Wang @alexandr_wang]

1/ today we're releasing muse spark, the first model from MSL. nine months ago we rebuilt our ai stack from scratch. new infrastructure, new architecture, new data pipelines. muse spark is the result of that work, and now it powers meta ai. 🧵

[Mansour Fall @moumaf]

@marryevan999 Polymarket

[Marry Evan @marryevan999]

THIS GUY MAKES $1,000 A DAY BY WAITING FOR THE LAST FEW SECONDS ON POLYMARKET.

No predictions, no drama, just buying 99 cent contracts when the outcome is already basically decided.

Giving This Free for 24 hours. To get it:

1. Comment the word 'PolyMarket'
2. Like and Retweet this post
3. Follow me @marryevan999 (so i can DM you)

[Mansour Fall @moumaf]

@Danny_Crypton BTC

[Mansour Fall @moumaf]

@AndrewBolis Send

[Mansour Fall @moumaf]

@tomreiai Guide

[Mansour Fall @moumaf]

@Shruti_0810 Send

[Mansour Fall @moumaf]

@JobFound5 Interested

[Mansour Fall @moumaf]

@AndrewBolis YT

[Mansour Fall @moumaf]

@AndrewBolis Money

[Chris Paxton @chris_j_paxton]

I absolutely believe real world RL is the future

[Elon Musk @elonmusk]

As I was saying …

[Fuli Luo @_LuoFuli]

MiMo-V2-Pro & Omni & TTS is out. Our first full-stack model family built truly for the Agent era.

I call this a quiet ambush — not because we planned it, but because the shift from Chat to Agent paradigm happened so fast, even we barely believed it. Somewhere in between was a process that was thrilling, painful, and fascinating all at once.

The 1T base model started training months ago. The original goal was long-context reasoning efficiency. Hybrid Attention carries real innovation, without overreaching — and it turns out to be exactly the right foundation for the Agent era. 1M context window. MTP inference for ultra-low latency and cost. These architectural decisions weren't trendy. They were a structural advantage we built before we needed it.

What changed everything was experiencing a complex agentic scaffold — what I'd call orchestrated Context — for the first time. I was shocked on day one. I tried to convince the team to use it. That didn't work. So I gave a hard mandate: anyone on MiMo Team with fewer than 100 conversations tomorrow can quit. It worked. Once the team's imagination was ignited by what agentic systems could do, that imagination converted directly into research velocity.

People ask why we move so fast. I saw it firsthand building DeepSeek R1. My honest summary:

— Backbone and Infra research has long cycles. You need strategic conviction a year before it pays off.
— Posttrain agility is a different muscle: product intuition driving evaluation, iteration cycles compressed, paradigm shifts caught early.
— And the constant: curiosity, sharp technical instinct, decisive execution, full commitment — and something that's easy to underestimate: a genuine love for the world you're building for.

We will open-source — when the models are stable enough to deserve it.

From Beijing, very late, not quite awake.

[Brett Adcock @adcock_brett]

Today, Figure is showing another major milestone towards a robot in every home

Running Helix 02, cleaning a living room fully autonomously

[Nic Cruz Patane @niccruzpatane]

NVIDIA CEO Jensen Huang today: “The Chat-GPT moment for self-driving cars has arrived.

NVIDIA announced that they’re now expanding their self-driving partnership to BYD, Nissan, Hyundai, and Geely.

Their automotive partners:
• GM
• Toyota
• Mercedes-Benz
• Jaguar Land Rover
• Volvo
• Rivian
• Hyundai
• BYD
• Geely
• XPENG
• Polestar
• NIO
• Lucid
• Li Auto
• Nissan
• Isuzu
• Zeekr
• Xiaomi
• Stellantis

Additionally, NVIDIA and Uber are partnering to launch a global network of Level 4 autonomous robotaxis, beginning in LA and San Francisco in early 2027 before expanding to 28 cities by 2028.

The robotaxi race is well underway.

[NVIDIA GTC @NVIDIAGTC]

AI factories are moving from build-first to simulate-first.

With NVIDIA DSX Air, organizations can validate complex AI infrastructure deployments in simulation—reducing time to first token from months to days.

DSX Air brings together the AI factory ecosystem, enabling partners across servers, networking, storage, orchestration, and security to test and validate solutions alongside NVIDIA infrastructure.

🔗 Learn more: https://t.co/QyVf5TgFro
#NVIDIAGTC

[Mistral AI @MistralAI]

🚀Announcing a strategic partnership with NVIDIA to co-develop frontier open-source AI models, combining Mistral AI’s frontier model architecture and full-stack AI offering with NVIDIA’s leading compute infrastructure and development tools.