Alex Lau

Software Engineer @ Aicadium 🧑‍💻 | Born and raised in 🇭🇰 | Currently based in 🇸🇬

[Degen Talk @degentalk_hk]

【聊聊 Polymarket 一個「暗黑」但可以大幅提高勝率的策略】

要在 @Polymarket 長期賺錢，除了看資訊差，更要看“操縱成本”，也就是「改變結果的成本」是否低於「賭局的潛在利潤」

舉個例子，如果賭盤是「BTC 下一小時能否突破 $93,000」，，那最簡單的就是去觀察一下要把BTC 價格拉到 93,000 的成本是多少。如果在Polymarket 上的賭注盈利大於拉盤成本，那這樣就有利可圖。

但考慮到BTC 這麼龐大的市場，你要操縱這個結果，可能需要幾億美金的資金去拉盤，而Polymarket 上的流動性頂多讓你下幾十萬 U。

按照這個思路，我們可以把目光轉向流動性差的「長尾市場」。例如最近 Polymarket 上有個賭盤：

「寶可夢卡牌 Kabuto 1st Edition 年底前能否突破 $100？」

當前數據：
- 地板價 (Floor Price)： $38
- 賭盤賠率： "Yes" 的價格是 $0.40 (也就是 40% 機率)，賠率 1:2.5。

經過簡單計算，要把市場上所有低於 $100 的掛單全部「掃貨」(Sweep the floor)，成本不到 $10,000 U。

如果你在 Polymarket 上下了 $50,000 的重注買 "Yes"，預期回報是 $125,000 (淨賺 $75k)。

只要你的 「掃貨成本 ($10k)」 < 「賭局淨利潤 ($75k)」，這就是一個 +EV 的必勝策略。你買回來的卡片甚至可以之後慢慢賣，稍微虧本也無所謂，因為賭局的大頭已經賺到了。

這不是一個必勝策略，只是提供一個視角，如何可以提高勝率。現在 40% 的賠率明顯是低估了「人為干預」的可能性。

[Alex Lau @AlexLau314]

@eugeneyan I just started using Claude code recently and am curious to know about the workflow from others. Would you rather accept all of their change by default? Or supervising each change and tool call one by one?

[Emm | scenario.com @emmanuel_2m]

How I generated a high quality 3D model in Blender, from a simple 2D image, using Claude 3.7!

Screen capture + link + step-by-step process below👇

[Emm | scenario.com @emmanuel_2m]

"Blender MCP" vs. "Generative 3D" – The Showdown ⚔️

Side-by-side comparison of two generated buildings:

⬅️ The 1st one (left) was created using Blender MCP (powered by Claude 3.7).
➡️ The 2nd one (right) was generated with a generative 3D model (direct 2D-to-3D).

Check out the differences below 👇

[pedma @pedma7]

ChatGPT now helps you backtest Simple Trading Strategies.

No more wasting 100's of hours building code from scratch for bad ideas.

Here’s how to do it for free, in less than 10-minutes:

[Sakana AI @SakanaAILabs]

Sakana AIは、顧客へのサービス提供を開始するため、「事業開発本部」を立ち上げました。最先端のAI技術を使って、日本企業のビジネス課題や、防衛や公共サービスなど社会的課題の解決に注力していきます。大規模な採用も開始します。どうぞご期待ください！

https://t.co/HTqDY7UteQ

[Sakana AI @SakanaAILabs]

Introducing The AI Scientist: The world’s first AI system for automating scientific research and open-ended discovery!

https://t.co/jC7g5GPVsE

From ideation, writing code, running experiments and summarizing results, to writing entire papers and conducting peer-review, The AI Scientist opens a new era of AI-driven scientific research and accelerated discovery.

Here are 4 example Machine Learning research papers generated by The AI Scientist.

We published our report, The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery, and open-sourced our project!

Paper: https://t.co/lTQ8UenFHk
GitHub: https://t.co/Im53whVeAq

Our system leverages LLMs to propose and implement new research directions. Here, we first apply The AI Scientist to conduct Machine Learning research. Crucially, our system is capable of executing the entire ML research lifecycle: from inventing research ideas and experiments, writing code, to executing experiments on GPUs and gathering results. It can also write an entire scientific paper, explaining, visualizing and contextualizing the results.

Furthermore, while an LLM author writes entire research papers, another LLM reviewer critiques resulting manuscripts to provide feedback to improve the work, and also to select the most promising ideas to further develop in the next iteration cycle, leading to continual, open-ended discoveries, thus emulating the human scientific community. As a proof of concept, our system produced papers with novel contributions in ML research domains such language modeling, Diffusion and Grokking.

We (@_chris_lu_, @RobertTLange, @hardmaru) proudly collaborated with the @UniOfOxford (@j_foerst, @FLAIR_Ox) and @UBC (@cong_ml, @jeffclune) on this exciting project.

[elvis @omarsar0]

my new favorite coding workflow:

gpt-4.5 for brainstorming and planning

claude 3.7 sonnet for building

windsurf for all the agentic stuff

watch the video for a quick example

[abhishek @abhi1thakur]

I'm thrilled to announce my new role at Arcee AI, where I will primarily be working on enhancing the AI features of their products: Orchestra and Conductor; and research and development of new features and small language models! As always, my focus will be no-code solution to create AI workflows and to make everyone's life easy when it comes to training or using AI models and building AI Agents 🚀😉

Arcee has a world-class team of researchers and developers who have built:
🎼 Arcee Orchestra: Custom AI workflows. When tasks get specific, LLMs fall down. With Arcee Orchestra, you create customized AI agents to handle specific processes and data that deliver real business results.
🎯 Arcee Conductor: Conductor intelligently routes any query to the best model, efficiently delivering precise results for any task. The right models for the right use case, every time.
🤖 SLMs: State-of-the-art small language models built for every job. Small, specialized, and secure language models are the optimal solution for the majority of business use cases across industries like healthcare, financial services, legal, education, and more.
🔗 Model Merging: Blend models, skip the training. by combining two or more neural network models into a superior, unified model – retaining the strengths or qualities of each.
⚡️ Spectrum: Turbocharge model training. A novel training methodology pioneered by Arcee AI researchers and built into our Small Language Model (SLM) training platform from the ground up.
💥 Supernova: 70B parameters of pure power.
And much more!

Excited to dive in and shape the future of AI! Let’s connect! 🔥

And yes, we are hiring!

[Awni Hannun @awnihannun]

I'm committing to only use local LLMs for the next few weeks to get a real vibe-check on the gap in perf between closed/server-side and open/local (powered by MLX of course).

My favorite tools for that right now are:

- The raw terminal (mlx_lm.generate / mlx_lm.chat)
- LM Studio

[xyzeva @xyz3va]

how to gain code execution on millions of people and hundreds of popular apps
and of course, firebase was (partially) the cause

https://t.co/U7j7YcYS18

[Ki Young Ju @ki_young_ju]

We’re in the late stage of the #Bitcoin bull market, but I believe there’s still room for growth.

I’d say we’re in the early distribution phase, as new retail investors are entering. Trump’s global promotional impact could extend this bull run for another couple of quarters.

----------
Typical BTC distribution:
Whales → Retail Investors

This cycle:
Retail Investors (OG) + Whales (OG) → Retail Investors (Paper Bitcoins) + Whales (Institutions)
----------

OGs leave footprints through on-chain activity and crypto exchanges, while paper Bitcoin (ETFs, corporate stocks) leaves custody wallet on-chain footprints at settlement.

----------
Final phase of distribution:
Retail Investors (OG) + Whales (OG) + Whales (Institutions) → Retail Investors (Paper Bitcoins)
----------

I expect this won't happen until at least mid-year. It might even extend into next year.

[Anthropic @AnthropicAI]

A few researchers at Anthropic have, over the past year, had a part-time obsession with a peculiar problem.

Can Claude play Pokémon?

A thread:

[Alex Lau @AlexLau314]

@asxn_r @HyperliquidX What is assistance fund?

[ltrd @ltrd_]

OK, the reaction for last analysis surprised me. It seems that I have a lot more to share you in next weeks. What breakdown in-depth analysis would you love to see?

[Fede Cardoso @cardosofede]

Initial balance: 51 USD 💸
Total Traded Volume: 25600 USD 🏦
Trading time: 3:21 hours ⏰
PNL: not important for this demonstration but you can see it in the video.
What's interesting is that I paid 5.12 USD in fees in 3 hours, if I have 5 fees or rebates will be all profit 🤖

[Prajwal Tomar @PrajwalTomar_]

I’ve been thinking of building an MVP from idea to deployment using AI tools like Lovable, Cursor AI, etc.

I’ll share the entire process, from product planning and UI/UX to deployment, showing how I build MVPs FAST for my clients.

Would you be interested in watching this?

It’ll take a good chunk of my time, so I’ll only do it if enough people are interested. Otherwise, I’ll focus on my agency hahha!

[Grindery | The currency of AI @grindery_io]

We're thrilled to unveil our first Quest for our community: Grindery Road to 100k on X @Galxe 🎉

📅 Duration - 7 days
🏆 Exclusive Loyalty Points
💎 Complete the quest to be eligible for a portion from the special pool of Grindery $G1 tokens!

Join here: https://t.co/obsiBPjgCj

[Alex Lau @AlexLau314]

@bera_buppets @berachain 🐻