Offshore
Photo
Offshore
Photo
Illiquid
This is why you are early to Seikoh Giken.
tweet
This is why you are early to Seikoh Giken.
We makes 1.6T from 2026 onwards
We makes 3.2T from 2028 onwards https://t.co/4mCVa2gc5T tweet
Offshore
Photo
God of Prompt
RT @free_ai_guides: I stopped using ChatGPT for daily tasks.
Switched to an open-source agent that:
→ Remembers every conversation I've had
→ Knows my preferences without being told
→ Adds new skills from a community library
→ Runs 24/7 without me touching it
→ Costs almost nothing
It's called OpenClaw.
And it's what AI assistants should have been from the start.
I wrote the setup guide. Free.
Comment "OpenClaw" and I'll DM it to you
tweet
RT @free_ai_guides: I stopped using ChatGPT for daily tasks.
Switched to an open-source agent that:
→ Remembers every conversation I've had
→ Knows my preferences without being told
→ Adds new skills from a community library
→ Runs 24/7 without me touching it
→ Costs almost nothing
It's called OpenClaw.
And it's what AI assistants should have been from the start.
I wrote the setup guide. Free.
Comment "OpenClaw" and I'll DM it to you
tweet
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
RT @douglaskimkorea: In the next several years, 6G networks are expected to be rolled out aggressively in Korea and Samji Electronics could be one of the beneficiaries of this 6G network expansion. https://t.co/OkxBfIwOqg
tweet
RT @douglaskimkorea: In the next several years, 6G networks are expected to be rolled out aggressively in Korea and Samji Electronics could be one of the beneficiaries of this 6G network expansion. https://t.co/OkxBfIwOqg
tweet
Offshore
Photo
DAIR.AI
On evaluating multi-step scientific tool use in LLM agents.
SciAgentGym provides an interactive environment with 1,780 specialized tools across 4 scientific disciplines.
The core finding: even advanced models like GPT-5 see success rates drop sharply from 60.6% to 30.9% as tasks require more interaction steps.
Multi-step execution remains a fundamental bottleneck.
To address this, the researchers developed SciForge, a data synthesis method that models tool interactions as dependency graphs. Their fine-tuned SciAgent-8B outperformed much larger competing models on scientific workflows.
Scientific automation requires reliable multi-step tool use. Targeted training on graph-structured trajectories is more effective than raw model scale for these tasks.
Paper: https://t.co/Z9u1zi5K0U
Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c
tweet
On evaluating multi-step scientific tool use in LLM agents.
SciAgentGym provides an interactive environment with 1,780 specialized tools across 4 scientific disciplines.
The core finding: even advanced models like GPT-5 see success rates drop sharply from 60.6% to 30.9% as tasks require more interaction steps.
Multi-step execution remains a fundamental bottleneck.
To address this, the researchers developed SciForge, a data synthesis method that models tool interactions as dependency graphs. Their fine-tuned SciAgent-8B outperformed much larger competing models on scientific workflows.
Scientific automation requires reliable multi-step tool use. Targeted training on graph-structured trajectories is more effective than raw model scale for these tasks.
Paper: https://t.co/Z9u1zi5K0U
Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c
tweet
Offshore
Photo
Clark Square Capital
RT @ClarkSquareCap: Would love a RT if you found this useful! Thank you!
tweet
RT @ClarkSquareCap: Would love a RT if you found this useful! Thank you!
Here is this week's special situations digest.
281 situations in activist campaigns, M&A/divestments, management changes, and other corporate events.
Make sure to check it out https://t.co/NZBTGqVC6P tweet
Offshore
Photo
DAIR.AI
RT @omarsar0: Interesting new work on adaptive reasoning depth for LLM agents.
Not every agent step requires the same level of thinking. Some steps need strategic planning. Others are routine execution.
This research introduces CogRouter, a framework inspired by ACT-R cognitive theory that dynamically adjusts reasoning depth at each decision step across four hierarchical cognitive levels.
Appropriate cognitive depth should maximize the confidence of the resulting action. Training combines supervised fine-tuning for stable cognitive patterns with policy optimization for step-level credit assignment.
A 7B parameter model achieved 82.3% success rate on agent benchmarks, outperforming GPT-4o while consuming 62% fewer tokens.
Why does it matter?
Adaptive reasoning is a more practical path to efficient agents than simply scaling model size. Think fast when you can, slow when you must.
Paper: https://t.co/kYLqeHaY8p
Learn to build effective AI agents in our academy: https://t.co/1e8RZKs4uX
tweet
RT @omarsar0: Interesting new work on adaptive reasoning depth for LLM agents.
Not every agent step requires the same level of thinking. Some steps need strategic planning. Others are routine execution.
This research introduces CogRouter, a framework inspired by ACT-R cognitive theory that dynamically adjusts reasoning depth at each decision step across four hierarchical cognitive levels.
Appropriate cognitive depth should maximize the confidence of the resulting action. Training combines supervised fine-tuning for stable cognitive patterns with policy optimization for step-level credit assignment.
A 7B parameter model achieved 82.3% success rate on agent benchmarks, outperforming GPT-4o while consuming 62% fewer tokens.
Why does it matter?
Adaptive reasoning is a more practical path to efficient agents than simply scaling model size. Think fast when you can, slow when you must.
Paper: https://t.co/kYLqeHaY8p
Learn to build effective AI agents in our academy: https://t.co/1e8RZKs4uX
tweet
Offshore
Video
Dimitry Nakhla | Babylon Capital®
RT @DimitryNakhla: Bill Ackman on $GOOG $GOOGL:
“Our view on Google, one way to think about it when a business becomes a verb, that's usually pretty good sign about the moat around the business... You know the Google advertising search, YouTube franchise is one of the most dominant franchises in the world... Now AI of course is a risk if all of a sudden people start searching or asking questions of Chat GPT...
Our view based on work we had done and talked to industry experts is that Google if anything had a, by virtue of the investment they've made, the time, the energy that people put into it, 𝙬𝙚 𝙛𝙚𝙡𝙩 𝙩𝙝𝙚𝙞𝙧 𝘼𝙄 𝙘𝙖𝙥𝙖𝙗𝙞𝙡𝙞𝙩𝙞𝙚𝙨 𝙬𝙚𝙧𝙚 𝙞𝙛 𝙖𝙣𝙮𝙩𝙝𝙞𝙣𝙜 𝙥𝙤𝙩𝙚𝙣𝙩𝙞𝙖𝙡𝙡𝙮 𝙜𝙧𝙚𝙖𝙩𝙚𝙧 𝙩𝙝𝙖𝙣 𝙈𝙞𝙘𝙧𝙤𝙨𝙤𝙛𝙩 𝘾𝙝𝙖𝙩 𝙂𝙋𝙏 𝙖𝙣𝙙 𝙩𝙝𝙖𝙩 𝙩𝙝𝙚 𝙢𝙖𝙧𝙠𝙚𝙩 𝙝𝙖𝙙 𝙤𝙫𝙚𝙧𝙧𝙚𝙖𝙘𝙩𝙚𝙙.”
___
𝐓𝐡𝐞 𝐋𝐞𝐬𝐬𝐨𝐧: Markets constantly swing between fear and enthusiasm, often driven by stories rather than facts. The real advantage comes from maintaining independent thought — focusing on underlying business economics instead of prevailing sentiment. Contrarian investing is about staying grounded in the facts when narratives become loud, emotional, and disconnected from fundamentals.
___
1️⃣ 𝐒𝐞𝐩𝐚𝐫𝐚𝐭𝐢𝐧𝐠 𝐒𝐞𝐧𝐭𝐢𝐦𝐞𝐧𝐭 𝐅𝐫𝐨𝐦 𝐅𝐚𝐜𝐭𝐬
At the time:
• AI fear dominated headlines
• Bard’s early demo shaped perception
• The market concluded Google was “behind”
But Ackman did something most investors struggle to do:
He 𝐢𝐠𝐧𝐨𝐫𝐞𝐝 𝐬𝐞𝐧𝐭𝐢𝐦𝐞𝐧𝐭 & 𝙛𝙤𝙘𝙪𝙨𝙚𝙙 𝙤𝙣 𝙛𝙖𝙘𝙩𝙨:
• Dominant Search economics
• YouTube — a global advertising powerhouse
• Rapidly scaling Cloud business
• Extraordinary profitability
• Massive data advantages
• Deep engineering talent
• Enormous financial resources
• Potential for margin expansion
• 7% earnings yield with a large margin of safety
• Optionality of the businesses within
• Clean balance sheet and ton of cash
Narratives create volatility.
Fundamentals create value.
___
2️⃣ 𝐔𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝𝐢𝐧𝐠 𝐖𝐡𝐚𝐭 “𝐑𝐢𝐬𝐤” 𝐀𝐜𝐭𝐮𝐚𝐥𝐥𝐲 𝐌𝐞𝐚𝐧𝐬
Ackman openly acknowledges AI as a risk.
But notice the nuance:
•Risk ≠ Headlines
•Risk ≠ Fear
•Risk ≠ Price declines
•Risk = Probability of long-term impairment
A critical distinction.
Most investors confuse uncertainty with danger.
𝘠𝘦𝘵 𝘥𝘰𝘮𝘪𝘯𝘢𝘯𝘵 𝘧𝘳𝘢𝘯𝘤𝘩𝘪𝘴𝘦𝘴 𝘰𝘧𝘵𝘦𝘯 𝘢𝘱𝘱𝘦𝘢𝘳 𝘮𝘰𝘴𝘵 “𝘳𝘪𝘴𝘬𝘺” 𝘱𝘳𝘦𝘤𝘪𝘴𝘦𝘭𝘺 𝘸𝘩𝘦𝘯 𝘯𝘢𝘳𝘳𝘢𝘵𝘪𝘷𝘦𝘴 𝘱𝘦𝘢𝘬.
___
3️⃣ 𝐌𝐚𝐫𝐤𝐞𝐭 𝐎𝐯𝐞𝐫𝐫𝐞𝐚𝐜𝐭𝐢𝐨𝐧𝐬 𝐂𝐫𝐞𝐚𝐭𝐞 𝐎𝐩𝐩𝐨𝐫𝐭𝐮𝐧𝐢𝐭𝐲
“The market had overreacted.”
This sentence captures the essence of investing. Markets constantly swing between:
Overconfidence ↔ Panic
Euphoria ↔ Fear
When perception collapses faster than fundamentals… Valuation gaps emerge.
𝘞𝘩𝘢𝘵 𝘭𝘰𝘰𝘬𝘴 𝘰𝘣𝘷𝘪𝘰𝘶𝘴 𝘪𝘯 𝘩𝘪𝘯𝘥𝘴𝘪𝘨𝘩𝘵 𝘳𝘢𝘳𝘦𝘭𝘺 𝘧𝘦𝘦𝘭𝘴 𝘰𝘣𝘷𝘪𝘰𝘶𝘴 𝘪𝘯 𝘳𝘦𝘢𝘭 𝘵𝘪𝘮𝘦. 𝘉𝘦𝘤𝘢𝘶𝘴𝘦 𝘤𝘰𝘯𝘵𝘳𝘢𝘳𝘪𝘢𝘯 𝘪𝘯𝘷𝘦𝘴𝘵𝘪𝘯𝘨 𝘪𝘴 𝘱𝘴𝘺𝘤𝘩𝘰𝘭𝘰𝘨𝘪𝘤𝘢𝘭𝘭𝘺 𝘶𝘯𝘤𝘰𝘮𝘧𝘰𝘳𝘵𝘢𝘣𝘭𝘦 𝘣𝘺 𝘥𝘦𝘴𝘪𝘨𝘯.
___
4️⃣ 𝐅𝐚𝐜𝐭𝐬 𝐯𝐬 𝐅𝐞𝐞𝐥𝐢𝐧𝐠𝐬
Perhaps the most important takeaway:
Successful 𝐢𝐧𝐯𝐞𝐬𝐭𝐢𝐧𝐠 is rarely 𝐚𝐛𝐨𝐮𝐭 superior intelligence.
It is about 𝐬𝐮𝐩𝐞𝐫𝐢𝐨𝐫 𝐞𝐦𝐨𝐭𝐢𝐨𝐧𝐚𝐥 𝐝𝐢𝐬𝐜𝐢𝐩𝐥𝐢𝐧𝐞.
Facts are stable.
Sentiment is unstable. Yet sentiment is louder.
𝙄𝙣𝙫𝙚𝙨𝙩𝙤𝙧𝙨 𝙬𝙝𝙤 𝙖𝙣𝙘𝙝𝙤𝙧 𝙩𝙤 𝙣𝙖𝙧𝙧𝙖𝙩𝙞𝙫𝙚𝙨 𝙞𝙣𝙝𝙚𝙧𝙞𝙩 𝙫𝙤𝙡𝙖𝙩𝙞𝙡𝙞𝙩𝙮. 𝙄𝙣𝙫𝙚𝙨𝙩𝙤𝙧𝙨 𝙬𝙝𝙤 𝙖𝙣𝙘𝙝𝙤𝙧 𝙩𝙤 𝙛𝙪𝙣𝙙𝙖𝙢𝙚𝙣𝙩𝙖𝙡𝙨 𝙞𝙣𝙝𝙚𝙧𝙞𝙩 𝙤𝙥𝙥𝙤𝙧𝙩𝙪𝙣𝙞𝙩𝙮.
___
Final Thought:
The world often feels most certain at extremes. Yet, 𝘼𝙘𝙠𝙢𝙖𝙣’𝙨 𝙚𝙙𝙜𝙚 𝙬𝙖𝙨 𝙣𝙤𝙩 𝙥𝙧𝙚𝙙𝙞𝙘𝙩𝙞𝙣𝙜 𝘼𝙄. 𝙄𝙩 𝙬𝙖𝙨 𝙧𝙚𝙛𝙪𝙨𝙞𝙣𝙜 𝙩𝙤 𝙡𝙚𝙩 𝙚𝙭𝙩𝙧𝙚𝙢𝙚 𝙛𝙚𝙖𝙧 𝙙𝙞𝙨𝙩𝙤𝙧𝙩 𝙧𝙚𝙖𝙡𝙞𝙩𝙮. 𝙎𝙩𝙤𝙘𝙠 𝙥𝙧𝙞𝙘𝙚𝙨 𝙢𝙤𝙫𝙚 𝙤𝙣 𝙚𝙢𝙤𝙩𝙞𝙤𝙣 𝙞𝙣 𝙩𝙝𝙚 𝙨𝙝𝙤𝙧𝙩 𝙩𝙚𝙧𝙢. 𝘽𝙪𝙨𝙞𝙣𝙚𝙨𝙨 𝙦𝙪𝙖𝙡𝙞𝙩𝙮 𝙥𝙧𝙚𝙫[...]
RT @DimitryNakhla: Bill Ackman on $GOOG $GOOGL:
“Our view on Google, one way to think about it when a business becomes a verb, that's usually pretty good sign about the moat around the business... You know the Google advertising search, YouTube franchise is one of the most dominant franchises in the world... Now AI of course is a risk if all of a sudden people start searching or asking questions of Chat GPT...
Our view based on work we had done and talked to industry experts is that Google if anything had a, by virtue of the investment they've made, the time, the energy that people put into it, 𝙬𝙚 𝙛𝙚𝙡𝙩 𝙩𝙝𝙚𝙞𝙧 𝘼𝙄 𝙘𝙖𝙥𝙖𝙗𝙞𝙡𝙞𝙩𝙞𝙚𝙨 𝙬𝙚𝙧𝙚 𝙞𝙛 𝙖𝙣𝙮𝙩𝙝𝙞𝙣𝙜 𝙥𝙤𝙩𝙚𝙣𝙩𝙞𝙖𝙡𝙡𝙮 𝙜𝙧𝙚𝙖𝙩𝙚𝙧 𝙩𝙝𝙖𝙣 𝙈𝙞𝙘𝙧𝙤𝙨𝙤𝙛𝙩 𝘾𝙝𝙖𝙩 𝙂𝙋𝙏 𝙖𝙣𝙙 𝙩𝙝𝙖𝙩 𝙩𝙝𝙚 𝙢𝙖𝙧𝙠𝙚𝙩 𝙝𝙖𝙙 𝙤𝙫𝙚𝙧𝙧𝙚𝙖𝙘𝙩𝙚𝙙.”
___
𝐓𝐡𝐞 𝐋𝐞𝐬𝐬𝐨𝐧: Markets constantly swing between fear and enthusiasm, often driven by stories rather than facts. The real advantage comes from maintaining independent thought — focusing on underlying business economics instead of prevailing sentiment. Contrarian investing is about staying grounded in the facts when narratives become loud, emotional, and disconnected from fundamentals.
___
1️⃣ 𝐒𝐞𝐩𝐚𝐫𝐚𝐭𝐢𝐧𝐠 𝐒𝐞𝐧𝐭𝐢𝐦𝐞𝐧𝐭 𝐅𝐫𝐨𝐦 𝐅𝐚𝐜𝐭𝐬
At the time:
• AI fear dominated headlines
• Bard’s early demo shaped perception
• The market concluded Google was “behind”
But Ackman did something most investors struggle to do:
He 𝐢𝐠𝐧𝐨𝐫𝐞𝐝 𝐬𝐞𝐧𝐭𝐢𝐦𝐞𝐧𝐭 & 𝙛𝙤𝙘𝙪𝙨𝙚𝙙 𝙤𝙣 𝙛𝙖𝙘𝙩𝙨:
• Dominant Search economics
• YouTube — a global advertising powerhouse
• Rapidly scaling Cloud business
• Extraordinary profitability
• Massive data advantages
• Deep engineering talent
• Enormous financial resources
• Potential for margin expansion
• 7% earnings yield with a large margin of safety
• Optionality of the businesses within
• Clean balance sheet and ton of cash
Narratives create volatility.
Fundamentals create value.
___
2️⃣ 𝐔𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝𝐢𝐧𝐠 𝐖𝐡𝐚𝐭 “𝐑𝐢𝐬𝐤” 𝐀𝐜𝐭𝐮𝐚𝐥𝐥𝐲 𝐌𝐞𝐚𝐧𝐬
Ackman openly acknowledges AI as a risk.
But notice the nuance:
•Risk ≠ Headlines
•Risk ≠ Fear
•Risk ≠ Price declines
•Risk = Probability of long-term impairment
A critical distinction.
Most investors confuse uncertainty with danger.
𝘠𝘦𝘵 𝘥𝘰𝘮𝘪𝘯𝘢𝘯𝘵 𝘧𝘳𝘢𝘯𝘤𝘩𝘪𝘴𝘦𝘴 𝘰𝘧𝘵𝘦𝘯 𝘢𝘱𝘱𝘦𝘢𝘳 𝘮𝘰𝘴𝘵 “𝘳𝘪𝘴𝘬𝘺” 𝘱𝘳𝘦𝘤𝘪𝘴𝘦𝘭𝘺 𝘸𝘩𝘦𝘯 𝘯𝘢𝘳𝘳𝘢𝘵𝘪𝘷𝘦𝘴 𝘱𝘦𝘢𝘬.
___
3️⃣ 𝐌𝐚𝐫𝐤𝐞𝐭 𝐎𝐯𝐞𝐫𝐫𝐞𝐚𝐜𝐭𝐢𝐨𝐧𝐬 𝐂𝐫𝐞𝐚𝐭𝐞 𝐎𝐩𝐩𝐨𝐫𝐭𝐮𝐧𝐢𝐭𝐲
“The market had overreacted.”
This sentence captures the essence of investing. Markets constantly swing between:
Overconfidence ↔ Panic
Euphoria ↔ Fear
When perception collapses faster than fundamentals… Valuation gaps emerge.
𝘞𝘩𝘢𝘵 𝘭𝘰𝘰𝘬𝘴 𝘰𝘣𝘷𝘪𝘰𝘶𝘴 𝘪𝘯 𝘩𝘪𝘯𝘥𝘴𝘪𝘨𝘩𝘵 𝘳𝘢𝘳𝘦𝘭𝘺 𝘧𝘦𝘦𝘭𝘴 𝘰𝘣𝘷𝘪𝘰𝘶𝘴 𝘪𝘯 𝘳𝘦𝘢𝘭 𝘵𝘪𝘮𝘦. 𝘉𝘦𝘤𝘢𝘶𝘴𝘦 𝘤𝘰𝘯𝘵𝘳𝘢𝘳𝘪𝘢𝘯 𝘪𝘯𝘷𝘦𝘴𝘵𝘪𝘯𝘨 𝘪𝘴 𝘱𝘴𝘺𝘤𝘩𝘰𝘭𝘰𝘨𝘪𝘤𝘢𝘭𝘭𝘺 𝘶𝘯𝘤𝘰𝘮𝘧𝘰𝘳𝘵𝘢𝘣𝘭𝘦 𝘣𝘺 𝘥𝘦𝘴𝘪𝘨𝘯.
___
4️⃣ 𝐅𝐚𝐜𝐭𝐬 𝐯𝐬 𝐅𝐞𝐞𝐥𝐢𝐧𝐠𝐬
Perhaps the most important takeaway:
Successful 𝐢𝐧𝐯𝐞𝐬𝐭𝐢𝐧𝐠 is rarely 𝐚𝐛𝐨𝐮𝐭 superior intelligence.
It is about 𝐬𝐮𝐩𝐞𝐫𝐢𝐨𝐫 𝐞𝐦𝐨𝐭𝐢𝐨𝐧𝐚𝐥 𝐝𝐢𝐬𝐜𝐢𝐩𝐥𝐢𝐧𝐞.
Facts are stable.
Sentiment is unstable. Yet sentiment is louder.
𝙄𝙣𝙫𝙚𝙨𝙩𝙤𝙧𝙨 𝙬𝙝𝙤 𝙖𝙣𝙘𝙝𝙤𝙧 𝙩𝙤 𝙣𝙖𝙧𝙧𝙖𝙩𝙞𝙫𝙚𝙨 𝙞𝙣𝙝𝙚𝙧𝙞𝙩 𝙫𝙤𝙡𝙖𝙩𝙞𝙡𝙞𝙩𝙮. 𝙄𝙣𝙫𝙚𝙨𝙩𝙤𝙧𝙨 𝙬𝙝𝙤 𝙖𝙣𝙘𝙝𝙤𝙧 𝙩𝙤 𝙛𝙪𝙣𝙙𝙖𝙢𝙚𝙣𝙩𝙖𝙡𝙨 𝙞𝙣𝙝𝙚𝙧𝙞𝙩 𝙤𝙥𝙥𝙤𝙧𝙩𝙪𝙣𝙞𝙩𝙮.
___
Final Thought:
The world often feels most certain at extremes. Yet, 𝘼𝙘𝙠𝙢𝙖𝙣’𝙨 𝙚𝙙𝙜𝙚 𝙬𝙖𝙨 𝙣𝙤𝙩 𝙥𝙧𝙚𝙙𝙞𝙘𝙩𝙞𝙣𝙜 𝘼𝙄. 𝙄𝙩 𝙬𝙖𝙨 𝙧𝙚𝙛𝙪𝙨𝙞𝙣𝙜 𝙩𝙤 𝙡𝙚𝙩 𝙚𝙭𝙩𝙧𝙚𝙢𝙚 𝙛𝙚𝙖𝙧 𝙙𝙞𝙨𝙩𝙤𝙧𝙩 𝙧𝙚𝙖𝙡𝙞𝙩𝙮. 𝙎𝙩𝙤𝙘𝙠 𝙥𝙧𝙞𝙘𝙚𝙨 𝙢𝙤𝙫𝙚 𝙤𝙣 𝙚𝙢𝙤𝙩𝙞𝙤𝙣 𝙞𝙣 𝙩𝙝𝙚 𝙨𝙝𝙤𝙧𝙩 𝙩𝙚𝙧𝙢. 𝘽𝙪𝙨𝙞𝙣𝙚𝙨𝙨 𝙦𝙪𝙖𝙡𝙞𝙩𝙮 𝙥𝙧𝙚𝙫[...]