Offshore
Video
God of Prompt
84% of developers use AI coding tools daily.
25% of new startups ship codebases that are almost entirely AI generated.
But here's the stat nobody talks about: AI assisted developers produce 3-4x more code... and 10x more security issues.
Your vibe coded app isn't broken. It's unfinished.
Here's the gap nobody's solving (and why it matters now): 🧵
tweet
84% of developers use AI coding tools daily.
25% of new startups ship codebases that are almost entirely AI generated.
But here's the stat nobody talks about: AI assisted developers produce 3-4x more code... and 10x more security issues.
Your vibe coded app isn't broken. It's unfinished.
Here's the gap nobody's solving (and why it matters now): 🧵
tweet
Offshore
Photo
NecoKronos
So... who got the trade?
Once we filled the selling imbalances, it was lights out.
An instant 2% drop to start the week. Thanks for playing. 🤝
#BTC https://t.co/Olwk8sMFx1
tweet
So... who got the trade?
Once we filled the selling imbalances, it was lights out.
An instant 2% drop to start the week. Thanks for playing. 🤝
#BTC https://t.co/Olwk8sMFx1
The dream scenario for #BTC right now?
✅ Range VAH deviation confirmed.
✅ Selling stacked imbalance cleared.
If we get both, the trade sets itself up https://t.co/pOqLSlFcrG tweet
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