
Building artificial intelligence for the public sector | Co-Founder @traddoo | Technical AI Consultant | Christian
Joined February 2024
- Tweets557
- Following315
- Followers504
- Likes502
Pinned Tweet
This trader passed April 9th
He’s gotten 18 payouts
Total ~ $26,437
Really awesome to see how beneficial having the ability to withdraw at anytime has been for our traders
#TeamTraddoo
Thomas retweeted
Ranking of LLMs by the % of JSON violations we detected over the past week, for the top Structured Output requests
⭐️ Qwen, Mistral, and GPT-4o-mini generally doing the best
‼️DeepSeek v3, Sonnet 4 seeing violations north of 20%
Thomas retweeted
Agentis-Omni is the general purpose powerhouse
Capable of pulling data from anywhere on the web and transforming it into actionable insights
Development, Research, Analysis and more
Thomas retweeted
Introducing the next generation: Claude Opus 4 and Claude Sonnet 4.
Claude Opus 4 is our most powerful model yet, and the world’s best coding model.
Claude Sonnet 4 is a significant upgrade from its predecessor, delivering superior coding and reasoning.
Everything we build for @Traddoo will be open source
That way if you see anything you like and would like to implement it at your own firm you have it available
I’ve been functionally retired since my early 20s and will continue to use my time to build things that benefit everyone in this space
Super excited to show everyone what we’ve been working on the last 6 months
3
27
1,655
Agent based support now online @Traddoo
Built an entire agent framework from scratch to avoid reliance on any external frameworks or dependencies
Will post updates here as we continue to enable new capabilities
5
406
Thomas retweeted
They don't want to see us win.
8 months in the making. Building out of selflessness, all culminated into one relaunch with some kick ass partners. I couldn't have it any better way .
thankyou Max, Connor, Zaeem, and Thomas for believing in a vision that has only just started.
10% and loading...
Thomas retweeted
Agentis-MCP framework prioritizes MCP servers opposed to traditional tools
Here we have a Claude 3.7 agent with filesystem + brave mcp researching stocks and then creating a file with an investment plan
Persistent connection enables multi-turn tool interaction in each server
PyPI package releasing this weekend
Thomas retweeted
Introducing AEGIS - AI enabled gateway for intelligent Solana operation
Here we use a mixture of agents to perform token analysis.
You can also execute with natural language using Wexley!
#AgentisLabs
Free tool to help you avoid scam firms
Pulls real-time data from social media and various review sites to generate a report
As of now it’s ranked @FTMO_com #1 and @breakoutprop #2
Will be adding a few more data sources that were suggested by the community!
propview.net
Thomas retweeted
It was time to future-proof our operations
Introducing Traddoo Futures
The same live capital allocations we brought to FX - now available for Futures Traders
#TeamTraddoo
There are some firms claiming to allocate real capital that are not at all
In fact they are actively slipping traders on demo accounts using “real capital” as a guise to do so
Over the next few months a lot is going to come out about these operations
Traders deserve to know the truth, even if that means some firms get hurt in the process
Thomas retweeted
Swarms - experimental multi-agent framework
Here we create a GPT-4 chain with enhanced reasoning and planning capability
@OpenAIDevs
Use this prompt to give Claude o1 level reasoning
Dynamic COT + Reflection + Verbal Reinforcement Learning
Works surprisingly well
3
10
1,191
Can @AnthropicAI Claude 3.5 sonnet outperform @OpenAI o1 in reasoning? Combining Dynamic Chain of Thoughts, reflection, and verbal reinforcement, existing LLMs like Claude 3.5 Sonnet can be prompted to increase test-time compute and match reasoning strong models like OpenAI o1. 👀
TL;DR:
🧠 Combines Dynamic Chain of thoughts + reflection + verbal reinforcement prompting
📊 Benchmarked against tough academic tests (JEE Advanced, UPSC, IMO, Putnam)
🏆 Claude 3.5 Sonnet outperformes GPT-4 and matched O1 models
🔍 LLMs can create internal simulations and take 50+ reasoning steps for complex problems
📚 Works for smaller, open models like Llama 3.1 8B +10% (Llama 3.1 8B 33/48 vs GPT-4o 36/48)
❌ Didn’t benchmark like MMLU, MMLU pro, or GPQA due to computing and budget constraints
📈 High token usage - Claude Sonnet 3.5 used around 1 million tokens for just 7 questions
399



















