I’m excited to finally launch this. My MCP server now offers real-time stock market news. Available tools: • get news • get stock prices • get key metrics • get financial statements Took about a month to get the MCP server right, but it's working well now. More financial tools coming soon.

Big changes coming to AI hedge fund. I’m Perplexifying every agent. • query routing • task planning • financial data retrieval • trading signal generation, etc. Launching this weekend.

I’m very excited for this. Our AI hedge fund now: 1 • analyzes stocks 2 • analyzes portfolios I’ve enjoyed having the Buffett and Munger agents point out the faults in my portfolio. Also, been a great way to pick new stocks for further research.

I’m open sourcing my financial LLM evals The initial github repo is set up My vision: 1 • evaluate single LLM 2 • evaluate chain of LLMs 3 • evaluate agentic LLM flows All tasks will be on investment research.

I’m enjoying this financial assistant use case Comparing fundamentals of stocks My usual approach: • get financials for company A • get financials for company B • compare strengths vs. weaknesses I like being able to see the underlying data and then the LLM’s reasoning. Best part is that our demo app works well on mobile browsers. Code is open source below, as usual.

I really love this chart It shows today's top 6 frontier LLMs My initial takeaways: 1 • gemini is a fantastic LLM 2 • both k2 and r1 are insane value 3 • quality battle is tight, pricing is not I’ll admit I haven’t used gemini 2.5 pro as much as sonnet 4. This is about to change. Pricing for k2 and r1 is mind boggling. k2 is based on the same architecture as r1 - this feels like the future of LLM design? Finally, k2 is same ELO as opus, but 29x cheaper. Tons of room for improvement here.

Our MCP server now uses streamable HTTP. Streamable HTTP has become the default transport protocol for MCP. We support both: 1 • SSE (legacy) 2 • Streamable HTTP (new) Excited to get this live 🚀

I added an agent swarm to our AI hedge fund This one takes large, contrarian bets. The Market Mavericks: 1 • Bill Ackman 2 • Michael Burry 3 • Stan Druckenmiller Michael Burry: known for making bold, unconventional bets against overvalued markets. Bill Ackman: activist investor who takes large stakes to influence corporate strategy. Stan Druckenmiller: finds major macro trends to make big, directional investments.

I’m restarting my LLM evaluations. My focus will be on financial research tasks. Initial ideas: 1 • news sentiment analysis 2 • financial calculations 3 • 10-K analysis, etc All code will be shared - as goal is to learn. Will test both small and large models. I’ll also finetune small open source models and see how they compare to large ones on specific tasks! Experiment ideas are welcome.

I upgraded our AI hedge fund’s backtester. This adds key evals to our research system. How it works: 1 • enter stocks 2 • select AI analysts 3 • set available cash 4 • run backtester You can run evals across any time horizon. The backtester simulates daily market activity using historical data. Let me know what other evals to add.