
@miramurati @tinkerapi This looks good! Can we test it? 👀
The model helps financial analysts filter and prioritize documents.
Many users praised Bridgewater partnering with Tinker API to fine-tune AI for financial analysis, calling it a cool and smart real-world use of LLMs by quant experts.
No Digg Deeper questions have been answered for this story yet.

@miramurati @tinkerapi This looks good! Can we test it? 👀

@miramurati @tinkerapi Love it when finance leverages AI! Here's some work we did a couple of years ago with BlackRock on fine-tuning specialized small models for extracting insights from financial knowledge graphs: https://developer.nvidia.com/blog/insights-techniques-and-evaluation-for-llm-driven-knowledge-graphs/

@miramurati @tinkerapi Experts improve the AI, AI amplifies the experts.
Human judgment is always the real edge.

@miramurati @tinkerapi There's something so hypnotic about the marching squares algorithm

@miramurati @tinkerapi custom models >
refreshing to see the radical transparency around the open source model used and why

@miramurati @tinkerapi Brilliant! I hope you're optimising for independent consumers as well. I see that your organisation is enterprise-centric for now. Congratulations on the new partnership

@miramurati @tinkerapi Any plans to add GLM-5.2 to Tinker?

@miramurati @tinkerapi this is why dario has been crashing out and asking for ban on open source models... he knows finetuned models outperform both on cost and accuracy

@miramurati @tinkerapi good work mira

@miramurati @tinkerapi Always happy to hear about the progress of the "Thinking Machines " and tinkerapi that we are already using it.

@miramurati @tinkerapi Knowledge + AI = exponential impact. @miramurati Hoping to one day contribute to work like this at Bridgewater AIA Labs.

@miramurati @tinkerapi hardest part of domain fine-tuning isnt the model - its having domain experts who actually know what to prioritize. thats the real edge here

@miramurati @tinkerapi https://youtu.be/9_qGeV1XRDU

@miramurati @tinkerapi Expert knowledge as the training signal is more durable than any benchmark... Fine-tuning only earns its place when the people closing the loop actually know what matters.

@miramurati @tinkerapi Thats actually a smart use case. Love seeing domain experts shaping the tech instead of the other way around.
@miramurati @tinkerapi Mira, where can we test it?

@miramurati @tinkerapi 金融场景里这类微调确实挺吃专业判断的

@miramurati @tinkerapi Great use of expertise

@miramurati @itsandrewgao @tinkerapi Congrats to the team!

@miramurati @tinkerapi 🔥