• A "best-in-class" financial AI that is also ultra competitive in performing general tasks.
• 50bln parameters LLM trained on very robust underlying financial data.
• The largest domain-specific dataset with 363bln tokens consisting of English financial documents (complemented with > 345bln tokens from general-purpose datasets).
• Already exceedingly outperforms current models exclusively focused on financial tasks.
• Flexibly able to provide real-time analysis of financial data from Bloomberg’s proprietary terminal, assisting in a variety of tasks for analyzing financial markets and making investment decisions.
• Powerfully analyzing and processing large underlying financial data.
• Aids deep sentiment analysis.
• High intrinsic accuracy in named entity recognition.
• Aptly aiding the news classification routines.

According to Shawn Edwards, Bloomberg’s CTO, "BloombergGPT will enable us to tackle many new types of applications, while it delivers much higher performance out-of-the-box than custom models for each application, at a faster time-to-market."

These are very exciting times. And perhaps, scary times as well! With more embedded use cases being baked in gradually for AI in finance, the potential disruptive power can be far-reaching, including its wall-to-wall ramifications for the traditional finance functions. Do you see massive restructurings and/job losses on the horizon in the coming months/years? Let me know in the comment session. One thing is clear for now: just like the internet, AI is here to stay!