TiredBanker has taken 10 years-worth of earnings reports, and call transcripts and summarized them using GPT-4. We provide a weekly email digest containing new calls that took place, or you can peruse our comprehensive library of previous calls.
Now THIS is an excellent use of GPT/LLM technology...! Looking backwards across a huge but uniform dataset to pull out insights that human analysts may not be able to spot simply because — how can a human keep so many variables in mind at one time?
This use case is narrow, specific, and useful. Love it.
Report
@chrismessina Came here to say this. Agreed, there is beauty in the uniformity of earning reports + LLM’s
Hello! I’m Sunil Rajaraman - I’ve been a serial entrepreneur/writer in the tech world for the past 10 years. I co-founded a writer marketplace called Scripted, and ran a local SF publication called The Bold Italic which I sold to Medium a few years back. Along the way I’ve invested in startups - both personally and as a Sequoia scout, and ran marketing at various companies (was CMO at Metromile, and VP of Marketing at GoodRx).
Early in my career, I was a pure numbers guy. I spent time as an M&A analyst at a large company, and one of the most thankless tasks was writing summaries of earnings reports of competing companies. I had to listen to calls for hours, and dissect 10-Qs to get relevant takeaways.
I decided to start TiredBanker as a small project - TiredBanker takes every earnings report for the past 10 years of every S&P 500 companies and turns it into something you can easily understand (using GPT-4). TiredBanker is one of many projects I plan on launching as part of my new company called Hamlet. I don’t think AI will ever replace the role of journalism - but it will take complex and convoluted datasets and turn them into something usable.
Feel free to reach out to me directly on Twitter at subes01 - would love your ideas on how I can improve this. Eventually, I do plan on including more companies + add voice transcription support… Thanks for your feedback!
Time to get back onto wallstreetbets 😆 But in all seriousness, this makes personal finance & investment knowledge a lot more accessible to people. Great launch ?makers!
Love how easy it is to actually just understand a company at a high level! Looking forward to receiving the weekly digests…
Is all the data shown about companies from LLM trained on the quarterly reports etc or, do you also plug into ‘real’ data sources?
Congrats on the launch, Sunil! It would be super cool if a user was able to make specific questions against dataset, e.g. "what is common for all companies which grew their revenue at least 20% each quarter"?
I think this application of AI is pretty unique. The key problem in stocks is not a large amount of data, but *how* do you look at that data. The fractal nature of the fluctuations makes different people look at the data differently and this is where AI plays the role. Summaries look great. Pixel art adds a cherry on top. Great job!
Raycast
Tired Banker
Tired Banker
Wudpecker
Tired Banker
Zeed
Tired Banker
Lumos
Quilgo
Tired Banker