Two weeks ago, I achieved a breakthrough in stock trading that backtests at 973% CAGR on S&P 500 stocks (and an improved version still in training has already surpassed its validation loss.) The model is detailed in other markets, such as this one:
/SteveSokolowski/will-my-stock-trading-market-exceed
I am preparing marketing materials to solicit a businessman (or woman) to set up an organization and raise funds for this model. Please offer your opinion on this logo. Do you like it? If not, why not?
I would also be interested to find out if people believe that the slogan's claim is perceived as too bold, or whether a model that is currently tracking with a Sortino ratio of 100 has indeed achieved superintelligence in this limited domain. Should the slogan be changed because the potential audience simply won't believe it?
Note: as with all good things, this logo was entirely developed by GPT-4 and DALL-E 3.
1,000
3.00I basically agree with the other comment that "Seregon" "AI investing" and "Superintelligent trading has arrived" should all have exactly the same width, and the vertical space between "AI investing" and "Superintelligent trading has arrived" should be smaller.
I don't really have a problem with the font or color, the wizard is cool. I think people will not be expecting that you think your model is literally superintelligent and that "superintelligent" in the tagline is meant as a playful exaggeration. (Though maybe that will change when people see you are claiming 900% CAGR, I don't really know anything about stock trading but isn't 30% CAGR supposed to be very good?)
@BoltonBailey The name was generated by GPT-4 as well. The idea is that Clarke said that superintelligence is indistinguishable from magic.
@BoltonBailey Also, in regards to the CAGR numbers, yes, 30% was considered very good. But, considering today, we are up 4.2% in six days - before compounding (a trading year has about 220 days.)
At first, I thought that such returns were impossible. But then I realized that in every field, there is a moment when AI "clicks" and goes from useless to superintelligent within months. We saw that with chess in 1997 and Go in 2016 and image generation in 2023.
In another poll, I discussed that this architecture might lend itself to music generation. If I had more GPUs, I would like to try that out to see if this is a genuine breakthrough that could be generalized to other one-dimensional sequences of floating point values like music where "good enough" is sufficient for prediction.
It rings amateur for me, maybe because the font of the letters (too common/simple), the color choices (too many -green, musk green, black and grey), too much white space, the names are not aligned too (you can tell seregon is more to the left than AI investing). I like the slogan and the magician though
@drcat Thanks for the feedback. I will create a second draft tomorrow and re-post it with your improvements.