16 comments

[ 3.2 ms ] story [ 58.6 ms ] thread
[flagged]
given the breath and depth and history of Bloomberg, that seems eminently possible to me. Why does that seem preposterous to you?

Bloomberg news has been around since 1990 but Bloomberg itself was founded in 1981. They would have the rights to do whatever they want for all the reports, is which there are multitudes, both internal and external, to use as a training corpus, not to mention all stock market movements they've recorded. not only can they go back and see which reports and analysis panned out, but they can analyze why the ones that didn't failed based on all their market data.

Giving the GPT a brokerage account and telling it to go make money is one of the four horsemen for an AGI takeover of the human race, and even though that's hyperbola, it's still a move towards the machines taking over. Whether they'll be harnessed for the good of humanity, just a select few, or even for anybody's good, we'll just have to wait and see.

I wonder if they have the ML wizards to make this happen
I mean, that table certainly doesn’t sell it.
For all the hype, the benchmarks they report don't seem super compelling. Financial questions need to have 100% accurate responses (not 60% or so they report) otherwise they are worse than useless and can cost billions of dollars in losses
Isn't it enough to win 51% times?
It’s not trading anything, in which case the above still wouldn’t be true. It’s just answering financial questions, so if it hallucinates and tells you Tesla isn’t a public company 40% of the time then what’s the point? I’m using a very simple example here, I’m sure the model makes more intricate mistakes.
As it's a closed model, only comparision with GPT-4 matters, and it's missing.