|
Today's Brief
It's like the hard drive market all over again: AI running costs are dropping so fast that 'is it worth automating?' is no longer a real question, which is why you're seeing labs and companies rush to release new, giant models like DeepSeek-V4 that can handle one-million-token contexts. This lands on trading desks as a scramble to integrate these new models into existing workflows, because the cost savings are too big to ignore - and it's not just about the models themselves, but about how they're changing the way we think about event contracts and prediction markets. Now you need to ask your team: are we using the right model for our specific problem, or are we just using the biggest one because it's the newest shiny thing?
|
