Nice to be mentioned by efinancialcareers in this article covering life after quant. Although, as you'll see below, GPT Astra was so mind-blowing that we couldn't quite resist leaving trading fully :-p efinancialcareers.co.uk/news…

Exchanges are S-tier software businesses. With AI rapidly eroding engineering moats (we run one with a grand total of 5 engineers), network effects are the only things that matter for long-term retention. Any trading product that is building a UI/frontend/broker/B-Book is doomed to compete in an ever-tightening noose of CAC. Liquidity is the only option for viability. Matching models that do not maximize for passive liquidity from a differentiated user base are unviable. CLOBs are one example that work, put they suffer from a 'leaky hand' problem. Sophisticated trading firms don't want to place alpha-based orders into an order book with public market data, for fear of 'leaking' their model edge before they get filled. In this case, alternative trading systems (aka dark pools) can provide genuinely better liquidity inside the spread. Everyone wins. Something for @QFEX to think about as we scale.

I think crypto’s biggest contribution will be better financial systems. Perps are just one example, but there are lots of ideas borrowed from 'crypto' that don't need a blockchain at all.