AI researcher Herbie Bradley argues it is easier to build a $100B data company than a $100B AI lab: "I actually think it probably is much easier to build a 100B data company than a 100B neo lab." "Part of the reason there are so few is that structurally it doesn't appear to be a good business. It's sort of like a consulting-style business where you're not really selling SaaS, you're just grinding out this massive workforce, or you're scaling quite linearly in some cases." "From a VC's point…