AI startups are going to get squeezed between the foundation models that now want to build applications themselves + the systems of record that have all the context data and also want to build the applications themselves curious how others think this competition will play out 1) systems of record still lack a lot of context outside of the record, so your company could stitch together the outside context to make a new system of record 2) AI is part of your process but includes IRL components to deliver services, capture data, etc. 3) You AI applications are best in class for a really specific workflow that the systems of record or foundation models just can't make good enough (e.g. making a consumer application/interface, which the enterprise companies have 0 idea how to do well)