Curation over aggregation
Major tech companies have generally take an aggregation approach to building technology platforms. They gather as much information as possible, list as many products as available, and work to serve nearly everyone by curating results for each individual. It's worked in terms of growth for the companies, but it's far from apparent that most of the major tech companies have the users best interests in mind when curating results.
With millions upon millions of websites, products, apps, etc available, aggregation platforms are forced to curate content for users. It's not feasible to serve the same content to an audience that has no connection besides common use of a platform. To scale to an infinite amount of users and data, tech companies turn to machines to help many of the decisions by developing algorithms. By nature, algorithms are only as good as the data supplied, and the quality of the algorithm, which naturally leads to biases and miscues that are readily apparent to humans.
Seth Godin wrote about the role of discovery among various tech platforms. He notes the problems with Apple and Google, and compares that to the impressive 'discovery' features of Netflix:
Apple has carefully guarded the podcast directory, persuading podcasters that ‘winning’ here is the shortcut to building a popular podcast. But they’re terrible at introducing podcasts to new listeners, terrible at developing a point of view that enables the industry to thrive or even grow.
Compare this to Netflix. They’re terrific at surfacing content and helping people find things that they end up liking. Netflix, for economic and marketing reasons, has limited the number of ‘shows’ they have to promote, but within their set, they continue to delight. Compare this to YouTube–if you follow the ‘recommended’ path for just a handful or two of clicks, you’ll end up with something banal or violent.
There's opportunity for more curation that requires manual work. It's not scalable and it's far more meaningful to a core audience. Build something better for the people who care enough about the subject to want more than an algorithm's results. Of course, make use of technology - and add in the hard work of real curation.