I think the play was something along the lines of being "the new iOS". Lots of new apps (see: LLMs) are browser-first, forgoing building platform-targeting apps entirely. When they couldn't see any app devs lining up to build Arc-specific extensions, apps, or such, they pivoted. Dia was more of an LLM-host browser, with the play being they're hoping for OpenAI or Perplexity or one of the big foundation-players to "pay" to be the "exclusive" AI provider, like old-Google pays to be the default on Safari. But ... both plays didn't find any audience or customer, rightly so, as they didn't try to "fix" anything anyone actually considered a problem, just tried to build a niche their own. :(