It's funny that the article is a list of predictions that the author got wrong followed by a new prediction. I'm willing to bet they'll be wrong about this one

It's funny that the article is a list of predictions that the author got wrong followed by a new prediction. I'm willing to bet they'll be wrong about this one as well. I'll note that just because websites ended up needing SEO it doesn't mean creating websites was a mistake. Early on in internet history there were two important fights: the first was about whether we needed websites at all. One view held that the _internet_ was valuable but websites were a novelty. Applications on your computer that you install would leverage your internet connection to transmit data. The second view held that you'd eventually only have a single application on your computer: your web browser (essentially every computer would be a chromebook). Obviously we ended up much closer to view 2 but it's not hard to imagine a world where view 1 won - just look at the way people install apps on phones. The other important fight was whether the web should be more like broadcast or cable. One view held the web would be agnostic of your provider - you pay for access to the internet and then you can access whatever you want. The other view held that each provider would have its own unique set of websites that subscribers would get access to. If you are old enough to remember AOL keywords you might be able to see how this vision was in the works. It seems like the author was upset the open internet (first view) won. The author and I agree about one thing: the first time ever I feel like voice can be a first order input mechanism. I know that we've had Siri and Alexa for a while but they've been a chore to use as perfectly articulated in the author's section "Language Has Been the Logjam". Now it seems their world view is that we'll all be writing plugins for Siri or Alexa or whatever voice platform is dominate. That sounds to me like saying "don't bother making websites or apps - just make APIs" (by the way that was another real worldview that got some traction about 15 years ago). So let me present two alternative scenarios that could play out: First - we could end up with a proliferation of bots each bespoke. I know many different people. I wouldn't ask my friend from IT about organic chemistry and I wouldn't ask my biology professor buddy about directions around my town he's never visited. If our programs become more lifelike it's possible that same circuitry in our brains that allows us to remember people can be coopted to remember programs and they'll want each one to be distinct. Second - maybe there will be a common voice interface but each bot will still generate its own text. Just like not every company builds a web browser but everyone still makes their own websites. Like the SEO and app stores mentioned in the article you might have to play by the rules of the platform you're on but ultimately you're still running your own LLM for your application.