The most entertaining outcome is the most likely
Remember the keypads on smart phones in the early 2000s? Even before I was able to use one I had an iPhone in my hands, playing Youtube with the old CRT TV logo. I didn't need an instruction manual like I did with the Wii. I picked it up and played Pocket Frogs. Times change quickly and before the next generation understands how to setup their AI pen pal I think we'll be past the strenuous work of touching, swiping on a glass sandwich.
Keyboards can transmit around 6.5 bits per second whereas humans can consciously think and process info at 41 bits per second and a couple years ago Zuck said "We’re working on a system that will let you type straight from your brain about 5x faster than you can type on your phone today." The point is that everytime I type in my thoughts into a text-box into an interface that is hosted on a server a few hundred miles away from my screen, I'm both bandwidth(keyboard) and latency(wait for site to send to inference server, back to me, etc) constrained. For something that can be extremely intuitive and dynamic, I feel deeply frustrated everytime I can't type in gibberish and the LLM just understanding the entire context of my life and stream high quality tokens straight into my mind.
There are probably a few solutions, new brain interfaces, new web interfaces, ai hardware that sits on you all the time, etc. And most contrarily will work. Some of these founders understand deeply what the existing and new era of interfaces and interaction will be yet most people online shut it down trying to grasp why their glass sandwich will be the end all be all for all humanity.
The most absurd outcome will be more likely
A fantasy coin backed by anonymous miners makes no tangible sense, a glass sandwich where we can take photos, type, and consume media makes no tangible sense, and a device that can either read our minds, listen to our thoughts and can initiate and let us consume media with our own preferences also doesn't make tangible sense, yet the people do not understand what they want yet -- The future is now, old man, its time to see something new.
So what is new?
Filtering perception. We'll be able to tweak and fine-tune our consumption of feeds. Social feeds, news feeds, and communication channels(slack, texts, etc)
New, old interfaces. Interfaces will say the same; chat boxes, search pages, but you won't have to interact with them so often. They will know your intentions before you arrive and do actions in the async -- auto reply to setup up meetings, show you cooking videos near lunch time, order ahead of time. Ideally you will just need to watch and confirm, sort of like TikTok if it had auto scroll and you consume entertainment
Closer brain-computer interfaces. Non-invasive interfaces will be able to read, process, and relay information directly from/into your brain. If TikTok is the ideal app for the screen, apps will be able to feed you info subliminally.
Downfall of Apple. A companion that lives on a dedicated device will strip the need of a phone. The Vision Pro feels more of a last-ditch attempt for a multi-trillion dollar company to diversify their revenues from the iPhone, and the sooner we finalize the correct interface and abstraction for this companion, the sooner the glass sandwich will be viewed as the 2000's keypad on a phone. Hardware, software, brainware?
The kids don't talk to ChatGPT, a survey I've done informally says that they don't. They use it to cheat on their homework and thats kinda it. If you want an AI companion that won't exist in that interface; that interaction only emerges via a completely different domain, either via an anime waifu app or hardware device.
The most absurd outcome is most likely. the more angry, frustrated, and belligerent people are towards new ideas and quickly growing products that these new products will survive. No product has never had no haters. Part of me feels frustrated using Perplexity and seeing these new AI devices, frustrated that I wasn't able to execute on those before.
At the same time at this point in the market I feel hesitant to say we're not at the top; because we definitely are at a local maxima. The reason I don't say global yet is because more AI companies would be getting acquired if we were. They're all well capitalized already yet either the IP/moat isn't there or there's a couple more years until there's more prey ready to be killed.
The pieces are already on the board and figuring out what you are doing to accelerate already existing infra is the only bet to make it out alive. For example no one can make a specialized chip just for Transformers. No one can make a new computer. Billions have been put to make these existing things the absolute best they will and ever can be. Putting your time towards letting new interfaces being built would be better rather than rehashing on the same idea.