New trends like LLMs and Apple Vision may make linux/BSDs harder to justify for productivity

I had this thought today: what takes over the tech world now (LLMs and VR) make it harder for 'normal' users to choose Linux or BSD.

LLMs

MS will integrate generative AI on every product. Even on the OS itself. Standard users will learn they can talk to the OS/tools to get what they want. Linuxes and BSDs can hardly do the same, because it requires lots of compute power that is usually centralized (as of mid 2023; perhaps in a few years we can run LLMs locally).

Incentives to learn 'how things work', edit config files, figuring out dependencies… are less obvious. If you can ask the machine in plain English what you want it to do, there's no reason to get inside settings. If the machine helps you produce better word docs, powerpoints etc… this is a difficult value proposition to replicate from the open source world. Libreoffice would look primitive in comparison. And there's a clear price tag attached to using the FOSS version of productivity software (less productive), that managers will not want.

The same way people don't think about writing code to solve their problems anymore ('there's an app for that)… we might be the last generation that actually knows how things work inside, and cares about it. But end users will produce better word docs, powerpoints etc…

Even if Linux/BSD would get comparable features in the future, I suspect there's a gap of 1-2 years where they would look comparatively (and objectively) inferior for productivity for the 'normal' office users. Which you may argue, were never the target audience for those operating systems.

Apple Vision

Apple vision is Apple's new mixed reality headset announced at WWDC in June 2023, which is planned to be shipped in early 2024. The headset, priced at $3,499, has been described by Ben Thomson (who I consider to be a decent thinker) as an extraordinary experience exceeding his already high expectations.

Spatial computing needs extreme integration between hardware and software. I suspect it'll be very difficult to make Linux run on Apple vision or similar products by other bigtech companies.

Even if eventually there's linux running on this hardware, it feels like using the native features will take an inordinate amount of work, and the advantages of using a great shell/terminal on linux (compared to the alternatives!) disappear in a world that is designed for point and click, in this case pointing with your eyes.

This headset is really a two way interface for personal AI. while I think we are still ways to go to have personalized AIs, imagine the possibilities for children to learn by playing interacting with an agent that knows exactly what they like, what they have been exposed to, etc. A curriculum with a a single user in mind. No need for exams (as the machine knows what you know, in real time, building a model of not only your knowledge but your learning strategies).

The last thing we need is another passive device to watch 3D TikTok. Another "walled in" device where the only way to solve problems is to use someone else's solution in the form of a very specific app, that abstracts away 'details' such as files in a filesystem.

The core question is: Does the Apple vision pro make professionals and hobbyists substantially more productive? If so, it will take off. Then it will reach mainstream success, and with that, eventually, the education system. Will it combine with generative AIs to produce personalized learning experiences? If so, we really need these devices to run open source OSs when they go mainstream. Do you want your child's future to depend on the whims of a bigTech company?

As Ben Thomson says:

Consider that the iPhone expanded computing from the desktop to every other part of a person’s life. It turns out that was an even larger opportunity than the desktop, and the product category was an even larger success.

Will this new category of smart glasses produce better learners? If so, will it depend on knowing the user intimately? if so: is it a good idea to let a bigTech company dominate this market? I hope not.

Comments on HN:

https://news.ycombinator.com/item?id=36224947