Not to say you're wrong about commoditization. I don't think these companies will be able to raise their prices and keep them there to make enough money to keep building models like they've been doing.

There are people working on this.

e.g. https://github.com/haykgrigo3/TimeCapsuleLLM

It's the same in other industries. In your example, if labor went into building a whole new assembly line that pays itself off over several years, capex. If it just goes into building products to sell, opex.

https://www.hibob.com/blog/r-and-d-capitalization/

https://fred.stlouisfed.org/series/IHLIDXUSTPLEGA

No, not quite. It really comes down to opex vs capex and the depreciation schedule for your investment.

Software development is typically categorized as capex, on a 3-5 year depreciation schedule. You assume the software you write today will be generating value for you that long.

If a big, expensive model training project only gives you value for a year or less, that is not like most companies.

Source: My wife works IT for our school district.

But other things will:

- Liability rules

- Regulations that you get audited on (esp. for companies already heavily regulated, like banks, credit agencies, defense contractors, etc)

If you get the legal responsibility part right, then the education part flows from that naturally.

It's kinda shitty to steal someone's works, then use them to build a machine to also steal their jobs.

It works in isolation and fails miserably when trying to do a big demo of it in a conference talk when attempted by dozens of people in the same room.

What you own is the image copyright. But the right to copy is only one of the rights at issue.

Under various state laws (California in particular), you might not be entitled to do all the things with that picture that you could do of one that doesn't have my likeness. Privacy laws like the CCPA are one possible carve-out. A "right of publicity" is another.

There's an old saying about property law that "property is a bundle of sticks". The bundle can be subdivided.

https://www.law.cornell.edu/wex/publicity

In the beginning Twitter was very free and open with API access. There were plenty of alternative apps. Of course, that changed when they got serious about monetizing.

Would it really be any harder for Bluesky to switch from ATProto to a proprietary API than it was for Twitter to close their API? How many users are realistically going to download their archives and upload them to some other provider? If most people are using the website or official app, that's where the stickiness is. There would be a blog post with a title like "Supporting the Bluesky Community for the Next Century" and how it's better to have a centralized site that can feed its employees than an idealistic decentralized one that disappears. Things would seem OK at first. But enough years of chasing quarterly KPIs would put them in the same spot as Twitter and Facebook.

If cable television or restaurants or ice cream start causing harm that we want to deal with, we can vote on that when the time comes.

> I assumed the writer was a journalist or author with a non-technical background trying to explore a more "utopian" vision of where trends could go.

If you assume you're reading something from a person with intention and a perspective, who you could connect with or influence in some way, then that affects the experience of reading. It's not just the words on the page.

This is an interesting reversal in itself. If you make the specification protected under copyright, then the whole practice of clean room implementations is invalid.