I predict open source is going to have a day of reckoning. Devs are already asking "why should I put code on GitHub for free?" and they're right to. Microsoft owns it and other big corporations have essentially built giant projects off of code stolen from there (AWS is an example). Equally, they are starting to ask "why should I do uncompensated work on open source projects if megacorps will be the main beneficiaries?". Given the abuses of devs by the corporations of late, loyalty to employers is shot. I think we are looking at a major change in the software world over the next 10-15 years.
Discussion
i think its the opposite, its time to doubledown on opensource hardware and firmwares.. not being able to even run a pre-trained llm without begging some proprietary api to let you.. or make a phone call without begging for a spyware device.. thats the end of times.
For practical reasons, yes, I'm just saying I think there will be more reluctance from devs to help big corporations for no compensation.
Should have made that clearer.
The opensource I look up to, is the source that enabled the takedown of Microsoft. Back in the day, Microsoft was on the same track as Apple, and OpenAI is now. To dominate the entire server market. I hated them, took every opportunity to replace them. It felt like paying a tax just to exist, just to run a website or a file share. Opensource was what got us into position to run our own servers, with Linux, Apache, Mysql, Php, Samba, and be paid (jobs created) to contribute to and run opensource at companies.
I think those projects did well for themselves, and for their users/contributors.. What really scares me now is not just that we let google and apple dominate mobile and wall the garden, but that AI is the new battleground and quickly heading toward dystopia where we again, aren't able to run our own computers.
Interestingly, without china releasing these opensource models, it would appear the battle was already lost.. Zuck will/would cave to closed source eventually, but not if china keeps releasing. Sadly, the gathering of all the data from the internet and running it through billion dollar data center training, isn't something a startup or small operation can do and therefor, we are at the mercy of the big corporations like never before.. We can only hope that the models themselves will remain useful and available.
With AI, closed source doesn't work, anymore, unless you only offer code as webservices and keep the family jewels hidden in the servers.
We do half/half, with some of our code on GitHub and public, but the cooler stuff on a private git server. It doesn't really matter, in the end, as we're pretty much the only people who can read and understand the code, so it being open or closed is largely irrelevant.
There's now so much open-source code, out there, that we've run out of humans to examine it, so code gets rated by the WoT of the developer. As if it were closed-source.
Ain't nobody got time to read all that. 😂