private/secret repositories where they worked on things without sharing with the world
There’s a name for this. Closed source. It might not technically be closed source, but it is in principle. Namely, getting to some result and making an impact on the world, without sharing the process of getting there.
They didn’t want to post to a Stack Overflow site because of “harvesters and scrapers digesting my answer and not giving me one cent or even credit”
they didn’t want to deal with the assholes of “Open Source”
These are almost the exact reasons why many choose to go closed source. Why FUTO has a weird license. Why Bitwarden just changed the license of all their apps.
“Open source” acts like people are willing to give things for free, and ask for nothing in return. In reality that is rarely true. Everybody has different ideas of what is proper compensation. Maybe it’s money. Maybe attribution. But right now everybody disagrees with each other, and people will choose closed source to be safe and avoid feeling exploited.
If this continues, then AI will destroy open source. Hopefully we figure out a solution soon.
I’ve been pretty pissed off at “helping bad people” with my freely given information since roughly the summer of 2022.
The true test of our industry and community; will or how much will the current felt theft non-attribution and non-appreciation of good-will work influence or destroy good-will work like FOSS, free public resources and commons, and their communities.
Good-will work was, presumably, always thankless for most people. No monetary gain, little direct feedback and appreciation or acknowledgement.
Now that, seemingly, AI becomes the consumer and interface instead of people, it feels even worse. I would understand if this tips the scales for many people.
I also think the attitude toward it would be more positive if those in the spotlight for training these models weren’t actively creating global economic crises, bringing down random websites with sustained high-volume traffic, filibustering collaboration with low quality issues and PRs, and completely disregarding the licenses for the code they’re using.
In fact, were these models only used to better the world, I’m sure that many developers would be more than happy to support the effort even if it meant licensing the code for training under more permissive licenses (assuming they could get their contributors on board as well).
Instead, we all get to hear about how some egotistical billionaire is trying to become a trillionaire, or how the prices of everything skyrocketing is a good thing because some rich people got more rich.
Definitely. These bad-actor activities further taint them and paint them as bad actors you don’t want to enable.
I understand the sentiment. It’s frustrating knowing that every word you write and every line of code you put on the internet is going to be used to train models.
This post reminded me of the spirited discussion regarding Mark Pilgrim’s deletion of his website back in the day.
I don’t thing the fsf deserve to be catching strays on this
The advent of Turing-test-passing AI was such an enormous mistake. There’s no shortage of people who want to do creative, intellectual work. People even do it in their spare time and give the product away. They’re trying to automate the wrong kinds of jobs.
Turns out the wrong kinds of jobs are easier to automate. But once we get those done away with, it’s a lot of more labour in physical jobs and we can drive those costs down. Great, eh?
Tbf we’ll need to automate everything we can because we aren’t producing enough kids to keep this upside down pyramid going for much longer unfortunately. By we, I mean most of the developed world, and the developing world will get there in a few decades too.
Just to add an analysis of UN numbers on this.
Wait until the printing press arrives, all the scribe monks will become jobless and no one will have experience anymore.
Wait until the cotton gin arrives, all the hand weavers will become jobless and no one will have experience anymore.
Wait until the computer arrives, people will become obsolete and no one will have experience anymore.
Wait until the steam engine, automobiles, planes, calculators, telephones, bla bla
“Slippery slope” arguments like these assume that once a technology introduces a specific risk, society lacks the agency to create counter-measures, new norms, or alternative uses for that technology.
History shows that when a “slope” appears, regulation steps in, technology evolves to solve the problem, or the culture shifts to reinterpret the tool.
In almost every case, the feared “bottom” of the slope was never reached because humans constantly built ramps or bridges along the way.
None of those tools could do the decision making part.
I see this comparison so often and it’s incredibly disingenuous. An LLM can iterate over a problem and resolve it over time. It can debug, fact check, create and test hypotheses.
That is distinctly different from arithmetic but faster or writing but faster.
That is distinctly different from arithmetic but faster or writing but faster.
But not “problem-solving but faster”.
Correct, yes.
Prior to this technology there was some “higher” or more complex skill that could be learned that encorporated the new tool. The human brain stays relevant only in a new and more expansive space.
This tool does not do that. There is nothing more advanced the human brain can do beyond problem solving.
There is an entire field of research dedicated to this; RBV. Things like a lathe or a calculator are “complementary tools” under RBV, humans can utilize them without loss of the skill that enables them. LLM’s are “substitutive” tools under RBV, they replace human skills without enhancement.
Here, there’s a whole paper about it.
https://sms.onlinelibrary.wiley.com/doi/full/10.1002/smj.3387
I’m a bit afraid to rely on common sense or regulations being reasonable, but yes, we never reached the bottom yet
This resonates with me. I feel like the world is worse off for it, but I don’t see a solution for it.
I definitely commit a lot more to my private forgejo than codeberg account anymore despite everything I do being specifically AGPL3.
