Likely progressive, but definitely not luddite [0]. Anti-capitalist for sure.
I struggle with this discourse deeply. With many posters like OP, I align almost completely - unions are good, large megacorps are bad, death to facists etc. It's when we get to the AI issue that I do a bit of a double take.
Right now, AI is almost completely in the hands of a few large corp entities, yes. But once upon a time, so was the internet, so were processing chips, so was software. This is the power of the byte - it shrinks progressively and multiplies infinitely - thus making it inherently diffuse and populist (at the end of the day). It's not the relationship to our cultural standards that causes this - it's baked right into the structure of the underlying system. Computing systems are like sand - you can melt them into a tower of glass, but those are fragile and will inevitably become sand once again. Sand is famously difficult to hold in a tight grasp.
I won't say that we should stop fighting against the entrenchment of powers like OpenAI - fine, that's potentially a worthy fight and if that's what you want to focus on go ahead. However, if you really want to hack the planet, democratize power and distribute control, what you have to be doing is working towards smaller local models, distributed training, and finding an alternative to backprop that can compete without the same functional costs.
We are this close to having a guide in our pocket that can help us understand the machine better. Forget having AI "do the work" for you, it can help you to grok the deeper parts of the system such that you can hack them better - and if we're to come out of this tectonic shift in tech with our heads above water, we absolutely need to create models that cannot be owned by the guy with the $5B datacenter.
Deepseek shows us the glimmer of a way forward. We have to take it. The megacorp AI is already here to stay, and the only panacea is an AI that they cannot control. It all comes down to whether or not you genuinely believe that the way of the hacker can overcome the monolith. I, for one, am a believer.
Not true for the Internet. It was the open system anyone could join and many people were shocked it succeeded over the proprietary networks being developed.
It started as ARPANET, which was not an open system (slightly sloppy use of the term i know, not technically the internet but I consider one the extension of the other).
How are unions any better than mega corps? My brother is part of a union and the leaders make millions.
He's pigenholed at the same low pay rate and can't ever get a raise, until everyone in the same role also gets a raise (which will never happen). It traps people, because many union jobs can't or won't innovate, and when they look elsewhere, are underskilled (and stuck).
You mention 'deepseek'. Are you joking? It's owned by the Chinese government..and you claim to hate fascism? Lol?
Big companies only have the power now, because the processing power to run LLMs is expensive. Once there are break throughs, anyone can have the same power in their house.
We have been in a tech slump for awhile now. Large companies will drive innovations for AI that will help everyone.
That's not a union - that's another corporate entity parading as a union. A union, operating as it should, is governed by the workers as a collective and enriches all of them at the same rate.
Deepseek is open source, which is why I mention it. It was made by the Chinese government but it shows a way to create these models at vastly reduced cost and was done with transparent methodology so we can learn from it. I am not saying "the future is Deepseek", I am saying "there are lessons to be learned from Deepseek".
I actually agree with you on the corporate bootstrap argument - I think we ought to be careful, because if they ever figure out how to control the output they will turn off outputs that help develop local models (gotta protect that moat!), but for now I use them myself to study and learn about building locally and I think everyone else ought to get on this train as well. For now, the robust academic discourse is a very very good thing.
I've been on both sides of it (in a union and working with union members when there was no other choice).
They all work the same way. I'm fundamentally against the idea of unions after seeing how they stifle innovation in nearly all industries they control.
I struggle with this discourse deeply. With many posters like OP, I align almost completely - unions are good, large megacorps are bad, death to facists etc. It's when we get to the AI issue that I do a bit of a double take.
Right now, AI is almost completely in the hands of a few large corp entities, yes. But once upon a time, so was the internet, so were processing chips, so was software. This is the power of the byte - it shrinks progressively and multiplies infinitely - thus making it inherently diffuse and populist (at the end of the day). It's not the relationship to our cultural standards that causes this - it's baked right into the structure of the underlying system. Computing systems are like sand - you can melt them into a tower of glass, but those are fragile and will inevitably become sand once again. Sand is famously difficult to hold in a tight grasp.
I won't say that we should stop fighting against the entrenchment of powers like OpenAI - fine, that's potentially a worthy fight and if that's what you want to focus on go ahead. However, if you really want to hack the planet, democratize power and distribute control, what you have to be doing is working towards smaller local models, distributed training, and finding an alternative to backprop that can compete without the same functional costs.
We are this close to having a guide in our pocket that can help us understand the machine better. Forget having AI "do the work" for you, it can help you to grok the deeper parts of the system such that you can hack them better - and if we're to come out of this tectonic shift in tech with our heads above water, we absolutely need to create models that cannot be owned by the guy with the $5B datacenter.
Deepseek shows us the glimmer of a way forward. We have to take it. The megacorp AI is already here to stay, and the only panacea is an AI that they cannot control. It all comes down to whether or not you genuinely believe that the way of the hacker can overcome the monolith. I, for one, am a believer.
0 - https://phrack.org/issues/7/3