Are large language models a threat to digital public goods?
1–10 of 163 posts
Re: Are large language models a threat to digital public goods?
#2I think one thing that would tip this in a positive way is easy re-sharing of learning. I have plenty of "source code" now for conversations where I've taken something that was hard to find out (or code that was previously hard to write) through a non-LLM route. Publishing those conversations isn't as easy or re-adoptable into the commons as it could be.
(I'll note that proprietary models would be first-order disincentivised to stop this because such conversations can be used to better train other models, but of course they will benefit more generally for maintaining a commons of knowledge to draw from.)
Re: Are large language models a threat to digital public goods?
#3Does this mean we can’t adopt new languages since ChatGPT is frozen to 2022 coding knowledge?
Re: Are large language models a threat to digital public goods?
#4Of course they are. Experts will stop posting for free when their output is stolen and displayed without attribution.
Re: Are large language models a threat to digital public goods?
#5This is a very intriguing study as it would appear the fact that LLMs need public data to train on, its incentive on the marketplace of ideas is to reduce the very thing that gives it its power. It contains the seeds of its own destruction, destroying the web and open data ethos, and yet another data point pointing at another AI winter.
Re: Are large language models a threat to digital public goods?
#6Does this mean we can’t adopt new languages since ChatGPT is frozen to 2022 coding knowledge?
Yes, exactly. It's all over.
Re: Are large language models a threat to digital public goods?
#7This is a very intriguing study as it would appear the fact that LLMs need public data to train on, its incentive on the marketplace of ideas is to reduce the very thing that gives it its power. It contains the seeds of its own destruction, destroying the web and open data ethos, and yet another data point pointing at another AI winter.
AI winter? Labs haven't even began to scratch the surface of the data available to train models on. We may be running out of quality human-written text but we have yet to dive into:
Video Audio Images Heat Motion/acceleration Lidar RF Sonar Radar Network traffic Atmospheric pressure Wind vectors Magnetic fields System/application logs Electrical current UV X-ray Microwave Ionizing particle emissions
Re: Are large language models a threat to digital public goods?
#8This is a very intriguing study as it would appear the fact that LLMs need public data to train on, its incentive on the marketplace of ideas is to reduce the very thing that gives it its power. It contains the seeds of its own destruction, destroying the web and open data ethos, and yet another data point pointing at another AI winter.
For reasons I can't articulate I see LLMs as a vehicle for removing the creators from their ideas. This is very different than search engines. If a search engine generates traffic for documented ideas it creates a community. An LLM based internet seems to remove the creator and shim itself in between for the sake of business.
Re: Are large language models a threat to digital public goods?
#9This is a very intriguing study as it would appear the fact that LLMs need public data to train on, its incentive on the marketplace of ideas is to reduce the very thing that gives it its power. It contains the seeds of its own destruction, destroying the web and open data ethos, and yet another data point pointing at another AI winter.
AI winter? Labs haven't even began to scratch the surface of the data available to train models on. We may be running out of quality human-written text but we have yet to dive into: Video Audio Images Heat Motion/acceleration Lidar RF Sonar Radar Network traffic Atmospheric pressure Wind vectors Magnetic fields System/application logs Electrical current UV X-ray Microwave Ionizing particle emissions
Interesting, so we chuck in exabytes and more of data generated each day and then what?
Re: Are large language models a threat to digital public goods?
#10This is a very intriguing study as it would appear the fact that LLMs need public data to train on, its incentive on the marketplace of ideas is to reduce the very thing that gives it its power. It contains the seeds of its own destruction, destroying the web and open data ethos, and yet another data point pointing at another AI winter.
For reasons I can't articulate I see LLMs as a vehicle for removing the creators from their ideas. This is very different than search engines. If a search engine generates traffic for documented ideas it creates a community. An LLM based internet seems to remove the creator and shim itself in between for the sake of business.
If a handful of firms are indeed allowed to harvest he collective knowledge for rent seeking that would indeed be a shame.