11 ms·

Is there any difference between a human watching 1000 videos on a topic and producing content based on that information, and a machine learning model? If it onl

Is there any difference between a human watching 1000 videos on a topic and producing content based on that information, and a machine learning model? If it only creates original work, what does it matter?

__loam 3y ago
There's a pretty big difference but the aspirant neuroscientists and legal scholars on hackernews would have you believe it's the same thing.
raincole 3y ago
The difference is that one is a human and one is a machine. Just because they do something kinda similar it doesn't make them the same. Selling human organs will never be the same as selling beef liver, while the difference between a human's liver and a cow's is much smaller than the difference between your brain and a A100 tensor core.
supriyo-biswas 3y ago
Because humans are entitled to certain rights and privileges that computer systems are not, and coming up with systems that causes profits to be distributed within the top 1% while using the output of people without compensating them is inherently unfair.
- visarga 3y ago
  When we use gen-AI, the provider makes a few cents, and users get to set the prompt and use the outputs. So benefits go to the prompter more than the provider. It's like buying pills - yes, the drug manufacturer makes a profit, but the patient might save her health - a much larger benefit. Gen-AI is an empowering technology for the masses and easy to run locally. And the providers are in a race to the bottom on pricing.
  - supriyo-biswas 3y ago
    People hate the pharma industry but at least some of them perform their own research, and thus are, at least from a moral standpoint, more entitled to make a profit than the genAI companies where the value primarily comes from the training data. (Whether they should be able to hike drug prices by 100% overnight or sell drugs with insane markups is a separate question.)
  - josephwegner 3y ago
    That’s not how markets work, right? And I don’t think it’s true in either example. If you can charge more, you will. Price should match value. OpenAI is charging a _really_ high monthly fee for ChatGPT, and it’s quite popular. It’s very limited - there’s no way the cost incurred for usage is near that. Obviously their costs include the R&D that has already happened, but I still think it’s priced way over that. Drug manufacturers are famously making money hand over fist. Patients aren’t their customers. Insurance companies are. They are surely gouging insurance companies to their fullest capacity.
    - visarga 3y ago
      > If you can charge more, you will. Price should match value. And big LLM providers will certainly try, but they have competition now, all 3 tied up at roughly the same level. And then there is competition from open models that can handle 50% of what big models do. So they can only set a high price for very advanced/critical tasks, where usage will be much lower.
    - fshbbdssbbgdd 3y ago
      $20/month is high? For a tutor and research assistant who I can consult at any time of day or night? Seems like an incredible value to me. Would have probably cost thousands of dollars in 2020.
CSMastermind 3y ago
Humans learn and synthesize; machines often compress and reproduce. With every new technology, there needs to be legislation and judicial rulings to determine the exact definition of things so that the best people can offer you is speculation or their interpretation. The quality of those interpretations will vary by person; I suspect most will be relatively uninformed. One important factor to consider is the terms of use of the websites from which the data is obtained. For example, YouTube's terms clearly state that content may only be used for personal, non-commercial purposes. Additionally, the current generation of models does not derive core aspects of knowledge or build world models; instead, they compress vast amounts of data and use that information to generate content. This process likely involves storing copyrighted information, which may violate copyright law.
seanhunter 3y ago
From the point of view of the economics of the creation of videos the difference is enormous. People produce videos hoping or expecting to be monetized. People who monetize videos (by advertising or sponsoring) do so hoping for the audience to buy some product or service. People who host videos do so expecting to be paid by the monetizing people. That's how the current economic model for ad-supported video works. A machine learning model is not an economic actor[1]. It's not going to go out and buy a flat wallet or a new set of headphones or whatever else is being advertised on stream. So the current economic model under which that content produced (which is predicated on the assumption that the audience is human) totally falls apart if the audience is an ML model. [1] yet. And even when they routinely can take part in economic transactions they aren't the audience the advertisers are paying to reach.
- ecoquant 3y ago
  The irony is that youtube was largely built on copyright violation with taking everything that existed in video before youtube and giving it away for free.

__loam 3y ago

There's a pretty big difference but the aspirant neuroscientists and legal scholars on hackernews would have you believe it's the same thing.

raincole 3y ago

The difference is that one is a human and one is a machine. Just because they do something kinda similar it doesn't make them the same. Selling human organs will never be the same as selling beef liver, while the difference between a human's liver and a cow's is much smaller than the difference between your brain and a A100 tensor core.

supriyo-biswas 3y ago

Because humans are entitled to certain rights and privileges that computer systems are not, and coming up with systems that causes profits to be distributed within the top 1% while using the output of people without compensating them is inherently unfair.

visarga 3y ago
When we use gen-AI, the provider makes a few cents, and users get to set the prompt and use the outputs. So benefits go to the prompter more than the provider. It's like buying pills - yes, the drug manufacturer makes a profit, but the patient might save her health - a much larger benefit. Gen-AI is an empowering technology for the masses and easy to run locally. And the providers are in a race to the bottom on pricing.
- supriyo-biswas 3y ago
  People hate the pharma industry but at least some of them perform their own research, and thus are, at least from a moral standpoint, more entitled to make a profit than the genAI companies where the value primarily comes from the training data. (Whether they should be able to hike drug prices by 100% overnight or sell drugs with insane markups is a separate question.)
- josephwegner 3y ago
  That’s not how markets work, right? And I don’t think it’s true in either example. If you can charge more, you will. Price should match value. OpenAI is charging a _really_ high monthly fee for ChatGPT, and it’s quite popular. It’s very limited - there’s no way the cost incurred for usage is near that. Obviously their costs include the R&D that has already happened, but I still think it’s priced way over that. Drug manufacturers are famously making money hand over fist. Patients aren’t their customers. Insurance companies are. They are surely gouging insurance companies to their fullest capacity.
  - visarga 3y ago
    > If you can charge more, you will. Price should match value. And big LLM providers will certainly try, but they have competition now, all 3 tied up at roughly the same level. And then there is competition from open models that can handle 50% of what big models do. So they can only set a high price for very advanced/critical tasks, where usage will be much lower.
  - fshbbdssbbgdd 3y ago
    $20/month is high? For a tutor and research assistant who I can consult at any time of day or night? Seems like an incredible value to me. Would have probably cost thousands of dollars in 2020.

CSMastermind 3y ago

Humans learn and synthesize; machines often compress and reproduce. With every new technology, there needs to be legislation and judicial rulings to determine the exact definition of things so that the best people can offer you is speculation or their interpretation. The quality of those interpretations will vary by person; I suspect most will be relatively uninformed. One important factor to consider is the terms of use of the websites from which the data is obtained. For example, YouTube's terms clearly state that content may only be used for personal, non-commercial purposes. Additionally, the current generation of models does not derive core aspects of knowledge or build world models; instead, they compress vast amounts of data and use that information to generate content. This process likely involves storing copyrighted information, which may violate copyright law.

seanhunter 3y ago

From the point of view of the economics of the creation of videos the difference is enormous. People produce videos hoping or expecting to be monetized. People who monetize videos (by advertising or sponsoring) do so hoping for the audience to buy some product or service. People who host videos do so expecting to be paid by the monetizing people. That's how the current economic model for ad-supported video works. A machine learning model is not an economic actor[1]. It's not going to go out and buy a flat wallet or a new set of headphones or whatever else is being advertised on stream. So the current economic model under which that content produced (which is predicated on the assumption that the audience is human) totally falls apart if the audience is an ML model. [1] yet. And even when they routinely can take part in economic transactions they aren't the audience the advertisers are paying to reach.

ecoquant 3y ago
The irony is that youtube was largely built on copyright violation with taking everything that existed in video before youtube and giving it away for free.