4 ms·

I think you're describing an inevitable endgame of information technology: Value of software tends towards 0, as it can be copied with very low cost, while valu

I think you're describing an inevitable endgame of information technology: Value of software tends towards 0, as it can be copied with very low cost, while value of data tends towards the value given by scarcity and actionable information of that data. In there you've encoded the value of business processes within the data economy. Which is why long term, only the services model can endure while pure software vendors will have to either adapt or die. Among the traditional software companies, Microsoft has been the strongest in adapting to this new paradigm. I only wonder though, how long this economy is going to last, i.e. how long until AI is so commoditized that the value of data goes towards 0 as well, along with the value of most education in IT.

starbugs 7y ago
That doesn't quite match the reality that I can see. A static piece of code may tend to 0 (business) value, but software that is actively being used and maintained tends to have significant value. That is probably one of the key reasons why the service model is so successful. Static software's value rots away quickly, but software that is actively being used appears to be getting more valuable over time. If and when the point of advanced AI that you describe is reached, does it still matter? Everything will have lost its value by then unless society changes its mind about what a machine with that level of intelligence is. Just my 2 cents.
- m_mueller 7y ago
  I get the impression that we're actually in agreement, just with different wording. For me the term 'software' is describing exactly that, a static piece of code. The usage part is encoded in data that goes back to the developers - but if all they do with it is provide another static piece of code (e.g. an upgrade), we're at the same starting point again. Now, if you merge together development and operations, such that usage data is always there to grab, analyse and push down a development pipeline - that's where the services model is shining, but not thanks to a piece of code, but due to tons of business processes running that others can't easily replicate.
  - starbugs 7y ago
    Right, I guess we agree then. It's just that it doesn't work without writing that piece of software which gives it its value. But hasn't it been like that anyway all the time? Just in a slower way with more primitive distribution channels and less data flowing around?
    - m_mueller 7y ago
      I think the distribution channels are exactly what changed - 25 years ago they were only available to large organizations, while now they're free and commoditized.
bunderbunder 7y ago
> I only wonder though, how long this economy is going to last, i.e. how long until AI is so commoditized that the value of data goes towards 0 as well Well, first - AI in general is not very commoditized at all. It's just machine learning that is becoming commoditized. And the distinction is meaningful. Going from that - data is the stuff you use to build ML models, and it's absolutely essential to doing better ML. ML becoming commoditized increases the value of (good) data, because it's an essential ingredient for any ML company trying to gain an edge in an increasingly competitive market. It's not entirely unlike how oil being a commodity does not mean that oil drilling equipment is cheap. We see this play out in the companies that are big users of machine learning. Facebook, Microsoft, Google, etc. are all quite happy to publish the software they use to build their models. But the data is kept under lock and key and buried under the bottom of the sea. And a huge, huge, huge amount of their competition is centered around attracting the people who produce that data, because it's the ability to generate more data that gives them a competitive edge.
- m_mueller 7y ago
  I absolutely see the point about the value of data going up, and again, that's the paradigm we're living in now. It's just that I'm not sure this is going to last for long. What's new in ML is, at its core, increasingly powerful pattern matching. But with even some moderately strong cognitive functions on top of that, the data processing monster could start feeding on itself, i.e. go out to get new data on its own. We don't have enough images of donkeys feeding on grass for our object recognizer to match against? Let me coordinate with our drone network to go get a couple of terrabytes...
  - codesushi42 7y ago
    That's not how ML works. You need to acquire labeled data to train on.
    - m_mueller 7y ago
      Again, I'm talking about possible future technology. What if you can infer labels some other way? The fundamental element you need is not labels but a cost function - that could e.g. be curiosity and fear (of disappointing the end user). The most fundamental part that ML is missing to be AI, is to have a model of the world, with which such a cost function could evolve organically. Once you have that, AI can go and explore on its own.
justicezyx 7y ago
> Value of software tends towards 0 Is there anything that can be produced of which the value doesn't "tends towards 0"?
- m_mueller 7y ago
  Anything that's scarce, either in availability (raw materials) or time required using the most efficient method to acquire it. Edit: And for which there is demand.

starbugs 7y ago

That doesn't quite match the reality that I can see. A static piece of code may tend to 0 (business) value, but software that is actively being used and maintained tends to have significant value. That is probably one of the key reasons why the service model is so successful. Static software's value rots away quickly, but software that is actively being used appears to be getting more valuable over time. If and when the point of advanced AI that you describe is reached, does it still matter? Everything will have lost its value by then unless society changes its mind about what a machine with that level of intelligence is. Just my 2 cents.

m_mueller 7y ago
I get the impression that we're actually in agreement, just with different wording. For me the term 'software' is describing exactly that, a static piece of code. The usage part is encoded in data that goes back to the developers - but if all they do with it is provide another static piece of code (e.g. an upgrade), we're at the same starting point again. Now, if you merge together development and operations, such that usage data is always there to grab, analyse and push down a development pipeline - that's where the services model is shining, but not thanks to a piece of code, but due to tons of business processes running that others can't easily replicate.
- starbugs 7y ago
  Right, I guess we agree then. It's just that it doesn't work without writing that piece of software which gives it its value. But hasn't it been like that anyway all the time? Just in a slower way with more primitive distribution channels and less data flowing around?
  - m_mueller 7y ago
    I think the distribution channels are exactly what changed - 25 years ago they were only available to large organizations, while now they're free and commoditized.

bunderbunder 7y ago

> I only wonder though, how long this economy is going to last, i.e. how long until AI is so commoditized that the value of data goes towards 0 as well Well, first - AI in general is not very commoditized at all. It's just machine learning that is becoming commoditized. And the distinction is meaningful. Going from that - data is the stuff you use to build ML models, and it's absolutely essential to doing better ML. ML becoming commoditized increases the value of (good) data, because it's an essential ingredient for any ML company trying to gain an edge in an increasingly competitive market. It's not entirely unlike how oil being a commodity does not mean that oil drilling equipment is cheap. We see this play out in the companies that are big users of machine learning. Facebook, Microsoft, Google, etc. are all quite happy to publish the software they use to build their models. But the data is kept under lock and key and buried under the bottom of the sea. And a huge, huge, huge amount of their competition is centered around attracting the people who produce that data, because it's the ability to generate more data that gives them a competitive edge.

m_mueller 7y ago
I absolutely see the point about the value of data going up, and again, that's the paradigm we're living in now. It's just that I'm not sure this is going to last for long. What's new in ML is, at its core, increasingly powerful pattern matching. But with even some moderately strong cognitive functions on top of that, the data processing monster could start feeding on itself, i.e. go out to get new data on its own. We don't have enough images of donkeys feeding on grass for our object recognizer to match against? Let me coordinate with our drone network to go get a couple of terrabytes...
- codesushi42 7y ago
  That's not how ML works. You need to acquire labeled data to train on.
  - m_mueller 7y ago
    Again, I'm talking about possible future technology. What if you can infer labels some other way? The fundamental element you need is not labels but a cost function - that could e.g. be curiosity and fear (of disappointing the end user). The most fundamental part that ML is missing to be AI, is to have a model of the world, with which such a cost function could evolve organically. Once you have that, AI can go and explore on its own.

justicezyx 7y ago

> Value of software tends towards 0 Is there anything that can be produced of which the value doesn't "tends towards 0"?

m_mueller 7y ago
Anything that's scarce, either in availability (raw materials) or time required using the most efficient method to acquire it. Edit: And for which there is demand.