6 ms·

It was more than 10 years ... People reacted the same way to SHRDLU 40+ years ago, and 20 years ago when I encountered it: https://en.wikipedia.org/wiki/SHRDLU

by chubot 5y ago

It was more than 10 years ... People reacted the same way to SHRDLU 40+ years ago, and 20 years ago when I encountered it: https://en.wikipedia.org/wiki/SHRDLU https://en.wikipedia.org/wiki/SHRDLU There's definitely an illusion where we anthropomorphize very simple software, ascribe intention to it, etc.

p1esk 5y ago
I'm not 100% sure, but I think in 2012 chatbots still used handcoded rules. So we switched from handcoded rules to trainable neural networks in just a few years. Models like GPT-3 are interesting because they are conceptually simple, and are able to do (previously) complicated tasks without any explicit training (e.g. simply by processing lots of text and trying to predict the next word). This is a huge advancement in AI field, even if many more advancements are needed to get to the human level.
- fault1 5y ago
  I remember chatbots in 2000 using hidden markov models (and sometimes even neural nets), so from that standpoint, they aren't exactly new. The models are just larger now.
  - JamesBarney 5y ago
    Never underestimate the effects of scaling on intelligence. Caenorhabditis elegans neurons work a lot like our neurons, we just have a lot more of them.
    - fault1 5y ago
      I agree with you. I also believe "AI" will always be a moving target: https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect Certainly in the 1950s, most automatic control systems must have seemed magical, it was to Nobert Wiener, even if they were "just" an evolution of the steam governor. In the end, it depends on what you qualify as intelligence.
      - pdimitar 5y ago
        > I also believe "AI" will always be a moving target: https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect No, it absolutely is not. Everyone I spoke with in the 90s (and myself) still have the same requirements: be able to make sense of the 3D material world around you and interact with it without harming humans or valuable possessions. Maybe you spent too much time with critics that can never be satisfied. Sucks to be you if that's the case but don't think for a second that they represent most of humanity. Most of us want to spend less time cleaning or doing stuff in the kitchen, and us the programmers in particular would just be grateful for faster compilers and a bit more declarative programming that generates actual code.
        fault1 5y ago
        Of course, I love the spoils of technology and automation as much as anyone else. But it is absolutely human nature to get used to things and not consider them magical anymore, things simply become the new norm. This is exactly what happened with feedback controllers like airplane autopilot systems (1912 - https://en.wikipedia.org/wiki/Autopilot https://en.wikipedia.org/wiki/Autopilot) I've worked for a number of years on industrial robotics, which sounds very similar to your definition of "AI": real physical machines that take data from various sensors, including spatial sensors, make sense of them in real time, and decide how to optimally interact with the physical environment, with safety critical systems in mind. I hardly think about such systems as AI, more simply engineering, math, and a lot of coding.
        pdimitar 5y ago
        Hmmm. But I really didn't mean robots with mostly hardcoded requirements and a few excellent optical recognition algorithms (which might truly be the real boss here). I actually do mean a walking robot that can find the exact knife it needs to fillet a fish, regardless of whether it's in the sink, in its proper place on a stand, on the table, or dropped on the floor (and if it finds it on the floor it will wash it before usage first). I mean a robot that can open the fridge and find the tomatoes, regardless if your brother has moved them to the top shelf or if they are where they should be. Etc.
- adament 5y ago
  I think MegaHAL is (mostly?) just a markov model, and I think that was not exactly new when I looked at it around 2002. As I recall it was easier to distinguish from a human than Eliza, since it had a greater probability of spouting nonsense, but it was still amusing and fun to play with. Personal anecdote: I read one of Ray Kurzweils books in school back then, completely misunderstood how neural networks worked and ended up making a markov model chatbot with single word states.
- majormajor 5y ago
  > I'm not 100% sure, but I think in 2012 chatbots still used handcoded rules. So we switched from handcoded rules to trainable neural networks in just a few years. Models like GPT-3 are interesting because they are conceptually simple, and are able to do (previously) complicated tasks without any explicit training (e.g. simply by processing lots of text and trying to predict the next word). This is a huge advancement in AI field, even if many more advancements are needed to get to the human level. This seems like a very optimistic take on the timeline. It took us 50 years to go from ELIZA to GPT-3. And it still is "dumb" compared to human intelligence. So how long for the next major achievements? Are we talking years for each, or more decades?

by chubot 5y ago

p1esk 5y ago

I'm not 100% sure, but I think in 2012 chatbots still used handcoded rules. So we switched from handcoded rules to trainable neural networks in just a few years. Models like GPT-3 are interesting because they are conceptually simple, and are able to do (previously) complicated tasks without any explicit training (e.g. simply by processing lots of text and trying to predict the next word). This is a huge advancement in AI field, even if many more advancements are needed to get to the human level.

fault1 5y ago
I remember chatbots in 2000 using hidden markov models (and sometimes even neural nets), so from that standpoint, they aren't exactly new. The models are just larger now.
- JamesBarney 5y ago
  Never underestimate the effects of scaling on intelligence. Caenorhabditis elegans neurons work a lot like our neurons, we just have a lot more of them.
  - fault1 5y ago
    I agree with you. I also believe "AI" will always be a moving target: https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect Certainly in the 1950s, most automatic control systems must have seemed magical, it was to Nobert Wiener, even if they were "just" an evolution of the steam governor. In the end, it depends on what you qualify as intelligence.
    - pdimitar 5y ago
      > I also believe "AI" will always be a moving target: https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect No, it absolutely is not. Everyone I spoke with in the 90s (and myself) still have the same requirements: be able to make sense of the 3D material world around you and interact with it without harming humans or valuable possessions. Maybe you spent too much time with critics that can never be satisfied. Sucks to be you if that's the case but don't think for a second that they represent most of humanity. Most of us want to spend less time cleaning or doing stuff in the kitchen, and us the programmers in particular would just be grateful for faster compilers and a bit more declarative programming that generates actual code.
      - fault1 5y ago
        Of course, I love the spoils of technology and automation as much as anyone else. But it is absolutely human nature to get used to things and not consider them magical anymore, things simply become the new norm. This is exactly what happened with feedback controllers like airplane autopilot systems (1912 - https://en.wikipedia.org/wiki/Autopilot https://en.wikipedia.org/wiki/Autopilot) I've worked for a number of years on industrial robotics, which sounds very similar to your definition of "AI": real physical machines that take data from various sensors, including spatial sensors, make sense of them in real time, and decide how to optimally interact with the physical environment, with safety critical systems in mind. I hardly think about such systems as AI, more simply engineering, math, and a lot of coding.
        pdimitar 5y ago
        Hmmm. But I really didn't mean robots with mostly hardcoded requirements and a few excellent optical recognition algorithms (which might truly be the real boss here). I actually do mean a walking robot that can find the exact knife it needs to fillet a fish, regardless of whether it's in the sink, in its proper place on a stand, on the table, or dropped on the floor (and if it finds it on the floor it will wash it before usage first). I mean a robot that can open the fridge and find the tomatoes, regardless if your brother has moved them to the top shelf or if they are where they should be. Etc.
adament 5y ago
I think MegaHAL is (mostly?) just a markov model, and I think that was not exactly new when I looked at it around 2002. As I recall it was easier to distinguish from a human than Eliza, since it had a greater probability of spouting nonsense, but it was still amusing and fun to play with. Personal anecdote: I read one of Ray Kurzweils books in school back then, completely misunderstood how neural networks worked and ended up making a markov model chatbot with single word states.
majormajor 5y ago
> I'm not 100% sure, but I think in 2012 chatbots still used handcoded rules. So we switched from handcoded rules to trainable neural networks in just a few years. Models like GPT-3 are interesting because they are conceptually simple, and are able to do (previously) complicated tasks without any explicit training (e.g. simply by processing lots of text and trying to predict the next word). This is a huge advancement in AI field, even if many more advancements are needed to get to the human level. This seems like a very optimistic take on the timeline. It took us 50 years to go from ELIZA to GPT-3. And it still is "dumb" compared to human intelligence. So how long for the next major achievements? Are we talking years for each, or more decades?