4 ms·

When OpenAI started I don't think anyone would have thought that AI this powerful was around the corner.

edgyquant 3y ago
This is just more openAI marketing at work. ChatGPT is a chatbot, and it’s good at language for sure, but it isn’t some huge breakthrough that people were expecting would take a century. Its literally just a fancy RNN which we’ve had for awhile
- famouswaffles 3y ago
  No offense but this comment just shows a lack of understanding of the development that's happened with transformers. "Fancy RNN" is a pretty ridiculous assertion. And no, researchers didn't expect what GPT-3/4 has been shown to do to be around the corner at all. GPT's aren't chatbots. That's just a neat natural consequence that's happened. They're machines that reason, understand and follow instructions in plain language. And their abilities go far beyond what any expected language modelling to provide.
  - edgyquant 3y ago
    Nah not really, there’s just a lot of hyperbole going around right now. They are RNNs, the difference is they have a layer/layers being fed back instead of a single neuron. That’s a fancy recurrent network. They are not machines that reason, they are approximators. It’s all just token matching based on data we’ve fed it. Further it didn’t happen over night but through successive improvements and at no point was the next improvement considered some infeasible thing. Seems to me maybe you’ve bought into the hype here, and are confusing that with reality.
    - famouswaffles 3y ago
      >They are not machines that reason, they are approximators. And meaningless distinction of the year award goes to.. "It's not real [insert property]" is not an intelligent argument. By all means, divine the way to distinguish results of the two. After all, what kind of important distinction can't be distinguished by results? >Further it didn’t happen over night but through successive improvements The only difference between GPT-3 and GPT-2 was scale. They didn't even change the tokenizer until 4. There were no "successful improvements" to smoothen the massive gap in capabilities between the two. So to say that was expected just shows more lack of knowledge here.
      - airgapstopgap 3y ago
        > The only difference between GPT-3 and GPT-2 was scale. They didn't even change the tokenizer until 4. There were no "successful improvements" to smoothen the massive gap in capabilities between the two. "Other than that, how was the play, Mrs. Lincoln?" GPT-2 is two orders of magnitude smaller. In terms of forebrain neuron count this is the difference between a human and an elephant shrew or budgerigar. It was absolutely expected by reasonable people that 2 OOMs of scaling will provide a qualitative jump.
        famouswaffles 3y ago
        >GPT-2 is two orders of magnitude smaller. In terms of forebrain neuron count this is the difference between a human and an elephant shrew or budgerigar. It was absolutely expected by reasonable people that 2 OOMs of scaling will provide a qualitative jump. When GPT-3 was released, it was by far the largest artificial neural network ever trained. And I mean by far. Now there wasn't any big jump in hardware capabilities to spur this sort of gulf. It wasn't a case of "Oh now we can train a very large model" So Want to know why there was such a gap ? It's because most researchers assumed the models would overfit the data or display diminishing returns long before 175b. Brain neurons are not comparable to ann parameters.
      - Reki 3y ago
        You would expect a machine that can reason to improve in its responses as a conversation progresses, instead of hallucinating.
    - chpatrick 3y ago
      > They are not machines that reason, they are approximators. They approximate a function that performs reasoning...
  - airgapstopgap 3y ago
    Researchers aren't a monolith. Just because Gary Marcus (a complete fraud by the way, look up his "XProp" magitech he sold to Uber) pooh-poohed connectionist AI until ChatGPT came out doesn't mean nobody predicted gains from scaling; certainly many people deep in the know expected things to click. Certainly they expected that the basic trick of associative learning will be cracked. E.g. here's Shane Legg, December 2009 [1]: > Conclusion: computer power is unlikely to be the issue anymore in terms of AGI being possible. The main question is whether we can find the right algorithms. > One of the big things influencing me this year has been learning about how much we understand about how the brain works, in particular, how much we know that should be of interest to AGI designers. I won’t get into it all here, but suffice to say that just a brief outline of all this information would be a 20 page journal paper (there is currently a suggestion that I write such a paper next year with some Gatsby Unit neuroscientists, but for the time being I’ve got too many other things to attend to). At a high level what we are seeing in the brain is a fairly sensible looking AGI design. You’ve got hierarchical temporal abstraction formed for perception and action combined with more precise timing motor control, with an underlying system for reinforcement learning. The reinforcement learning system is essentially a type of temporal difference learning though unfortunately at the moment there is evidence in favour of actor-critic, Q-learning and also Sarsa type mechanisms — this picture should clear up in the next year or so. The system contains a long list of features that you might expect to see in a sophisticated reinforcement learner such as pseudo rewards for informative queues, inverse reward computations, uncertainty and environmental change modelling, dual model based and model free modes of operation, things to monitor context, it even seems to have mechanisms that reward the development of conceptual knowledge. When I ask leading experts in the field whether we will understand reinforcement learning in the human brain within ten years, the answer I get back is “yes, in fact we already have a pretty good idea how it works and our knowledge is developing rapidly.” > I suspect that for the next 5 years, and probably longer, neuroscientists working on understanding cortex aren’t going to be of much use to AGI efforts. My guess is that sometime in the next 10 years developments in deep belief networks, temporal graphical models, liquid computation models, slow feature analysis etc. will produce sufficiently powerful hierarchical temporal generative models to essentially fill the role of cortex within an AGI. > Right, so my prediction for the last 10 years has been for roughly human level AGI in the year 2025 (though I also predict that sceptics will deny that it’s happened when it does!) This year I’ve tried to come up with something a bit more precise. In doing so what I’ve found is that while my mode is about 2025, my expected value is actually a bit higher at 2028. This is not because I’ve become more pessimistic during the year, rather it’s because this time I’ve tried to quantify my beliefs more systematically and found that the probability I assign between 2030 and 2040 drags the expectation up. Perhaps more useful is my 90% credibility region, which from my current belief distribution comes out at 2018 to 2036. And here's Rich Sutton's famous Bitter Lesson, a month after GPT-2 [2]: > We have to learn the bitter lesson that building in how we think we think does not work in the long run. The bitter lesson is based on the historical observations that 1) AI researchers have often tried to build knowledge into their agents, 2) this always helps in the short term, and is personally satisfying to the researcher, but 3) in the long run it plateaus and even inhibits further progress, and 4) breakthrough progress eventually arrives by an opposing approach based on scaling computation by search and learning. The eventual success is tinged with bitterness, and often incompletely digested, because it is success over a favored, human-centric approach. > One thing that should be learned from the bitter lesson is the great power of general purpose methods, of methods that continue to scale with increased computation even as the available computation becomes very great. The two methods that seem to scale arbitrarily in this way are search and learning. Surprise indicates the mismatch of your mental model and reality, not the inherent weirdness of the latter. Both the surprised/alarmed people and people still in denial about the power of LLMs have to revisit their assumptions and ask if they were founded on any credible understanding to begin with. 1. http://www.vetta.org/2009/12/tick-tock-tick-tock-bing/ http://www.vetta.org/2009/12/tick-tock-tick-tock-bing/ 2. http://www.incompleteideas.net/IncIdeas/BitterLesson.html http://www.incompleteideas.net/IncIdeas/BitterLesson.html
ENGNR 3y ago
There’s an interview between Musk and Tucker Carlson where he says he knew Google had these powerful models internally and thought they should be out in the open, so the answer was a non-profit called Open AI, it was the whole premise of the org I believe he disagreed with the direction the board was taking it and tried to take over, which when failed he stopped funding it, and they went to investors

This is just more openAI marketing at work. ChatGPT is a chatbot, and it’s good at language for sure, but it isn’t some huge breakthrough that people were expecting would take a century. Its literally just a fancy RNN which we’ve had for awhile

famouswaffles 3y ago
No offense but this comment just shows a lack of understanding of the development that's happened with transformers. "Fancy RNN" is a pretty ridiculous assertion. And no, researchers didn't expect what GPT-3/4 has been shown to do to be around the corner at all. GPT's aren't chatbots. That's just a neat natural consequence that's happened. They're machines that reason, understand and follow instructions in plain language. And their abilities go far beyond what any expected language modelling to provide.
- edgyquant 3y ago
  Nah not really, there’s just a lot of hyperbole going around right now. They are RNNs, the difference is they have a layer/layers being fed back instead of a single neuron. That’s a fancy recurrent network. They are not machines that reason, they are approximators. It’s all just token matching based on data we’ve fed it. Further it didn’t happen over night but through successive improvements and at no point was the next improvement considered some infeasible thing. Seems to me maybe you’ve bought into the hype here, and are confusing that with reality.
  - famouswaffles 3y ago
    >They are not machines that reason, they are approximators. And meaningless distinction of the year award goes to.. "It's not real [insert property]" is not an intelligent argument. By all means, divine the way to distinguish results of the two. After all, what kind of important distinction can't be distinguished by results? >Further it didn’t happen over night but through successive improvements The only difference between GPT-3 and GPT-2 was scale. They didn't even change the tokenizer until 4. There were no "successful improvements" to smoothen the massive gap in capabilities between the two. So to say that was expected just shows more lack of knowledge here.
    - airgapstopgap 3y ago
      > The only difference between GPT-3 and GPT-2 was scale. They didn't even change the tokenizer until 4. There were no "successful improvements" to smoothen the massive gap in capabilities between the two. "Other than that, how was the play, Mrs. Lincoln?" GPT-2 is two orders of magnitude smaller. In terms of forebrain neuron count this is the difference between a human and an elephant shrew or budgerigar. It was absolutely expected by reasonable people that 2 OOMs of scaling will provide a qualitative jump.
      - famouswaffles 3y ago
        >GPT-2 is two orders of magnitude smaller. In terms of forebrain neuron count this is the difference between a human and an elephant shrew or budgerigar. It was absolutely expected by reasonable people that 2 OOMs of scaling will provide a qualitative jump. When GPT-3 was released, it was by far the largest artificial neural network ever trained. And I mean by far. Now there wasn't any big jump in hardware capabilities to spur this sort of gulf. It wasn't a case of "Oh now we can train a very large model" So Want to know why there was such a gap ? It's because most researchers assumed the models would overfit the data or display diminishing returns long before 175b. Brain neurons are not comparable to ann parameters.
    - Reki 3y ago
      You would expect a machine that can reason to improve in its responses as a conversation progresses, instead of hallucinating.
  - chpatrick 3y ago
    > They are not machines that reason, they are approximators. They approximate a function that performs reasoning...
- airgapstopgap 3y ago
  Researchers aren't a monolith. Just because Gary Marcus (a complete fraud by the way, look up his "XProp" magitech he sold to Uber) pooh-poohed connectionist AI until ChatGPT came out doesn't mean nobody predicted gains from scaling; certainly many people deep in the know expected things to click. Certainly they expected that the basic trick of associative learning will be cracked. E.g. here's Shane Legg, December 2009 [1]: > Conclusion: computer power is unlikely to be the issue anymore in terms of AGI being possible. The main question is whether we can find the right algorithms. > One of the big things influencing me this year has been learning about how much we understand about how the brain works, in particular, how much we know that should be of interest to AGI designers. I won’t get into it all here, but suffice to say that just a brief outline of all this information would be a 20 page journal paper (there is currently a suggestion that I write such a paper next year with some Gatsby Unit neuroscientists, but for the time being I’ve got too many other things to attend to). At a high level what we are seeing in the brain is a fairly sensible looking AGI design. You’ve got hierarchical temporal abstraction formed for perception and action combined with more precise timing motor control, with an underlying system for reinforcement learning. The reinforcement learning system is essentially a type of temporal difference learning though unfortunately at the moment there is evidence in favour of actor-critic, Q-learning and also Sarsa type mechanisms — this picture should clear up in the next year or so. The system contains a long list of features that you might expect to see in a sophisticated reinforcement learner such as pseudo rewards for informative queues, inverse reward computations, uncertainty and environmental change modelling, dual model based and model free modes of operation, things to monitor context, it even seems to have mechanisms that reward the development of conceptual knowledge. When I ask leading experts in the field whether we will understand reinforcement learning in the human brain within ten years, the answer I get back is “yes, in fact we already have a pretty good idea how it works and our knowledge is developing rapidly.” > I suspect that for the next 5 years, and probably longer, neuroscientists working on understanding cortex aren’t going to be of much use to AGI efforts. My guess is that sometime in the next 10 years developments in deep belief networks, temporal graphical models, liquid computation models, slow feature analysis etc. will produce sufficiently powerful hierarchical temporal generative models to essentially fill the role of cortex within an AGI. > Right, so my prediction for the last 10 years has been for roughly human level AGI in the year 2025 (though I also predict that sceptics will deny that it’s happened when it does!) This year I’ve tried to come up with something a bit more precise. In doing so what I’ve found is that while my mode is about 2025, my expected value is actually a bit higher at 2028. This is not because I’ve become more pessimistic during the year, rather it’s because this time I’ve tried to quantify my beliefs more systematically and found that the probability I assign between 2030 and 2040 drags the expectation up. Perhaps more useful is my 90% credibility region, which from my current belief distribution comes out at 2018 to 2036. And here's Rich Sutton's famous Bitter Lesson, a month after GPT-2 [2]: > We have to learn the bitter lesson that building in how we think we think does not work in the long run. The bitter lesson is based on the historical observations that 1) AI researchers have often tried to build knowledge into their agents, 2) this always helps in the short term, and is personally satisfying to the researcher, but 3) in the long run it plateaus and even inhibits further progress, and 4) breakthrough progress eventually arrives by an opposing approach based on scaling computation by search and learning. The eventual success is tinged with bitterness, and often incompletely digested, because it is success over a favored, human-centric approach. > One thing that should be learned from the bitter lesson is the great power of general purpose methods, of methods that continue to scale with increased computation even as the available computation becomes very great. The two methods that seem to scale arbitrarily in this way are search and learning. Surprise indicates the mismatch of your mental model and reality, not the inherent weirdness of the latter. Both the surprised/alarmed people and people still in denial about the power of LLMs have to revisit their assumptions and ask if they were founded on any credible understanding to begin with. 1. http://www.vetta.org/2009/12/tick-tock-tick-tock-bing/ http://www.vetta.org/2009/12/tick-tock-tick-tock-bing/ 2. http://www.incompleteideas.net/IncIdeas/BitterLesson.html http://www.incompleteideas.net/IncIdeas/BitterLesson.html

ENGNR 3y ago

There’s an interview between Musk and Tucker Carlson where he says he knew Google had these powerful models internally and thought they should be out in the open, so the answer was a non-profit called Open AI, it was the whole premise of the org I believe he disagreed with the direction the board was taking it and tried to take over, which when failed he stopped funding it, and they went to investors