4 ms·

Rather than seeing the rise of AI as a threat to my career, I now view it as an opportunity to augment my skills and deliver even greater value to my clients. B

Rather than seeing the rise of AI as a threat to my career, I now view it as an opportunity to augment my skills and deliver even greater value to my clients. By embracing AI tools like ChatGPT, I can automate routine tasks and focus my efforts on the areas where my expertise and creativity can truly shine. It is the only option available in the immediate term and I would argue is the correct short term perspective. However, if AI continues on its path of exponential technological acceleration, then this is going to be very problematic. You can't sprint ahead of an accelerating technological curve forever. We are replacing the old monotonous task rat race with the new rat race attempting to remain relevant against inescapable advancements. I've described this in more detail writing as "Climbing the skill ladder is going to look more like running on a treadmill at the gym. No matter how fast you run, you aren’t moving, AI is still right behind you learning everything that you can do." https://dakara.substack.com/p/ai-and-the-end-to-all-things https://dakara.substack.com/p/ai-and-the-end-to-all-things

orwin 3y ago
I'm not so sure about it. Neural networks and their models are built on theories made 50 years ago. The breakthrough happened in 2011 and we were able to build a lot on it, but we are close to our limits unless a company made another discovery and didn't publish it. I'm sure we have room for improvements but the acceleration stopped at GPT3.5, and imho the last version proved that the next increment will be specific but not as impressive as GPT3 was. I could be wrong, but imho our engineering caught up to our theory, or is close to, and while we'll be less stale than the aerospace sector, as our experiments are less materially expensive, I can see the field diversify but stop growing as fast.
- 13years 3y ago
  I don't disagree. However, those who believe we are still on the accelerating curve also believe that somehow we will remain relevant. It doesn't make logical sense. So, the only counter to my argument is that we hit a wall. This is very possible, but AI development thus far is certainly increasing the unpredictability of future events. As we move forward we are in some sense understanding it less. The disparity over the definition of AGI is only increasing.
  - circuit10 3y ago
    There are definitely a lot of people who believe that we are still accelerating and as a result believe that we will either become irrelevant (but hopefully live in some kind of utopia) or all die (I guess I’m one of them, but I haven’t really accepted it on a deep level)
    - 13years 3y ago
      On the grand scale we are definitely on an acceleration technological curve. And by that I mean humanity and all technology not constrained to only AI. AI might hit some walls, there may be a lot of hype at the moment, but it has already broken new ground and boundaries not previously thought possible. Unfortunately, many of those grounds already broken and proven are nefarious and dystopian uses. In the article I link above, I go into great detail on this topic. With all the concern of AGI (utopia/world ending) there is not much focus on the imminent impacts of primitive AI. They are no less concerning.
- circuit10 3y ago
  Neural networks are Turing-complete, so in theory a large enough one should be able to express any logic. So the main limitation here is the training process, but the way it’s done seems very flexible, so I think the main limitation at the moment is just model size, which is limited by hardware which is still improving exponentially (though it has been slowing down slightly)

orwin 3y ago

I'm not so sure about it. Neural networks and their models are built on theories made 50 years ago. The breakthrough happened in 2011 and we were able to build a lot on it, but we are close to our limits unless a company made another discovery and didn't publish it. I'm sure we have room for improvements but the acceleration stopped at GPT3.5, and imho the last version proved that the next increment will be specific but not as impressive as GPT3 was. I could be wrong, but imho our engineering caught up to our theory, or is close to, and while we'll be less stale than the aerospace sector, as our experiments are less materially expensive, I can see the field diversify but stop growing as fast.

13years 3y ago
I don't disagree. However, those who believe we are still on the accelerating curve also believe that somehow we will remain relevant. It doesn't make logical sense. So, the only counter to my argument is that we hit a wall. This is very possible, but AI development thus far is certainly increasing the unpredictability of future events. As we move forward we are in some sense understanding it less. The disparity over the definition of AGI is only increasing.
- circuit10 3y ago
  There are definitely a lot of people who believe that we are still accelerating and as a result believe that we will either become irrelevant (but hopefully live in some kind of utopia) or all die (I guess I’m one of them, but I haven’t really accepted it on a deep level)
  - 13years 3y ago
    On the grand scale we are definitely on an acceleration technological curve. And by that I mean humanity and all technology not constrained to only AI. AI might hit some walls, there may be a lot of hype at the moment, but it has already broken new ground and boundaries not previously thought possible. Unfortunately, many of those grounds already broken and proven are nefarious and dystopian uses. In the article I link above, I go into great detail on this topic. With all the concern of AGI (utopia/world ending) there is not much focus on the imminent impacts of primitive AI. They are no less concerning.
circuit10 3y ago
Neural networks are Turing-complete, so in theory a large enough one should be able to express any logic. So the main limitation here is the training process, but the way it’s done seems very flexible, so I think the main limitation at the moment is just model size, which is limited by hardware which is still improving exponentially (though it has been slowing down slightly)