I think we're at a point where AI may or may not make huge changes to how we live and think. I'm a bit skeptical on if it will be an emergent technology that creates new job sectors like the...
I think we're at a point where AI may or may not make huge changes to how we live and think. I'm a bit skeptical on if it will be an emergent technology that creates new job sectors like the Personal Computer did for the past 30 years, I think what might happen is that it will create new avenues of employment but it won't lead to a serious shift like some (or many) assume.
People like to get very 'gloom and doom' over this but I don't know if we will change as much. There is some recent controversy over Google AI and the failures of AI, and people oftentimes confuse traditional robotics with sensors as AI. I feel like people don't really know what is AI and what is not AI, and investors will flock to your product if it mentions AI which implies some form of blind faith. This could be a new computing paradigm much like GUI-based operating systems and the Internet, or it could be a bust like Crypto.
The hard thing about AI is they keep moving the goal posts to define what is and isn't AI. The marketing on this and selling hype is so high it won't move up to it. Likely what we'll see is...
The hard thing about AI is they keep moving the goal posts to define what is and isn't AI. The marketing on this and selling hype is so high it won't move up to it.
Likely what we'll see is further replacement of customer service to AI where you'll never get a human on the phone which is already getting harder and harder.
In my sphere I see companies advertising their forecasting models they've had for 10-20 years as AI. Many haven't changed anything, they may use regression but most just use historical days and apply a multiplier to adjust and ensure it falls on a Monday if it was on a Monday last year. Not sophisticated and definitely not AI.
Most companies data is so bad to find a use case will be extremely difficult. Likely there will be some niches and as we're seeing now AI generated articles for more and more pump and dump and adjusted marketing.
I'd love to see a strong use case outside of stuff we've already been augmenting that provides strong value beyond the incremental changes we've seen over the last 10 years.
I'm very skeptical. Perhaps a big one will be medical and molecule research? I'm having a hard time seeing where this replaces people en masse vs just making people more productive.
I think the issue has been that generative AI is really good at bluffing. So while it's genuinely useful at a decent number of tasks, it's got lots of people vastly overestimating what it's...
I think the issue has been that generative AI is really good at bluffing. So while it's genuinely useful at a decent number of tasks, it's got lots of people vastly overestimating what it's capable of. This means the hype is going to peak very high and make even a very good level of adoption look tiny by comparison. Meanwhile the public backlash caused by hype-based misuse is going to have it's own consequences.
If AGI is possible and is developed, I'm confident that we will hit a point where our economic systems will be forced, maybe violently to fundementally shift in massive ways. I'm also very...
If AGI is possible and is developed, I'm confident that we will hit a point where our economic systems will be forced, maybe violently to fundementally shift in massive ways.
I'm also very confident that the current ai "revolution" is not that point.
AI research follows a very familiar pattern. Incremental gains are positive enough that useful products are released with the state of the art of the time, a ridiculously massive amount of hype is generated alongside an eye watering amount of funding for AI projects, and soon enough people realize the limitations of those projects, excitement mostly dries up, AI is integrated into normally every day products that people stop being excited about, and it no longer considered AI until the next big product release. There were a few AI hype cycles and winters that I've lived through, most notably Watson and voice assistants in the late 00s. Self driving cars in the mid 10s, and there were many more that were before my time.
I don't see the current LLM craze as much different. They can be useful tools, but they're not going to upend our current way of life.
Yeah, I mean with any new useful tool there's going to be disruption, but we're probably talking about microwave oven levels of disruption here, not steam engine levels.
Yeah, I mean with any new useful tool there's going to be disruption, but we're probably talking about microwave oven levels of disruption here, not steam engine levels.
I think we're at a point where AI may or may not make huge changes to how we live and think. I'm a bit skeptical on if it will be an emergent technology that creates new job sectors like the Personal Computer did for the past 30 years, I think what might happen is that it will create new avenues of employment but it won't lead to a serious shift like some (or many) assume.
People like to get very 'gloom and doom' over this but I don't know if we will change as much. There is some recent controversy over Google AI and the failures of AI, and people oftentimes confuse traditional robotics with sensors as AI. I feel like people don't really know what is AI and what is not AI, and investors will flock to your product if it mentions AI which implies some form of blind faith. This could be a new computing paradigm much like GUI-based operating systems and the Internet, or it could be a bust like Crypto.
The hard thing about AI is they keep moving the goal posts to define what is and isn't AI. The marketing on this and selling hype is so high it won't move up to it.
Likely what we'll see is further replacement of customer service to AI where you'll never get a human on the phone which is already getting harder and harder.
In my sphere I see companies advertising their forecasting models they've had for 10-20 years as AI. Many haven't changed anything, they may use regression but most just use historical days and apply a multiplier to adjust and ensure it falls on a Monday if it was on a Monday last year. Not sophisticated and definitely not AI.
Most companies data is so bad to find a use case will be extremely difficult. Likely there will be some niches and as we're seeing now AI generated articles for more and more pump and dump and adjusted marketing.
I'd love to see a strong use case outside of stuff we've already been augmenting that provides strong value beyond the incremental changes we've seen over the last 10 years.
I'm very skeptical. Perhaps a big one will be medical and molecule research? I'm having a hard time seeing where this replaces people en masse vs just making people more productive.
I think the issue has been that generative AI is really good at bluffing. So while it's genuinely useful at a decent number of tasks, it's got lots of people vastly overestimating what it's capable of. This means the hype is going to peak very high and make even a very good level of adoption look tiny by comparison. Meanwhile the public backlash caused by hype-based misuse is going to have it's own consequences.
If AGI is possible and is developed, I'm confident that we will hit a point where our economic systems will be forced, maybe violently to fundementally shift in massive ways.
I'm also very confident that the current ai "revolution" is not that point.
AI research follows a very familiar pattern. Incremental gains are positive enough that useful products are released with the state of the art of the time, a ridiculously massive amount of hype is generated alongside an eye watering amount of funding for AI projects, and soon enough people realize the limitations of those projects, excitement mostly dries up, AI is integrated into normally every day products that people stop being excited about, and it no longer considered AI until the next big product release. There were a few AI hype cycles and winters that I've lived through, most notably Watson and voice assistants in the late 00s. Self driving cars in the mid 10s, and there were many more that were before my time.
I don't see the current LLM craze as much different. They can be useful tools, but they're not going to upend our current way of life.
They are already disrupting professional writers in particular, in marketing etc. but I agree with you
Yeah, I mean with any new useful tool there's going to be disruption, but we're probably talking about microwave oven levels of disruption here, not steam engine levels.
Mirror, for those hit by the paywall:
https://archive.is/KCQD4