Do you think we are mostly going to be replaced by AI? If so what timeline (5yrs, 10yrs, 20yrs, 30yrs, >30yrs)?
Basically the title, leaving the question as open ended as possible. It's now been 1yr since ChatGPT debuted so we have a little more perspective from the panic knee-jerk reactions and a better sense of what the AI can do today / what rate of progress has looked like. Open to any and all thoughts on the subject (what jobs get killed off first, what sticks around maybe forever, how finance as an industry fares, etc)
Where is that man, who was here a few times earlier this year, going all schiziod about AI lol. We need him to chime in.
No.
Fair enough - Why not?
I had a response on another thread about how if we showed up to ancient Mesopotamia with a combine harvester and a factory, they would've said the same thing we're saying today about technology replacing us all. I'm sure finance people said the same thing when computers and Microsoft Excel were introduced so companies didn't need floors of number crunchers anymore. AI, like any other technology, is just a tool that will help us get more things done. Obviously some jobs will go away, but we're not "mostly" going to be replaced by AI.
Real convincing, great argument bro!
ChatGPT's gonna run a HF on it's own and LPs run by ChatGPT are gonna allocate to it
Yes, I think AI will outperform most humans at most human tasks in 10yrs or so
F*ck it I'll bite. Here's what I think a progression of AI over the time horizon might look like. To lay the ground work, I don't think we're gonna see Skynet/Terminator type of pure rogue autonomous AI anytime soon. Not with the current processing chip technology. We basically have reached limitation on how small we can get the transistors with existing tech. I believe 3nm is basically the smallest silicon transistor we can "print" without running into an actual quantum physics issue with electrons hopping over. This is all of course written with the caveat this is just pure speculation after a certain point. Current AI models, to paint with a broad brush, is basically a regression model on steroids. Also this is just one person's opinion / creative imagination. Also also, I'm not an expert in any of this. Just an enthusiast who worked in tech companies and have some tangential understanding from it. So grain of salt and all that.
5 Years: Majority of basically routine task are automated. Think basic P&L statements, low level of financial forecast/analysis can be done with AI. There already software exist that can do these things. Roles like AR/AP, invoicing etc are automated are automated and mostly run by AI; maybe with a final human check/verification. Use cases will be limited to specific jobs & functions that the AI is designed to do. Broader everyday common civilian use will be limited or maybe just getting to a point of commercialization that it's approaching affordability. (Cloud processing ain't cheap, AWS isn't in the business of losing monies)
10 Years: Wide commercial adoption for everyday use. Processing is cheap and easy and affordable to most well off families. Most business will employ some form of AI to aid in running the business. Will still have certain limitations in what it can do. Use cases for AI will have expanded beyond basics. I think school will be teaching AI coding in classes instead of Java coding classes.
15 Years: now we're getting into conjuncture territory here. Maybe quantum computing might be more mature and available for purchase by research institutions and gov entities/companies and they starts to incorporate into their workflow. AI processing speeds will have an evolutionary jump in capabilities. Probably have base capabilities to manage aspect of our lives beyond business applications. Thinking along the lines of manufacturing design be aided with AI powered by quantum computing, or electric power generation with a one-to-one match to demand. Manufacturing will be largely automated with very limited human intervention. Providing forecast with increasing granularity such as predicting weather pattern with much much higher certainty than now.
20 Years: hopefully if we design the AI functionalities correctly, it can automate much of our human needs such as food production, how much housing we need. AI effectively optimizes most aspect of our lives. (yes, that's some optimistic view).
30 Years: throw up a pie in the air and see what lands and what doesn't. If doesn't land, start praying cuz an Arnold Schwarzenegger Terminator might actually be around the corner. Bit of a cheeky joke, but tech progresses so fast that it's difficult to even think about the realm of the possible.
Overall, I think everything more or less hinges on quantum computing or an alternative to increase processing speed. The faster our computers can run the better AI can process. How an AI processes and the parameters that are set determines the wider applicability and usefulness. And I pray the military industrial complex does not get their dirty hands on it. Be interested to see other's thoughts.
Great thoughts -- though the one element you didn't account for in those various scenarios (5yrs, 10yrs, 15yrs, etc) is the impact on jobs at each stage
I'm skeptical of the 'new job creation' argument. yes there will be new jobs, but will the quantum of new jobs be sufficient to replace jobs taken away? I'm very skeptical
I think humans will still have an edge in 5yrs, though in 10yrs the base case in my mind is at least 70% of white collar tasks can be done by AI. You just don't need that many people relative to what we have today in this scenario
Fair call out. I guess I started thinking about it and the more I thought it just pure making up sh*t. What I do think is that there will be a job transition to jobs that is more oriented AI and robotics. Swarm drone and more diverse purpose for drones. Maybe farming uses, maybe construction, hardware oriented. Jobs will naturally shift away to what the market demands. All the junior AP/AR analyst won't disappear overnight. There will be a transition. I mean ice deliveryman used to be a profession. I'm sure the question of "what will ice deliveryman do after everyone gets a refrigerator?" was probably asked.
In terms of quantum, I'm concerned as well. But I think new opportunities will open up that people can shift into without some type of catastrophic mismatch between available jobs & qualified people.
I'm all for the adoption of AI over time and improving our society. I too get worried about how quickly it happens and whether that means major disruption in job markets that we don't have enough time to adjust to. That said, I do think adoption will take considerably longer than we think. I think we look at things like Tik Tok or other apps and how quickly they are adopted and sort of extrapolate that to something like AI. In reality, AI applications that are going to be majorly disruptive to job markets are going to take time. Not only will it take time to even build out the necessary infrastructure (power grids, chips, etc.) for everyone to be running high power AI programs, but these things aren't going to be free (they will actually be quite expensive as whoever develops these disruptive applications first are going to want to be paid big time for them). They are going to require major investments which is another layer of complexity for companies and businesses to overcome in adopting it, especially in earlier stages. I think for those reasons we hopefully see the changes happen over 5 - 10 year periods instead of 1 - 2 year periods which allow us to have more time do adapt as a society.
Rerum ad minima omnis sint. Qui adipisci qui perferendis dolor aut quis. Delectus laborum veniam est et quia possimus. Porro libero odio animi.
Fugiat reiciendis porro sed quidem ipsam et ut. Quis corporis deleniti ad iure rerum quasi non. Eveniet officiis dolorem voluptatibus quaerat.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...