I've got 2 decades of experience in media, advertising & the creative industry as a whole. And I don't mean bush league wedding video's, B2B flyers or 'content': I work for major brands & institutions on a senior level. If you do not see the danger or potential in this, then you are not expert enough or too naïve. I have used this, in practice, already for clients. It has saved me countless hours in 1 week. That is @ $100 / hour. All people I know how 'doth protest too much' are under the illusion that creativity is some magical skill. It isn't.
This is a boon for people without skill or budget, a force multiplier for unicorns. Anyone in between needs to adjust or become jobless.
I can create in one day what used to take weeks. For a lot of use case, these tools create cheap starting points without needing to look for funding.
That is massive.
This allows me to pursue business idea's I didn't have the time or energy for. This will massively change prototyping in a year.
You’ll have to have a lot of clients to bill to make a living, especially since your competitors or even your clients will do this for themselves. I fail to see how this benefits you long term.
I'm not at a senior level as you are, but I can almost see what you are saying. I don't want to be left behind. I am a student of marketing and I'm deeply passionate about internet marketing with the use of, - Artificial intelligence.
Will you please teach me how to apply artificial intelligence in marketing?
Well, there go the jobs for all the 'kids' who became digital marketers as it was The Future. And all those 1-2 horse consultancies, primarily women at home, doing work for smaller businesses, unless those small businesses turn to them. It's likely just enough for small businesses because what I see here meets needs, but hasn't any original thought. I'm sure it can do a value prop (promise to deliver) too.
When prompted appropriately, the good models can produce amazing results. Especially if you start feeding your own corpus of information using vector search / langchain for better retrieval and output.
1. This is going to be great for entrepreneurs. Being an entrepreneur is difficult in part because of the need to "wear different hats", which denies aspiring entrepreneur´s the benefits of specialization. Now, that is a hurdle that will be much easier to pass.
2. This is going to make many business services jobs more like professional athletes: Some workers are going to get all of the money and all of the works, working for the big companies, and the rest are going to be struggling and failing.
From your example, I can see this with:
- Marketing.
- Front-end web development.
- Community managers.
Jobs that right now are competitive, but in the future will be hypercompetitive.
Personally, as a marketer, I'm grateful that I'm at end of the work career. A lot of marketing and what used to be called advertising has become like laying bricks or making a shed. It hasn't been fun or creative in a long time, and humans need creativity. Big, fresh ideas? The jaw droppers? Absent. Let's face it, we have a population that is happy with dreck. Where is the shock of the new?
As an entrepreneur, I've definitely been investing how this can help me. I've never paid for marketing help, so it's not putting anyone out of a job. I've used it to help me with web development, and have gotten pretty good results so far. I'm hoping to spend some time coming up with a marketing strategy that AI can help me be consistent with. That's been my biggest hurtle as a small business owner.
Why would you go to low paid authors?....the spirit of the experiment is what is coming our way, not how perfect it is in its baby stage. Jump on it, twist it, make it develop its potential....being negative is easy. Give it something....
I don’t see how this doesn’t lead to large job losses, and in demographics that are totally unused to losses of that scale. It will no longer be blue collar people in flyover states affected, but college educated white collar workers who think they made all the right decisions in life. This will have huge downstream effects on politics and a lot besides.
When there have been large productivity gains in the somewhat recent past, they have never led to declining work hours and increasing leisure hours. They have always led to fewer workers working similar hours.
Maybe my imagination is not strong enough, but this technology seems perfect for replacing workers rather than creating new jobs.
On a side note, I think people are putting too much hope in learning to prompt as some sort of secret skill that can lead to making a lot of money or building a career. Writing good prompts is *much easier* than learning to code proficiently, learning the myriad little skills that go into most white collar jobs, etc. And the fact that it is much easier will result in it being not very valuable a skill at all. Unskilled or semiskilled work in other words.
A lot of younger white collar workers will be out of jobs or working for minimum wage. If I were a mid-stage marketer, I'd be looking to train in a new career like plumbing or painting because Bing or ChatGPT is one step away from writing value props and plans. I thank God I'm at the end, and saw a lot of real creativity which is like great nutrition. The dive into mediocrity continues.
Writing largely may as well be done by ChatGPT as it is already poor. There's a glaring typo in the first paragraph of this though overall it's written well. I read a lot online and the basic grammar and conveyance of ideas is poor because it's is written too quickly.
AI still has a lot of quirks. It does save time, but it's also wrong a lot, so you can't just take what it gives you and use it. You have to make sure it's accurate. Like with coding, I had an entire conversation with Google Bard the other day during which it lied to me the entire time about the code I presented it. I think it even lied to me about being able to access the Google Drive link I gave it. It was funny but was a great example of the fact that AI is still in development and may never reach its potential.
"When we all can do superhuman amounts of work, what happens?"
you should see that you've already answered your question a sentence or two earlier:
"I am sure humans could have done better, but they could not have been as fast."
I've been reading your blog for a couple months now (which is, like, ten thousand years in AI time?), and I really appreciate that you by and large take an agnostic stance towards the tech. It's scary, but simultaneously AMAZING! The problem for me here is that you (and most other nominally critical observers) still conflate terms like "better" and "superhuman" with terms clearly linked to efficiency and productivity (like "faster" and "amounts of work").
You're continuing to promote a worldview (a very fucked-up one, from my perspective) that thinks of efficiency, quantification of human labor, maximization of profit, and similar bullshit with superlatives like "better" and "superhuman." If what you're after is just getting shit done faster and "better" (where "better" means "more efficiently"), then I hope you've purchased your ticket on the Singularity Express, because it's leaving the station soon.
I hear Friedman has a few tickets he's hawking outside the Skynet tent, but the price is dear. Still, I understand you'll get your money back--guaranteed--at this train's terminus.
(Dude! I mean, SERIOUSLY! Look at how humbly uncritical (and creepily reverent) most of your readers are in their responses! When I said that you take an agnostic stance, I hope you can see that virtually everyone here not only believes in the saving grace of AI: they have drunk all the Kool-Aid they can, and will blithely traipse on into that dystopia (and/or extinction level event), with pockets stuffed full of cash. Your glass of Kool-Aid seems to have a few drops left in it, and your pockets to be less lined with money--you keep seeding your posts with caveats and messages to take care, but check yoself before you wreck yoself. And the rest of humanity at the same time.)
Over the course of the blog, I have been alternating between excitement and worry, and I try to communicate both. But, look, I also am trying to view AI pragmatically from within the window of my specialties: education and management, and especially how they overlap in entrepreneurship. And much of my research focuses on the fact that talent is evenly distributed, opportunity is not. From that perspective, the ability of AI to allow people to achieve things they could not before is the most exciting breakthrough in a century. The fact that the cost to use AI is so low is the ticket to a lot of people's (and society's) benefit - it democratizes opportunity. I see it in my classes: people who weren't taken seriously because they were bad writers or English was their third language now are. People who had a knowledge or skills gap can fill it. And yes, that is about productivity and speed as well.
We live in a world of AI, now. If I really believed that the Singularity and dystopia were coming, I would have a different stance. If I thought that the likely path was doom and mass unemployment, I would say that. How could I not? I honestly don't know what is coming, so I consider myself more of a pragmatist than an agnostic. I want people to understand how a key element of their lives -- work -- is changing, and that involves both meeting them in the world they are in (the world of work) and trying to show them possibilities, good and bad. I think seeing that is the first step to inoculation against the Kool-Aid. I could be wrong (I might be wrong!) but I really am writing what I believe is important to understand, and what I think I can teach. Should I feel that doom approaches, I will certainly yell it from the rooftops (like I have been doing about deepfakes). I am figuring this out too.
.....Hopefully, we will be in a future where we do less boring work, offloading the annoying and unfulfilling tasks, so that we can focus on the more creative and generative work we like to do.....
The above statement is great if you as an individual have a creative flair, but let's not forget the millions of people around the world for whom this is not in their DNA, and excel best in doing standard functional tasks well.
We are moving to a world where the focus is moving from producing to curating
ChatGPT is great at curating. In fact that is one of its main roles. Instead of going through what will in 5 years seem like the incredibly primitive method of a google search, you can ask anything from Bing and get a response complete with citations, and then ask follow up questions in casual, natural language which it will fully understand.
That’s not to say it’s only about curation. At this point lots of uses are baked in the cake even with today’s technology and APIs. A disembodied yet almost human personal assistant following us around: “GPT, put Bluey on our living room TV, get me an Uber to Akiko’s apartment on 68th street in 5 minutes”, a “good enough” math tutor, a “good enough” mental health therapist ChatGPT. We are about to see 50,000 startups bloom, get replaced by the next generation that might include an expanded short-term memory, etc
Curating dreck. And meanwhile the human brain rots and we become nothing but 'consumers.' You see that in everyone from highly educated people to working class people---people are forgetting to think, exercise their brains, and aspire to something beyond themselves, even if it's faith. It's very handy for a certain group who want to control the rest.
If I were a potential customer for this game I would skip over all this and probably not look at the product itself because it is so painfully generic. I get excited by communications and products that show some creativity, style and quality. None of this qualifies. The internet is drowning in stuff like this already. Better to put some investment in standing out with a real human touch.
Tried to use GPT4 but still it tells me that it s based on GPT3 - even after joining the waitlist...anything I may have done wrong :-/? Any recommendations plss??
Thanks for sharing Ethan. This paints a vivid picture of just how disruptive AI tools will be (and, in fact, already are). It’s nearly impossible to keep up with how quickly this space is moving. You could probably repeat this experiment on a near-weekly basis, incorporating the latest advances in AI tools each time. What would this experiment look like 3 months from now? In 6 months. How will your inputs/prompts and the results differ?
Thank you so much Ethan, this article is very helpful and interesting, actually we are already in a world of superhumans, we just have to wait for the implications.
Great experiment, but it appears that all the content derives from web-based material you provided at the outset. Bing just repackaged your content into the forms you requested. That's impressive, but rests heavily on the effort of the human(s) who created the original copy.
I've got 2 decades of experience in media, advertising & the creative industry as a whole. And I don't mean bush league wedding video's, B2B flyers or 'content': I work for major brands & institutions on a senior level. If you do not see the danger or potential in this, then you are not expert enough or too naïve. I have used this, in practice, already for clients. It has saved me countless hours in 1 week. That is @ $100 / hour. All people I know how 'doth protest too much' are under the illusion that creativity is some magical skill. It isn't.
This is a boon for people without skill or budget, a force multiplier for unicorns. Anyone in between needs to adjust or become jobless.
I can create in one day what used to take weeks. For a lot of use case, these tools create cheap starting points without needing to look for funding.
That is massive.
This allows me to pursue business idea's I didn't have the time or energy for. This will massively change prototyping in a year.
This is the steam engine for pro magnon.
Remember this comment.
You’ll have to have a lot of clients to bill to make a living, especially since your competitors or even your clients will do this for themselves. I fail to see how this benefits you long term.
I'm not at a senior level as you are, but I can almost see what you are saying. I don't want to be left behind. I am a student of marketing and I'm deeply passionate about internet marketing with the use of, - Artificial intelligence.
Will you please teach me how to apply artificial intelligence in marketing?
Thank you.
David.
Well, there go the jobs for all the 'kids' who became digital marketers as it was The Future. And all those 1-2 horse consultancies, primarily women at home, doing work for smaller businesses, unless those small businesses turn to them. It's likely just enough for small businesses because what I see here meets needs, but hasn't any original thought. I'm sure it can do a value prop (promise to deliver) too.
When prompted appropriately, the good models can produce amazing results. Especially if you start feeding your own corpus of information using vector search / langchain for better retrieval and output.
Thank you for this experiment.
Some predictions:
1. This is going to be great for entrepreneurs. Being an entrepreneur is difficult in part because of the need to "wear different hats", which denies aspiring entrepreneur´s the benefits of specialization. Now, that is a hurdle that will be much easier to pass.
2. This is going to make many business services jobs more like professional athletes: Some workers are going to get all of the money and all of the works, working for the big companies, and the rest are going to be struggling and failing.
From your example, I can see this with:
- Marketing.
- Front-end web development.
- Community managers.
Jobs that right now are competitive, but in the future will be hypercompetitive.
Personally, as a marketer, I'm grateful that I'm at end of the work career. A lot of marketing and what used to be called advertising has become like laying bricks or making a shed. It hasn't been fun or creative in a long time, and humans need creativity. Big, fresh ideas? The jaw droppers? Absent. Let's face it, we have a population that is happy with dreck. Where is the shock of the new?
As an entrepreneur, I've definitely been investing how this can help me. I've never paid for marketing help, so it's not putting anyone out of a job. I've used it to help me with web development, and have gotten pretty good results so far. I'm hoping to spend some time coming up with a marketing strategy that AI can help me be consistent with. That's been my biggest hurtle as a small business owner.
This is excellent work. Thank you for sharing.
This is an interesting experiment but I'm not sure of the effectiveness of what was created. How much editing does it need?
It make me think of low paid authors. They can output lots of words for a few dollars. But then you have to spend hours editing.
Why would you go to low paid authors?....the spirit of the experiment is what is coming our way, not how perfect it is in its baby stage. Jump on it, twist it, make it develop its potential....being negative is easy. Give it something....
I'm experimenting with ChatGPT almost every day. I mentioned the low paid authors as an analogy.
The point of the article is that AI is fast. I'm not sure about the time savings in these examples.
I don’t see how this doesn’t lead to large job losses, and in demographics that are totally unused to losses of that scale. It will no longer be blue collar people in flyover states affected, but college educated white collar workers who think they made all the right decisions in life. This will have huge downstream effects on politics and a lot besides.
When there have been large productivity gains in the somewhat recent past, they have never led to declining work hours and increasing leisure hours. They have always led to fewer workers working similar hours.
Maybe my imagination is not strong enough, but this technology seems perfect for replacing workers rather than creating new jobs.
On a side note, I think people are putting too much hope in learning to prompt as some sort of secret skill that can lead to making a lot of money or building a career. Writing good prompts is *much easier* than learning to code proficiently, learning the myriad little skills that go into most white collar jobs, etc. And the fact that it is much easier will result in it being not very valuable a skill at all. Unskilled or semiskilled work in other words.
A lot of younger white collar workers will be out of jobs or working for minimum wage. If I were a mid-stage marketer, I'd be looking to train in a new career like plumbing or painting because Bing or ChatGPT is one step away from writing value props and plans. I thank God I'm at the end, and saw a lot of real creativity which is like great nutrition. The dive into mediocrity continues.
Writing largely may as well be done by ChatGPT as it is already poor. There's a glaring typo in the first paragraph of this though overall it's written well. I read a lot online and the basic grammar and conveyance of ideas is poor because it's is written too quickly.
AI still has a lot of quirks. It does save time, but it's also wrong a lot, so you can't just take what it gives you and use it. You have to make sure it's accurate. Like with coding, I had an entire conversation with Google Bard the other day during which it lied to me the entire time about the code I presented it. I think it even lied to me about being able to access the Google Drive link I gave it. It was funny but was a great example of the fact that AI is still in development and may never reach its potential.
Especially when we get to a point where you can just prompt GPT for better prompts.
When you ask
"When we all can do superhuman amounts of work, what happens?"
you should see that you've already answered your question a sentence or two earlier:
"I am sure humans could have done better, but they could not have been as fast."
I've been reading your blog for a couple months now (which is, like, ten thousand years in AI time?), and I really appreciate that you by and large take an agnostic stance towards the tech. It's scary, but simultaneously AMAZING! The problem for me here is that you (and most other nominally critical observers) still conflate terms like "better" and "superhuman" with terms clearly linked to efficiency and productivity (like "faster" and "amounts of work").
You're continuing to promote a worldview (a very fucked-up one, from my perspective) that thinks of efficiency, quantification of human labor, maximization of profit, and similar bullshit with superlatives like "better" and "superhuman." If what you're after is just getting shit done faster and "better" (where "better" means "more efficiently"), then I hope you've purchased your ticket on the Singularity Express, because it's leaving the station soon.
I hear Friedman has a few tickets he's hawking outside the Skynet tent, but the price is dear. Still, I understand you'll get your money back--guaranteed--at this train's terminus.
(Dude! I mean, SERIOUSLY! Look at how humbly uncritical (and creepily reverent) most of your readers are in their responses! When I said that you take an agnostic stance, I hope you can see that virtually everyone here not only believes in the saving grace of AI: they have drunk all the Kool-Aid they can, and will blithely traipse on into that dystopia (and/or extinction level event), with pockets stuffed full of cash. Your glass of Kool-Aid seems to have a few drops left in it, and your pockets to be less lined with money--you keep seeding your posts with caveats and messages to take care, but check yoself before you wreck yoself. And the rest of humanity at the same time.)
Over the course of the blog, I have been alternating between excitement and worry, and I try to communicate both. But, look, I also am trying to view AI pragmatically from within the window of my specialties: education and management, and especially how they overlap in entrepreneurship. And much of my research focuses on the fact that talent is evenly distributed, opportunity is not. From that perspective, the ability of AI to allow people to achieve things they could not before is the most exciting breakthrough in a century. The fact that the cost to use AI is so low is the ticket to a lot of people's (and society's) benefit - it democratizes opportunity. I see it in my classes: people who weren't taken seriously because they were bad writers or English was their third language now are. People who had a knowledge or skills gap can fill it. And yes, that is about productivity and speed as well.
We live in a world of AI, now. If I really believed that the Singularity and dystopia were coming, I would have a different stance. If I thought that the likely path was doom and mass unemployment, I would say that. How could I not? I honestly don't know what is coming, so I consider myself more of a pragmatist than an agnostic. I want people to understand how a key element of their lives -- work -- is changing, and that involves both meeting them in the world they are in (the world of work) and trying to show them possibilities, good and bad. I think seeing that is the first step to inoculation against the Kool-Aid. I could be wrong (I might be wrong!) but I really am writing what I believe is important to understand, and what I think I can teach. Should I feel that doom approaches, I will certainly yell it from the rooftops (like I have been doing about deepfakes). I am figuring this out too.
.....Hopefully, we will be in a future where we do less boring work, offloading the annoying and unfulfilling tasks, so that we can focus on the more creative and generative work we like to do.....
The above statement is great if you as an individual have a creative flair, but let's not forget the millions of people around the world for whom this is not in their DNA, and excel best in doing standard functional tasks well.
We are moving to a world where the focus is moving from producing to curating
ChatGPT is great at curating. In fact that is one of its main roles. Instead of going through what will in 5 years seem like the incredibly primitive method of a google search, you can ask anything from Bing and get a response complete with citations, and then ask follow up questions in casual, natural language which it will fully understand.
That’s not to say it’s only about curation. At this point lots of uses are baked in the cake even with today’s technology and APIs. A disembodied yet almost human personal assistant following us around: “GPT, put Bluey on our living room TV, get me an Uber to Akiko’s apartment on 68th street in 5 minutes”, a “good enough” math tutor, a “good enough” mental health therapist ChatGPT. We are about to see 50,000 startups bloom, get replaced by the next generation that might include an expanded short-term memory, etc
Curating dreck. And meanwhile the human brain rots and we become nothing but 'consumers.' You see that in everyone from highly educated people to working class people---people are forgetting to think, exercise their brains, and aspire to something beyond themselves, even if it's faith. It's very handy for a certain group who want to control the rest.
If I were a potential customer for this game I would skip over all this and probably not look at the product itself because it is so painfully generic. I get excited by communications and products that show some creativity, style and quality. None of this qualifies. The internet is drowning in stuff like this already. Better to put some investment in standing out with a real human touch.
Tried to use GPT4 but still it tells me that it s based on GPT3 - even after joining the waitlist...anything I may have done wrong :-/? Any recommendations plss??
Thanks for sharing Ethan. This paints a vivid picture of just how disruptive AI tools will be (and, in fact, already are). It’s nearly impossible to keep up with how quickly this space is moving. You could probably repeat this experiment on a near-weekly basis, incorporating the latest advances in AI tools each time. What would this experiment look like 3 months from now? In 6 months. How will your inputs/prompts and the results differ?
Thank you so much Ethan, this article is very helpful and interesting, actually we are already in a world of superhumans, we just have to wait for the implications.
Would you describe yourself as a “prompt engineer “? Seeing this title lately…
Great experiment, but it appears that all the content derives from web-based material you provided at the outset. Bing just repackaged your content into the forms you requested. That's impressive, but rests heavily on the effort of the human(s) who created the original copy.
CRAZY
Ethan, Congrats on being the number 1 story on Ben's Bites today!! (https://www.bensbites.co/p/30min-ai-challenge)
Excellent use of the tools available to us. I enjoy your teaching style!