It’s Saturday morning in California, and I’m a little bleary-eyed after a trip to Mexico City in connection with my day job. I was going to sleep in this morning, but I ended up reading a Forbes article that Mitch Wagner shared. Then I thought about it. Now I’m thinking about it some more.
What the Forbes article said
The article in question is ‘I’ve Never Hired A Writer Better Than ChatGPT’: How AI Is Upending The Freelance World. The credited author is Rashi Shrivastava, and it’s obvious that the article wasn’t written by a bot.
Shrivastava obtained the quote in the article’s title from Melissa Shea of Fashion Mingle, who has used (and still uses) freelancers to write marketing copy and perform other tasks.
Why did Shea say what she said?
At $0 an hour, the chatbot can crank out more content much faster than freelancers and has replaced three content writers (Shea) would have otherwise hired through freelancing platform Upwork.From https://www.forbes.com/sites/rashishrivastava/2023/04/20/ive-never-hired-a-writer-better-than-chatgpt-how-ai-is-upending-the-freelance-world/?sh=1d556b7262be
Shrivastava then goes on to say that this affects the Upwork, Fiverr, and related platforms in two significant ways:
- People like Shea are less likely to go to Upwork and Fiverr to seek people for non-specialized tasks. After all, even the cheapest content creators from the most impoverished developing nations cost more than today’s free generative AI tools.
- People seeking jobs on the platforms are trying to get an edge by using…you guessed it, AI.
In early April, business consultant Sean O’Dowd uploaded two job postings on Upwork and within 24 hours he received close to 300 applications from freelancers explaining why they should be hired. Of the 300 proposals, he suspects more than 200 were done by ChatGPT, he says.
Obviously O’Dowd isn’t going to pay for something he can get for free, and others who hire freelancers caution in advance that they won’t pay for work in which the freelancer uses AI.
(I assume in this instance “use” means “generate the majority of the content automatically.” Shrivastava mentions a graphic designer from Nigeria “who uses ChatGPT to help him ideate,” and I’ve used the same technique myself. For this post, I asked a generative AI tool to describe the benefits of case studies for Inland Empire businesses. I threw out everything that the tool wrote except for five short bullets. Which I then changed to six short bullets.)
What the Forbes article didn’t say
Rashi Shrivastava was examining the relationship between generative AI and freelancing TODAY. That in itself is a handful to analyze, but there’s an entirely different question to address.
I addressed this other question in a Bredemarket LinkedIn post earlier this morning providing my thoughts on Wagner’s share of Shrivastava’s article. Now I’m going to be lazy: not by using generative AI to write the following paragraphs, but to simply reprint what I already said on LinkedIn. (I guess I could use generative AI to rewrite the post using different words, but…well, we’ll get to that.)
There are so many things to unpack here, but I will confine myself to one (for now):
What happens when (not if) the better generative AI tools are no longer free? What happens if the better tools cost MORE than a beginning writer in a developing country? Or the worst (best?) case scenario: what happens if the highest quality tools cost more than an experienced writer in a developed country?
As time passes, we will see significant differences in these tools.
Some will crank out basic work with outdated data; these will be “good enough” and will be cheap and may even continue to be free.
Other tools will be constantly be retrained and provide proper attribution of training data. Because of significant investments in R&D, the better tools will provide higher quality work – maybe even better than Mitch Wagner’s work (or my own). Perhaps you may even need to “program” them – an order of magnitude more complex than proper prompt phrasing.
At that point, someone requiring high quality content may decide that it’s cheaper to let a human do it.
What generative AI says
So that was what I’ve written about a future in which the quality and cost of generative AI tools will vary dramatically. But what would generative AI write about its own future? Since the service is still free, I thought I should ask.
In the past I posed a question to generative AI asking if it was after my job. If I recall correctly, I posed the question to Google Bard, and it diplomatically answered that humans can never be replaced. (But what was the source of its information? Google executives who want to minimize the chances that Congress will clamp down on generative AI use?)
In this case, I’m turning to ChatGPT. My prompt was “What happens to the economics and usage of generative AI when the higher quality services are no longer free?” The remainder of this section is ChatGPT’s response.
If higher quality generative AI services are no longer free, it is likely that the economics and usage of such services will be affected in several ways:
- Reduced usage: Some users may be deterred from using the service if it is no longer free. They may look for alternative solutions or reduce their usage of generative AI altogether.
- Increased cost: Users who continue to use the service may have to pay for it. This could increase the cost of using generative AI, which may limit its adoption.
- Competition: If one provider starts charging for generative AI services, it may create opportunities for other providers to offer similar services at a lower cost or with better features. This could increase competition in the market and ultimately benefit consumers.
- Quality: If providers start charging for their services, they may have more resources to invest in improving the quality of their generative AI. This could lead to better results and more satisfied users.
In summary, the economics and usage of generative AI may be affected if higher quality services are no longer free. However, the impact will depend on the pricing strategies of providers and the behavior of users in response to the change.
Postscript: a little more about Mitch Wagner
Why did Mitch Wagner share the Forbes article in the first place?
Because generative AI affects him. Or maybe it doesn’t.
Read his LinkedIn biography (as of today; writers sometimes like to edit themselves) and judge for yourself if generative AI threatens someone with his broad experience.
I’m a marketing writer and tech journalist, with more than 25 years of experience telling stories about companies using technology to deliver business value.
I’m expert at interviewing, researching, writing and editing articles and newsletters for audiences at all levels of technology understanding on daily, weekly, and monthly deadlines. I’ve worked solo, led a small full-time team, and large teams of freelance contributors. I’m skilled in hosting podcasts, webinars, and video.
Previously, I worked at Oracle, part of a team of writers, editors, and designers telling stories about how the company and its customers are innovating and inspiring.
Prior to Oracle, I was an executive editor and writer at top tech journalism publications, covering the early days of the Internet, social media, and wireless communications.
View my portfolio of writing samples here: https://authory.com/mitchwagner
Another postscript, this one on the post’s title
I did not write the title to this post.
I went through several titles, settling on “Tomorrow’s Generative AI Landscape Will Differ From Today’s.”
I then fed the following prompt to ChatGPT: “Rewrite the title “Tomorrow’s Generative AI Landscape Will Differ From Today’s” to optimize clicks”
ChatGPT returned this: “Revolutionary Changes in the Generative AI Landscape Coming Tomorrow – Don’t Miss Out!”
Unfortunately, ChatGPT had no way of knowing that BuzzFeed shut down this week. I was busy in Mexico City and missed the news, but luckily for me Mitch Wagner shared that news item also.
Even though Buzzfeed’s issues were in its business model and not its headline style (if you’re unfamiliar with the Buzzfeed headline style, read “7 Insanely Clickable Buzzfeed Headlines and What You Can Learn From Them“), I thought I should tone things down just a bit.
After some tweaks, I ended up with “Revolutionary Changes in the Generative AI Pricing Model Coming Tomorrow: How Will They Affect the Output?”
Although I kept the word “revolutionary” (even though it’s not sage-like). Because reasons. (Hint: my day job’s archetype is NOT Sage.)