March 19th 2025
Glimpses of a Much More Useful Future
My fountain pen, which is simultaneously my messiest pen and my favorite pen.
I never felt that as a ghostwriter on Fiverr I was doing the work I would choose to do if the money didn't matter. Shortly before ChatGPT went live, I spent significant time wondering whether or not I should allot more of myself to the writing projects I kept tucked away in my “own_writing” folder and less on the ones in my “freelance_writing” folder. The money I made freelancing at the time was good, though – excellent, in fact. For all the thought and energy that I put into meaning and purpose, it seemed foolish to turn down all those big checks. A bird in the hand and so on.
I don't regret making the decision that I did. Because of the time that I devoted to high-paying freelance projects in my 20s and early 30s, by the time ChatGPT started eating into the business I had built as a Top-Rated Seller on Fiverr, I had set myself up well enough that I could take my time figuring out what I wanted to do next. I worked my way through the ideas that I had written down and never gotten to for Bingeclock, my side project for media tracking. I read more. I set aside evenings for dating and met the love of my life.
A month ago, Fiverr invited me to attend the launch of Fiverr Go, their own take on generative AI presented in part as a substitute for freelancers' creative skills. I met Fiverr's CEO, Micha Kaufman, whose passion for the company was striking and whose demeanor was warm and approachable. We stood together on the podium for the opening bell of the New York Stock Exchange. Because it was such a nice experience, I hesitated to spoil it by offering anyone my honest opinion of Fiverr's explicit shift into facilitating AI model deployment.
My honest opinion is this: Fiverr is making its only sensible move at a time when generative AI is causing a precipitous drop in the value of its core offering, and the company is thus in turn also making an admission that the system its business is predicated upon is broken.
In a recent talk with David Eagleman, Microsoft's Prime Unifying Scientist Jaron Lanier posited that as AI models continue to improve and proliferate, they will drive down the value of any individual content (writing, graphics, audio) even more substantially than they have already. He then argues that in order for creators to capture some of the small value of the data they produce or help produce, new systems will be necessary.
“You treat your data as if it were labor,” Lanier explains. “I want that future, and not just a future where everybody has to go on universal basic income and feel useless.”
I want that future too.
Trying to explain the problem inherent in eliminating human contribution from data production, one may slide into circular reasoning. (It's bad to cut out the humans because it's good to keep the humans involved and it's good to keep the humans involved because...) I again defer to Jaron Lanier's eloquence in this area: "I don't think we have any choice but to somewhat mystify and elevate humans."
Fiverr's choice to incentivize creators to use more generative AI in their work, rather than less, represents a misinterpretation of the size and direction of the change that is taking place in the market for human creativity. A future in which data is treated as labor means one in which the data itself, and not just the resulting virality of the service using that data, is incentivized by compensation.
Fortunately, the US judicial system seems to be taking steps to cultivate an environment that may encourage new human-centric economic systems to emerge. On March 18th, the US Court of Appeals for the District of Columbia Circuit ruled that art generated by AI is not eligible for copyright protection, upholding Judge Beryl A. Howell's 2023 Thaler v. Perlmutter decision that "human authorship is an essential part of a valid copyright claim."
Create a record of your human authorship by doing your writing with the iTypedMyPaper human writing authentication tool.