AI goes freelance

Daily Data: What happens when GPTs compete with freelancers?

Keeping track of the deus ex machina ending to solve all our worldly troubles.

---

Everything reads better in your browser or in the app. The footnotes especially, and Random Walk is really leaning into the footnotes. Plus, if you have the app, you can set delivery to “app only” and then my daily barrage will feel less like a barrage. Unfortunately, substack does not yet have a “Weekly Digest” option, but I’m hectoring them aplenty.

If this email was forwarded to you, please click the shiny blue button:

Daily Data

---

AI goes freelance

Part of the bull case for “getting out of this mess” is an AI-driven productivity boon. It’s not stated quite so explicitly, but it’s one of the few things that conceivably leads to a “new normal,” whereas the old normal is plainly not good enough.¹

With that in mind, this neat study of GPT’s effect on freelance knowledge workers is something of a mixed bag.

Using data from the freelancer marketplace UpWork, the researchers tried to measure the effect on both the number of jobs and earnings for freelancer designers and copywriters.

The effects were pretty profound:

A ~3% drop in number of jobs, and a ~10% drop in earnings.

The really surprising (or troubling) thing is that the most successful freelancers saw the biggest dropoffs. This was no culling of the under-performers—ChatGPT came for them all.

---

AI taking our jobs?

Random Walk generally views this as a good thing—productivity gains are good, and the skilled, hardworking people who used to do this work, will find other gainful uses for their skills and hardwork. There might be some interim disruption, but y’know, so too the horse n’ buggy.

But, if you want to be pessimistic, then you say “the AIs are taking our [white collar] jobs.”

Personally, I think the general shift to cost-cutting over growth will take our white collar jobs a lot sooner than that, but we shall see.

As of now, AI isn’t driving much revenue, at least according to a BofA survey:

Only semis and capital goods reported any meaningful lift, but this is still the early innings.²

---

Strategy?! We don’t need no stinkin’ strategy!

One last thing from the survey.

BofA asked everyone what, if anything, was their AI “Strategy.” Pretty much everyone answered at least something . . . I mean, it’s a silly corporate survey.

There’s one industry, however, that could not be bothered.

It’s Real Estate.

Real Estate ain’t got no stinkin’ strategy:

As always, the man without the plan.

You do you, Real Estate, you do you.

---

Random Walk is an idea company dedicated to the discovery of idea alpha. Find differentiated data, perspectives and people, and keep your information mix lively. A foolish consistency is the hobgoblin of small minds. Fight the Great Idea Stagnation. Join Random Walk. Follow me on twitter. Follow me on substack:

1. The big news last week was OpenAI’s introduction of “custom GPTs,” which are pretty cool at first gloss, bur Random Walk hasn’t had the chance to play around with them enough to say anything interesting or different. [Edit: The actual “big” news appears to be a ludicrously amateur takeover attempt by OpenAI’s independent directors. Riding sidecar too close to the sun, apparently they let delusions of maternalistic grandeur run wild. It’s still early, and most of the details are yet to come out, but at first gloss, it’s hard to imagine a more audacious power-grab by people less qualified or entitled to grab it.]

This too was pretty cool—apparently GPTs are good enough at navigating an iPhone to successfully execute ~75% of natural language prompts like “shop for a milk frother with a budget between $50-$100.”

2. Also, while Semis definitely makes sense, Capital Goods isn’t what I would have expected, so query how meaningful it all is.