DIck Tofel takes a semi-double wrong turn

 I usually agree with him very much, but his late August piece is doubly wrong. And, I don't want to get into an argument with him, though he may have done so with my quote-stacking.

First is his attempt to defend Stanley Druckenmuller using AI to help write an op-ed for the Wall Street Journal. Actually, he doesn't even defend it. He just makes a fiat statement that it's OK, links to a Shit by Semafor that other people do that and that's that. That is, ultimately, a modal variant on the old ad populum fallacy, only with it being this time, "everybody's doing it" rather than "everybody believes that."

The other is the weird hangup he has about still not linking to Wikipedia as a source. That's even though he then adds that Wikipedia is often a very good source.

The first subissue is that this gives another round of ammo to people who say "But Wikipedia."

The second is that, other than his caveat, it's a fiat statement.

The reality is that Wikipedia is generally very good. I tell people who pull out the "But Wikipedia" bullshit to look at how many footnotes an entry has and what they link to. If it's got in-depth citations of reputable sources, that's the bottom line. 

He did respond to my quote-stack:

The point on op-eds isn’t that everybody does it, but that everybody expects it. And I thought I was being self-critical on Wikipedia; sorry if that wasn’t clear.

I debated about responding, and eventually did:

Got it on Wiki, and I’ll buy that. Maybe I was being too critical of you not appearing to be self-critical. Today’s Wiki is much better than 20 years ago. For people who say “but Wikipedia,” I respond, how many footnotes does it have and what are the sources they link? 

== 

On the op-eds? I’m not part of that everybody. Maybe that’s something that’s wrong with “Acela corridor” media or whatever term you prefer, or something that’s wrong with critics of such media.  

ow, if a Druckenmiller had outsourced his op-ed to a human ghostwriter, a PR person or whatever, then stamped his signature on it? I’d expect that. I still wouldn’t find it totally desirable, but I’d expect it. And, at least it would be a human being, and one paid to be an amanuensis. 

So, starting with that sense of “expect,” no, not me — I wouldn’t expect a CEO type to have AI do that. And, I’d find it even less desirable than a human amanuensis to do it. 

I guess next, we’ll have AI memoirs “from” or “by” the rich and famous. You won’t have to pay a ghostwriter, and you won’t even have to worry about whether or not you put a second name on that.

Next, we'll have the enshittification of the AI ghostwriter, I guess.  

I did add that, a la Wittgenstein, Dick and I may not be using the word "expect" in exactly the same way. 

Craphouse/Gannett/USA Today hop in bed with Palantir?

Gee, what could go wrong with this? Plenty, starting with the "why" of the opening graf:

USA Today Co., the largest newspaper chain in the country, announced Thursday that it is partnering with technology company Palantir to better monetize its customer data.

Assuming Craphouse has little to no restrictions on what Palantir can and will do with that data, including using it outside of the bounds of the deal, shit.

Now, supposedly, all data remains Craphouse's:

[CEO Mike] Reed said that though the company will use Palantir’s tools on its data, the data will remain the property of USA Today Co. 

“All of our data remains our data,” Reed said. “It’s our data. The actions and the intelligence that we utilize takes place on our platform. It’s our (intellectual property). It’s our outcomes.”

Not mentioned? Analytics, metadata, etc.

The story goes on to note that both Fox and Politico parent Axel Springer have already partnered with Palantir. More reason to read them less, too. 

Per the bottom of the piece, it's a conscious decision to, in our AI world, chase AI "clicks" that aren't even real:

The advent of AI has meant that USA Today Co. is essentially serving two audiences — human readers and the AI platforms that surface content to those readers — Reed said. 

“We recognize that we have to create and format content for humans and for machines,” Reed said. “And while we see a change in search-driven behavior, we also see entirely new ways to license, distribute, and monetize the trusted content we produce every day.”

I've long said that "digital dimes" turned into "mobile nickels" with smartphones. To continue the alliteration, I guess it's now "Palantir pennies" coming up. 

Getty, Shutterstock merging; And?

I saw this a week ago via LinkedIn News. Per one commenter there, it's pretty much a nothingburger unless other things change at the merged company.

Smartphone improvements and easier to use DSLRs (which have gotten cheaper with the rise of mirrorless) all put as much pressure on both Getty and Shutterstock as do AI.

The press release says that the merged company will be "cutting edge" on its own image search and use of AI. But? I've used one of the photo-generative AI programs twice. Even in a small town, I had enough internet to create Woody Allen in a cowboy hat in less than 5 minutes. Something else, not "real world" in that sense but more dynamic, took a little longer but not that much. And, when I had Substack do me an AI image for a post? Something "acceptable," which gets back to a commenter, took 2 minutes or so.

The only real hope for this merged company is a mixture of totally redefining itself and the hope for massive copyright lawsuits against AI scraping.

So, the "and," as usual with me in such situations, is ultimately rhetorical.

There is a sidebar to all of this.

It's arguable that many photos now copyrighted should be in the public domain, not because they're photos, or photos of a certain type, but that copyright law gives way too many years of protection with its various lengthenings over the last couple of decades, and that this doesn't benefit individuals nearly as much as rich corporations ... like Getty. "Who Owns This Sentence" is a GREAT book on the history of copyright.

Yeah, Getty will stick it to the people it pays for images that are new images. It's still part of the problem.