Twinned

When I tell people that I don't do much with generative automation in my personal life, they'll sometimes ask if I'm in the "anti-AI" crowd. Which is reasonable... generative automation is catching on, and a lot of people see real potential in the technology.

And so do I, for that matter. But at present, I'm a bit cautious about it. Call it a variation on Erwin Knoll's "Law of Media Accuracy," which notes that the media is always right... except for the rare stories of which one has firsthand knowledge. And, interestingly, I've found this to be true. Media stories of which I have firsthand knowledge are exceedingly rare; but in the few cases I've encountered, I've always found errors in the details. Nothing major, but noticeable.

And I've noticed the same with generative automation, especially with Google, since Gemini is now incorporated into the search functionality. About six months ago, I noted that the LLMs would note that Aurora, Illinois was a fictional city. Today I was looking for something significantly more niche.

There is an old tabletop role-playing game called Gemini. It was published by a Scandavanian company named Cell Entertainment at around the turn of the century.

Now, the fact that it shares its name with Alphabet's generative automation offering is going to be confusing right there. So the fact that someone reported difficulty in finding it via a Google search isn't a surprise. But Google didn't really help matters any...

The fact that the AI Overview doesn't include Gemini in its list of "Tabletop role-playing games with 'Gemini' in the name" speaks to Knoll's Law. I know that there's a tabletop role-playing game named Gemini, only because I have a copy. If I didn't, it would be reasonable to conclude that it didn't exist, even though the AI Overview never actually says as much.

To be sure, if one includes "Cell Entertainment" in the search, then the AI Overview is right on the mark, noting the year of publication and a general description of the setting. But this requires knowing the publisher of a game that was obscure when it was first published, more than a quarter century ago.

And this is why I don't do a lot with generative automation in my personal life; the amount of information that I feel I need to have about a subject to assess the automation's accuracy (and hence, its usefulness) obviates the need for the automation in the first place. And the fact that, when the name of the game is paired with the name of the publisher, the AI Overview recognizes that it's a tabletop role-playing game calls into question the processing of the original search terms.

Sure, I could feed prospective Nobody In Particular posts into Perplexity, Gemini or Copilot and task them with editing things for me... but the whole point of this exercise is to make me a better writer. Outsourcing that task to a generative automation system defeats the purpose.

So, like I said, I do understand the potential of generative automation. But I think that it's going to be a while longer before that potential is realized. 

Autumn Fruit

Today I learned that there are apple varieties that mature in mid to late November. It hadn't occurred to me that apples grew to this late in the season, but I walked past a house that had a tree in the front yard. (The apple tree that overhangs my yard tends to have dropped all of its fruit by mid-September.) I presume that it has something to do with the local climate; the general lack of frost makes for a longer growing season.

Trolling for Cash

Almost immediately, “people started noticing that many rage-bait accounts focused on U.S. politics appeared to be based outside of the U.S.,” The Verge said.
X update unveils foreign MAGA boosters

Number of people surprised by this turn of events: Probably zero.

The piece points out that for people in Africa and Southeast Asia, the income provided by winding people up on X can be significant. And when people need income, they're not always going to be picky about how they get it. The modern Internet is awash with people who are pretending to be someone and/or somewhere they are not, because their real identities and locations aren't valuable. After all, no one really cases what a Russian or an Indian thinks of American politics. But this random fellow American that one knows nothing about who just happens to have something to say that aligns with one's preconceptions? They're certainly worth a follow and not all manipulating people.

Because, supposedly, whatever viewpoint or information is on offer helps prove that The Other Side is not only wrong, but deliberately perverse. And that's the important thing. With Americans desperate to score points on one another, and that desperation having advertising dollars attached to it, it's to be expected that people from other countries would want to get in on the act, even if they aren't part of a foreign influence campaign. Income, after all, is income. And the worldwide nature of the attention economy allows people from poor countries around the world to participate in it.

Of course, it's not just the Make America Great Again crowd that's willing to pay attention to anyone willing to provide them with ammunition in the culture wars. The tally may be lopsided, but both sides have their free agents. In part because if the offshore voices were all MAGA, that would have been suspicious, but even in a culture war, there's work for mercenaries on all sides of the conflict.

Now that the cat is out of the bag, I expect that a number of accounts will either suddenly shift to the United States, or disappear, some to quickly reappear with new, "onshore" identities. Those operators who can't afford a VPN that will allow them to pretend to be in the United States will simply go away, as the proof of their locations renders them useless as culture war combatants. So I suspect that the rollout of account locations on X will benefit the formal influence campaigns at the expense of the free agents. Either way, the rage-baiting will continue; there will still be money in it.

Wishcraft

A question, that I wonder about sometimes. 

Given the degree to which the public in much of the developed world (and especially the United States) seems willing to support policies that are based on little more than wishful thinking, why would anyone expect that politicians drawn from said populations to be exclusively the people who don't?

While the Trump Administration is often a fairly clear example of governance by wishful thinking, the main difference between the current President and his predecessors is really simply a matter of how often his preferred outcomes seem to be completely at odds with the facts on the ground. Because honestly, President Biden wasn't anywhere near above wishful thinking, either.

But there's a constant undercurrent, at least in the media, that people should (and often do) vote for people who approach the world clear-eyed and rationally, rather than with the same emotional attachments that the voters themselves have. Even when it seems pretty clear that policy matters are being decided on emotional considerations, the idea that the people calling the shots somehow know better is always in the background.

It's true that not everyone thinks of the world in terms of "this is what I want, so there must be a low or no-cost way of bringing it about." But I don't know that it's reasonable to presume that no one in the political class would be drawn from that group. Especially when it's connecting with people, rather than a firm grasp on the likely outcomes of policies, that tend to drive people's voting patters; especially the marginally-attached "swing" voters whose choice to show up or not on Election Day tends to drive the outcomes.

President Trump is unusual in the sense that he's often very public with his apparent belief that things his voter base think are easy actually are easy. He's aided in this by the devotion of said voter base; this allows him to scapegoat the Democrats, the "deep state" or what -have-you whenever things actually turn out to be difficult. So he doesn't have to fall back on "things are complicated" when policies don't produce the claimed results. But he's not unique in thinking that policies that could come across as obviously flawed were going to be winners.

After all, he's a member of a public that tends to see the world this way, and likes to see their worldview mirrored back to them by their elected officials. 

Survey Service

I’m tired of feeling baited-and-switched by companies that ask customers to “rate their experiences,” when they’re actually being requested to rate the individual employee(s) they were interacting with. I understand that companies don’t really want feedback from the public; a lot of it is negative, some of it is completely unreasonable and just about all of it is geared towards customer priorities, rather than business (or shareholder, if we’re being honest) priorities. No business wants to hear that their processes are clunky or their inventory selection is lacking. And when they’re the ones writing the surveys, they don’t have to. I get it.

But an interaction with an individual employee (or three) is not the same as an interaction with the company as a whole, or even some function or division of the company. A customer support representative can be perfectly charming and willing to help, but be stymied by disempowering policies or simply a lack of the resources needed to serve the customer. And that leaves aside all of the various touchpoints that come into play both before and after the interaction.

And for any rating system that works on a scale of 1 to X, so long as X is greater than 2, a score of X-1 shouldn’t be considered a black mark. The expectation that people are going to do absolutely amazing work, and blow people away with every interaction is unrealistic. If give someone four stars out of five, I shouldn’t have a supervisor contacting me to offer up their head on a plate. Five out of five is not my base level of expectation; leaving some room at the top of the scale for truly exceptional service should be an option.

In the end, it dissuades me from engaging with these surveys and ratings systems. Service jobs can be hard enough as it is, people don’t need the added stress of me being supposedly the one person on Earth who still sees a difference between “did the thing I asked of them,” and “completely blew me away.” Especially when the responses are not anonymous.

Okay, rant over. At least until I can corner someone a little higher up the food chain in one of these companies. Then, I’m going to give them a talking-to.