Emformation

The problem that I often have with media is that I look to it for information, not emotion. "Here's something to know about the world," is different from "here's how you should feel about this aspect of the world." Of course, given that, I should have known better than to read "Behind the Curtain: What AI architects fear most (in 2024)" when I saw it on Axios this morning. I suspected what was coming; vague warnings of an immanent threat, and nothing in the way of workable solutions that the reader could implement.

Actually, let me take part of that back. I deliberately read articles that I suspect will engage in fearmongering because I do find them somewhat informative; they tell me about what media outlets, and their sources, understand their incentives to be. And Axios was pretty clear about their incentives in this article. The conclusion of the column is a mix of anxiety triggers and flattery, directed at the reader. It's open about the idea that "realize a new problem is coming," while subtly insinuating that a large part of that problem will be the credulity of other people. Axios readers, it hints, are alert and know when someone is trying to trick them... even if "one leading AI architect" finds that they themselves "no longer can distinguish fake from real."

That "one leading AI architect" is one of five people that the Axios journalists spoke to for the column. The others were a "a former top national security official," who warned that Russian Federation President Vladimir Putin was up to no good, LinkedIn co-founder Reid Hoffman, who claimed that open-source "AI" models couldn't be trusted because only closed models were capable of self-policing, Open AI CEO Sam Altman, whose main contribution was a platitude concerning industry collaboration, and finally, "a senior White House official," who sounded the alarm over disinformation, fraud and cyberattacks.

The obligatory mention of the "sick use" of AI for revenge pornography was apparently thrown in by the authors of the column themselves. Maybe they couldn't find anyone else willing to invoke that particular hobgoblin.

In any event, no-one was saying anything particularly noteworthy or sensitive. So why were three of the column's five sources anonymous? The whole of the AI industry refuses to acknowledge advances in text, image or video generation? No one in the national security establishment would put their name to speculation that Vladimir Putin is looking to sow discord in the United States? In the entirely of the Biden Administration, there wasn't a single person authorized to say publicly that generative "artificial intelligence" could potentially make it easier for people to mislead people or mount cyberattacks? Sure, they likely wouldn't have been names that I, or most other members of the general public, would recognize. But at least then there would have been an avenue for some follow-up via Internet search.

To be sure, articles like this also give me insight into a another group of interest: the readership of these publications. This goes beyond Dahlia Lithwick's observation to Stephen Dubner that  "If you scare people’s face off, they will click." The reason why the commodification of people's anxieties works is that people don't see news coverage as playing on their fears. Instead, they see it as providing valuable information... that just happens to coincide with their existing anxieties, which are mainly about other people. In other words, the world is a terrible place because of the unintelligent, credulous or unethical among the populace (a group of people that, of course, excludes the reader), and there are people who seek out items that remind them of that, because they are also reminders of their own exceptionalism. To be sure, I don't know how large a group of people this is, in the end. All things considered, it doesn't have to be very large to keep any one publication afloat, so I'm going to go out on something of a limb and conjecture that it's not a majority. It's just large enough to keep the wheel spinning. Which, really, is all the wheel needs.

Felling The Pull

"The Israel-Hamas war is creating tensions at work" is a reasonable enough headline. It gets to the point, and deals with the subject at hand. Well, actually, part of the subject at hand. It might not be readily apparent, but the current fighting in Gaza is not the only conflict going on in the world. Humanity has always been something of a combative bunch, and there are fights going on across the globe. What differentiates the current fighting between Israel and Hamas from the war between Russia and Ukraine, the Somali Civil War or the border dispute between Armenia and Azerbaijan is twofold, a) the current level of attention paid to it by media outlets here in the United States (and elsewhere for that matter) and b) the lack of agreement on who is the aggressor and who is the victim. Although there is a third point, that lines up with both a) and b), both sides have large domestic constituencies who are making the case that they are in the right, and are deserving of support from both the public and the government of the United States. In the section of the Puget Sound region where I reside, backers of Israel have been making their case. Signs and flyers calling on people to call politicians on behalf of Israel or to donate money to Israeli causes or organizations have popped up in several areas. If there are people doing the same thing for the Palestinians, I haven't come across any yet. (Although it's possible that I've simply been unable to pick them out of the campaign signs as easily.)

Backers of both sides feel that each party has the right to be acknowledged as the single wronged party in this case, and for their own feelings as to who is correct to be validated by those around them. This isn't something that is unique to this particular conflict. What makes it different is the fact that the fighting between Israel and Hamas is happening overseas. There are any number of domestic disagreements that cause workplace (and other place) tensions, because the backers of various sides feel a sense of entitlement to the support of people who would rather be doing something else. It's one of the main drivers of what is now called "cancel culture," which goes way back in American history. The tendency is unlikely to go anywhere anytime soon, but if it is going to be addressed, it's likely better to deal with the phenomenon as a whole, rather than speaking of each of its discreet incarnations as if they were unrelated to one another.

No Problem

What gets you re-elected and what solves problems are sometimes like oil and water. And solving problems makes people mad. Nobody wants to make anybody mad because that's not how you get re-elected. But were going to have to do that if were going to fix this.

Senator Claire McCaskill. February 2011.

I was reminded of this statement from Senator McCaskill while I was on my way to drop off a ballot yesterday. While driving, I came across this campaign sign:

I don't live in Bothell at this point (although I used to), so I hadn't really been following the city council race there. But Bothell has been the site for a good amount of apartment construction. After all, it's how one creates housing for a number of people in a relatively small footprint. Given that even a relatively small single family residence can run more than $800,000 in the vicinity, for many people, apartments are going to be their only way into to living in the area.

The greater Seattle area has a homelessness problem in large part to the relative dearth of a apartment, and other multiple-family, homes in the area, which has resulted in very high housing costs. But for the people who own the housing stock, that's not their problem. The lowering of their valuations that would result from increases in housing availability is their problem. And that's where people like Mr. Swanson come into the picture.

The "apartment explosion" is barely making a dent in the local supply of housing, given the number of people who want to live in the area. And so while I won't solve the lack of housing that's driving the high homelessness numbers, it can at least reduce the upward pressure on prices. Really solving the problem would require a lot more apartments, condominiums and townhouses than are currently under construction. But for local homeowners, at least when I've been present for the conversations, the answer to homelessness is basically to force people who live on local streets to live on distant streets, so that they'll be in someone else's cities and neighborhoods.

Fixing the housing supply in the Puget Sound area is going to make people mad. An adequate supply of housing, will, more or less by definition, lower housing prices locally. And people don't like solutions to problems that come with costs for them. Especially not when there's always someone who will run for office on a platform that says that it's only other people who should pay costs.

Something For Everyone

Recently, Elon Musk met with the Prime Minister of the United Kingdom, Rishi Sunak. And, being Elon Musk, said some things that drew people's attention. I'm going to briefly touch on a couple of them.

"We are seeing the most disruptive force in history here," Mr. Musk said, before speculating: "There will come a point where no job is needed - you can have a job if you want one for personal satisfaction but AI will do everything.

"It's both good and bad - one of the challenges in the future will be how do we find meaning in life."

Most people picked up on the "but AI will do everything" part, and, no surprise, that tended to cause worry.

Musk had a potential answer for that problem during Thursday's conversation with Sunak. "We won't have universal basic income, we'll have universal high income," he argued. "So, in some sense it will be some sort of level up, or an equalizer, because everyone will have access to this [technology]."

And that's the second thing that I wanted to touch on. What Mr. Musk is saying here is that new technology will render human labor obsolete; but that won't be a problem, because it will also create post-scarcity.

I'm dubious about the first part of that formulation, and have no belief at all in the second. Generative "Artificial Intelligence" will render a number of jobs more efficient than they are today. In practice, given the way businesses operate, this means that it will render some number of them obsolete. Sure, it might create a number of new jobs, but there are unlikely to be anywhere near as many prompt engineers as there were formerly, writers, technical support specialists and other roles where a sophisticated auto-complete function can manage the workload. But despite the broad portfolio of things that Generative "A.I." can handle, it can't do everything, given that, by definition, Large Language Models don't create novel solutions when there's already an answer in the database, and don't understand the world in a way that allows them to understand when they've auto-completed their way into erroneous statements. Accordingly, the very natures of the systems involved leave gaps that will have to be closed by people. And this leaves open the idea that new jobs could come along that only humans could do. Artificial General Intelligence, code that can handle any cognitive task a human could perform as well (or better) than a human could perform it likely could, on the other hand, would carry a substantial risk of rendering human labor obsolete; because it wouldn't leave any gaps that it would be unable to close, outside of simply preferences.

But decades of predictions notwithstanding, Artificial General Intelligence doesn't appear to be on the horizon, and there's no current indication, that I'm aware of, that Generative A.I. is bringing it closer. I admit to not being an expert on the topic, but given that AGI would render GAI completely obsolete in short order, one would expect that were it close, people would be talking about it.

The subject of post-scarcity is a little different, because it's not really about how much stuff there is to go around. Not in the end, anyway. What creates a post-scarcity society is the fact that any given individual has the ability to access any resources they need to sustain themselves, without needing to trade for them; people aren't directly in a position of needing to rely on other people to give them necessary resources. It doesn't matter if there's enough food to feed a group of people if that food is in the hands of someone who a) has an enforceable property right over it and b) isn't interested in parting with it. Ubiquitous access to automation and artificial intelligence isn't going to suddenly gift people with access to the raw materials needed to guarantee themselves access to food, clothing and shelter, nor would it suddenly make private ownership of property obsolete in places like the United States.

And this leaves a number of people with a gnawing anxiety that advances in automation, and an expansion of the types and numbers of jobs that will be automatable, will leave them without access to resources, and nothing of sufficient subjective value to trade for them. And Mr. Musk's utopian vision is not going to be enough to assuage those concerns.

And Effect

While there is a lot of chatter about how wrong economists (as if economists were some sort of collective intelligence) were about recession predictions, there are a fair number of people who believe that the United States is, in fact, in a recession. And they could be right; after all, the beginning of a recession is always declared after the fact.

And one could be forgiven for thinking that a recession is still in the cards for the near future, given the steady trickle of news about layoffs in company after company. I've spent a decent amount of time on LinkedIn recently, and my feed has been something of a downer recently. There have been a number of notifications of layoffs (and even company closings) recently. To be sure, there have also been people celebrating their new roles, but those are individual events, and they don't manage to balance the news of groups of people losing their jobs. I wondered what this would mean for confidence numbers, so I hopped over to The Conference Board's website, and, sure enough, Consumer Confidence in the United States is down by 1.7 points.

I'm curious as to the degree that this is a self-fulfilling prophecy. Because sometimes, there's no better way to be ahead of an event than to be part of its cause. One company's employees are another company's customers, and each round of layoffs makes other companies that rely on the public's discretionary income think that maybe they should start cutting production, costs or both, in order to not be caught flat-footed when demand takes a dive. And the snowball simply grows from there.

To be fair to businesses, they're simply following their incentives. After all, they have a responsibility to look after their stockholders' investments, and no real responsibility to the public at large, or their employees; at least not in the same way. And positioning themselves for a recession that doesn't come is considered more responsible than not being prepared for one that does. And labor costs are seen as something to be done away with, rather than a necessity for the economy at large to function.

It's also worth pointing out that government has it's own part to play in all of this. The United States government, being unwilling to tailor either spending or taxation to prevailing economic conditions, is unable to respond to those conditions without creating more problems down the road. Although, given that the United States does have a representative government, most of the blame for this can be laid at the feet of various constituencies, especially those that have the wherewithal to fund campaigns, and those that prefer to be the passive audiences of those campaigns, rather than seeking out information themselves. (Representative governments punish the non-participatory, regardless of the reasons for non-participation.)

In any event, there seem to be a lot of forces conspiring to ensure that a recession occurs late, rather than never. The business community is simply more open about its plans.