Days Without Number

A Product Director from Amazon posted on LinkedIn that he'd "asked Google, Bing, ChatCPT, Alexa, Copilot (Bing), and Gemini (Google): 'How many days between 2000-03-09 and 2002-09-30'," and "got four different answers."

When someone hands you an experiment that seems interesting, Step 1 is to replicate it. So I did. I don't have an Alexa device, so I posed the question "How many days between 2000-03-09 and 2002-09-30?" to the other four. (The base Google search doesn't try to answer the question, it simply points the user to other resources... it is a search engine, after all.)

Bing and Gemini returned the same answer: "There are 935 days from/between March 9, 2000 to September 30, 2002," only the syntax was slightly different, with Bing using "from" and Gemini using "between."

Bing Copilot gave a slightly different, and more detailed answer, noting: "There are 936 days between 2000-03-09 and 2002-09-30, excluding the end date. If you include the end date, there are 937 days." Copilot was also interesting in that it returned the dates in the same format that they were given.

ChatGPT (3.5) went to town, giving me the entire process by which it arrived at its answer:

To calculate the number of days between March 9, 2000, and September 30, 2002, we can use the following steps:

1. Calculate the number of days in full years between the two dates.
2. Add the number of days from March 9, 2000, to the end of 2000.
3. Add the number of days from January 1, 2002, to September 30, 2002.

Here's the calculation:

Full years between 2000 and 2002: 2001 and 2002 (2 years).
Total days for full years: 2 years * 365 days/year = 730 days.

Days from March 9, 2000, to the end of 2000:
March 9 to March 31 = 31 - 9 + 1 = 23 days.

Days from January 1, 2002, to September 30, 2002:
January 1 to September 30 = 31 (Jan) + 29 (Feb - leap year) + 31 (Mar) + 30 (Apr) + 31 (May) + 30 (Jun) + 31 (Jul) + 31 (Aug) + 30 (Sep) = 284 days.

Now, summing up all the days:

730 (full years) + 23 (days in 2000) + 284 (days in 2002) = 1037 days.

So, there are 1037 days between March 9, 2000, and September 30, 2002.

One of these answers is not like the others... ChatGPT was way off.

Which kind of gets to the thing with generative "A.I." The artifice is there, and fairly impressive. But they aren't intelligent systems, in the sense of having an understanding of what they're working with. The idea to start with full years in Step 1 was sound, but going January 1st through December 31st for both 2001 and 2002 was erroneous. Step 2 was flawed from the jump, as ChatGPT seemed to think that March 31st was the last day of the year 2000. (How's that for a Y2K bug?) And given that 2002 wasn't a leap year, Step 3 was also incorrect, although here it was only off by a single day. But this was the nice thing about Chat GPT showing its work. That work was wrong, but since it was possible to follow what it was doing, it was pretty clear why it was wrong; the errors were evident enough that one didn't need an idea of what the right answer should have been to realize that it was miscounting. The transparency that ChatGPT demonstrated was useful in that regard.

By the way, I did the math myself, and found that the answer is 935, or 937 days if you count both March 9th, 2000 and September 30th, 2002. So other than ChatGPT, the programs got it correct, but it does depend on how they read the question, since that determines the days on which they start and end their counts.

Escalations

Charles Darwin identifies three basic conflicts that made up the Struggle for Survival. They can be summed up as interspecies conflicts, intraspecies conflicts and conflicts versus the environment. I think one of the things that makes the concept of "evil" so compelling is that it elevates conflicts between people, either individuals or populations, from the personal to matters of the life and death of the species as a whole. And when the stakes are suddenly that high, it's easy to conclude that there's no such thing as extremism.

Aside:

For those of you who weren't here when I started The Short Form, it's mainly an antidote to my tendency to be wordy. Sometimes, thinking about what would otherwise be a long post helps me distill it down into a sentence or two, that says what I want to say, perhaps with a bit of context thrown in for added clarity.

Booted

A child's boot, abandoned in a parking lot.

The thing about found object photography is the knowledge that behind every object there is a story; one that will likely never be known.

Plotted Out

Generally speaking, people tend to write off those they consider to hold conspiracy theories as uncritical thinkers at best, and of dangerously questionable mental health at worst. Conspiratorial thinking, the logic goes, exists mainly to offer simple explanations to complex phenomena or questions.

But it's also worth noting that alleged conspiracies offer insight into the anxieties of the people who hold them. Consider the (relatively) new conspiracy theory concerning singing star Taylor Swift. You've likely heard of it, it's being covered all over the place. I won't bore you with the details (everyone else is looking to do that), but suffice it to say that for many people in the Make America Great Again crowd, the Deep State is at it again, in order to ensure that President Biden wins another illegitimate term in the White House.

While it wouldn't surprise me to learn that there was a drinking game at the bottom of all of this, I think that there is something of interest here. Namely, the fear among Trump supporters that Taylor Swift could manage to mobilize a significant number of people that Trump campaign was counting on not voting. After all, what tends to prevent the United States from sliding leftward is the fact that young people, who are the most likely to lean left, tend not to vote. Granted, there's been something of a bifurcation in "the youth vote" recently, with younger men drifting to the Right, while young women drift Left, but the fact remains that a substantial number of votes that would normally wind up in the "D" column are never cast. I suspect that people on the Right have realized that Taylor Swift has at least some power to change that.

And this is important in this election cycle because Donald Trump, while he can turn out the people who already want him to be President, has pretty much given up on making his case to people outside of that group. That is to say, for people who aren't already Trump-style populists, or deeply frightened of or enraged at the Democrats, the Trump movement has nothing to offer. In fact, one can make the case that Trumpism actively alienates people outside of Trump's base. A lot of those people were likely disinclined to vote anyway, given their general disaffection from the Democratic Party. But an endorsement from Taylor Swift could upend that.

If you'd asked me a month ago whether such an endorsement from Taylor Swift could have such an impact on this November's Presidential election, I would have told you "no." And quite honestly, I'm still mostly of that opinion. But the emergence of what looks like a state of open panic over the possibility tells me that there are a decent number of people who believe it. For the record, I'm not convinced that, say, Vivek Ramaswamy is dim enough to actually buy into this. But I think that he understands that if he's going to maintain enough support among Donald Trump's voters to land a cabinet position or the Vice-President's job, he has to talk the talk. And he's not the only one. And I think that as the people who seek to remain influential in the current Republican Party openly espouse the conspiracy, more and more Republican voters will see it as legitimate.

Which, to circle back, is part of their problem. I don't think that Taylor Swift really has the power to sink Donald Trump's chances of winning the November election. The insularity of the Trumpist base is doing that work for her. I know a few people who are big on Trump, and, even more so than most people I know who are politically active, outreach to people who don't already agree with them is simply not on the menu. I'm of the opinion that this is a bigger problem for them than any celebrity.

But Wait, There's More

I was out at a happy hour yesterday. For the food basically, I never learned to drink. The televisions were on, muted, and turned to the local broadcast channels. The evening news was on. The number of crime stories that ran was somewhat surprising. I know that news outlets have a tendency to run crime stories; "if it bleeds, it leads" and monetizing scaring people, and all that. But it seemed that more than half the newscast was simply one crime story after another. Given that the sound was off, I didn't get any of the details, but they wouldn't have been informative anyway.

But it explains why people in the local area claim to be afraid to go into Seattle unless they're armed. Which strikes me as ludicrous, but I suppose if  one sees "Man shot after walking past homeless encampment," multiple times in an evening (it seems that they run these stories on repeat), it starts to make a certain level of sense.

The faceless 'burb that I live in is dull by pretty much any standard. In the nearly two years I've lived in my current place I've seen the police mainly because they seem to show up across the street on a recurring basis. The only time other than that I'd seen officers in the neighborhood was when they were talking to the homeless guy who used to live in a van around the corner. (He cleared out just prior to the cold snap that landed in back in January, and hasn't returned.) So I'm always somewhat amused when people openly wonder if it's "safe." But then again, I'm a native of Chicago... and lived there when violent crime peaked nationwide back in the early 1990s. Nothing that happens here phases me.

But I've also stopped watching broadcast news, which always seems able to find a slew of random crime stories important enough that everyone needs to know about, but not enough so to be genuinely informative. This makes me less informed about "dangers" in the local area than some people like, but I find that I sleep better, so it's been a worthwhile trade.