tag:blogger.com,1999:blog-13966821991781225302024-03-17T17:09:39.675-07:00Nobody In ParticularJust Another Random American.Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.comBlogger3060125tag:blogger.com,1999:blog-1396682199178122530.post-86146330964566599752024-03-17T17:08:00.000-07:002024-03-17T17:08:52.866-07:00Blocked<p>I suspect that one of the barriers to reducing, let alone eliminating, poverty is simply this: The exploitation of poverty creates active incentives to perpetuate poverty, because poverty itself becomes a resource. And one that's much more renewable than many others one might name. As I've grown older, I've come to believe that the "wealthy world" including, or perhaps especially, the United States, has built its economy in such a way that were global poverty to go away tomorrow, the system would quickly become unsustainable. The idea that the world's leading economies would structure things in a way that could actually solve global poverty presupposes that people who have built their affluence on the exploitation of poverty are careless enough to kill even a shabby goose that lays golden eggs.<br /><br />And, to be sure, I'm not simply talking about nameless, shadowy "élites" in a figurative smoke-filled backroom somewhere. Irrespective of how precarious they feel their lives might be, a lot of middle class Americans (and likely some people for whom the middle class is out of reach) rely on the low labor costs, and thus poor returns on labor, that poverty enforces to have the comforts they do have in their lives. And people are loathe to surrender their comforts, even when it can be shown that they come at a cost to others.<br /><br />This strikes me as a recurring theme in life. When I was young, Generation X's "generational anxiety" was about the nation's level of debt, which began to climb sharply after Ronald Reagan's slashing of income (and other) taxes. But that anxiety only seemed to last long enough for the cohort to understand that if government spending was to be brought under control, "we" would need to be the ones on whom less was spent. And the political establishment understood that; so while there's political benefit in pointing the finger at someone else and vowing to raise their taxes, telling the public as a whole it's time to pay the buffet bill is a non-starter.<br /><br />There are constituencies for the continuation of all of life's big problems. I'm curious how many of them I'm a member of.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-42158712395909793622024-03-14T20:04:00.000-07:002024-03-14T20:04:21.318-07:00Rewrites<p>One of the worries that I'd heard voiced about Generative AI concerned the biases that might be present in the data. This prompted me to wonder, what could you understand about the dataset and/or the training that went into it, given the sorts of responses a system gave to prompts?<br /><br />To be sure, I have no idea of how to determine this... I simply don't know enough about how these systems work "under the hood" as it were. I'm a dabbler, not an engineer. But since the genesis of this series of experiments was someone finding that different systems gave different answers to the same prompts, another test came to mind. Last week's experiment had Copilot, Perplexity, Gemini and ChatGPT 3.5 translating a snippet of a document I'd written in romanized Japanese some two decades ago. This week, I gave each of them the entire translation, as created by Copilot, and had each system re-write the text. In a nutshell, the text is a brief narrative about Tom, who works for a bank in Sim City.<br /><br />Copilot and Gemini both did two things that stood out from the other two: 1) they created titles for the story, and 2) they re-ordered some of the details of the story. In the original, I start by noting that Tom works for a bank, but don't note that he's the branch manager until later. Copilot and Gemini note both Tom's workplace, and his position when they introduce him as a character. Perplexity and ChatGPT 3.5 had their own similarity: they created very similar text. The first sentence for each matched; literally word-for-word, and there is a sentence in the middle of the story where the two systems varied only by a single word.<br /><br />Gemini's rewrite was brief, managing to trim the text by almost 40 words, nearly a quarter of the total. Copilot, conversely, was the most verbose of the systems; it was the only system where the re-written text was longer than the text I'd submitted, by about a dozen words. Mainly because it tended to add little flourishes into the final document, but also because it cut the fewest corners in noting the details of the original text. To be sure, however, all of the systems had trouble with the details, sometimes appearing to miss the nuances.<br /><br />In the end, despite what I said earlier, I think I can start to understand something about the "interior" of each system from this test, given that I'm already starting to build a set of expectations of what each would do with a given input. I expect it will take several more trials to distill what seem like "personalities" into a distinct set of rules that each operates by. The fact that I'm not an engineer would make the task longer, but it seems doable. Which makes sense; it's the differentiator of systems that are otherwise doing the same thing.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-5733436828107895222024-03-13T09:11:00.000-07:002024-03-13T09:11:22.743-07:00Uncommitment<p>Today is the 13th of March, and that makes it, among other things, the day after the Washington State Presidential primary election. Not that it matters; both the Democratic and Republican primaries, to the degree that they were actually contested, were already over on the national level.</p><p>But there's always something interesting going on, and yesterday, I came across this sign, stuck in the parkway near a local grocery store:</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhJKlwPTd5-p8w_8MR4COtBn60PY7LvJznFi0AP88evMrOqz01asM1xwJR0t3jKnXnUyguApLESiJjv27COVhIHRGpSnxVgpygisnrrjMXgH1lF2AhgIRvvaFoIf8Qz_ne48vVM5A03WofVBRphPRJZom3lJI0Mrqaq2ZPhthtRGwPUVJn2CQKZ2iENHrs/s1503/Uncommit.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1002" data-original-width="1503" height="213" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhJKlwPTd5-p8w_8MR4COtBn60PY7LvJznFi0AP88evMrOqz01asM1xwJR0t3jKnXnUyguApLESiJjv27COVhIHRGpSnxVgpygisnrrjMXgH1lF2AhgIRvvaFoIf8Qz_ne48vVM5A03WofVBRphPRJZom3lJI0Mrqaq2ZPhthtRGwPUVJn2CQKZ2iENHrs/s320/Uncommit.jpg" width="320" /></a></div><p>I don't know when the idea that a vote for an "Uncommitted Delegate" became a sort of vote against continuing Israeli military action in Gaza. And I suspect that if I'm not clear on this, the campaign to re-elect President Biden to another four-year term isn't either.</p><p>It's a strange way to conduct politics: the Biden campaign is meant to take away from this that there are voters who are unhappy with the relationship between the United States and Israel, but then what? In the event that the conflict between Israel and Hamas simply drags on, an incoming Trump 2.0 administration is unlikely to request that Israel, dial things back... the Evangelical portion of the Republican/Trumpist base tends to be fairly Zionist and unconcerned with the plight of the Palestinians. So is the threat here that people will withhold their votes in November?</p><p>Given the nature of the Electoral College and Washington State politics, one may as well put Washington's 12 Electoral College votes in the "Joe Biden" column right now. This place hasn't been competitive for decades, and isn't likely to become so anytime soon. So there's no real leverage for a small number of ceasefire supporters. (If there were more of them, and they had more resources, their message would be more widespread.<br /></p><p>That makes this a signal that has a difficult time carrying any information. Politics tends not to work in the way that many people want it to; policymakers tend to have a level of insulation from all but the most intense levels of direct public opinion. Given this, signals have to be clear and unambiguous in a way that common public channels rarely are.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-25955359983660807602024-03-10T21:19:00.000-07:002024-03-10T21:19:48.590-07:00Unscriptured<p>Back when I first started this project, I <a href="https://aaronmclin.blogspot.com/2006/12/police-states-of-america.html">wrote</a> about a group of activists who met every Saturday to protest the wars in Iraq and Afghanistan. And I'd dropped in on them from time to time, to see how things were going. With the end of active military operations in the two nations, the protests wound down.</p><p>But with the war between Israel and Hamas, and the common idea that Israel is dependent on the United States to the degree that President Biden could effectively end Israel's ability to prosecute the war, protests are on again, in the same place at the same time.</p><p>I made time to drop in on the protests this week, mainly to get some photographs; it's the sort of thing that I find to be worth recording. I chatted with the protestors for a bit, noted both their small numbers and the absence of any counter-protests, and snapped a few photographs. Far from the paranoia of the mid-00s, a couple of the protestors wanted to make sure that I was able to get clear pictures of their messages.</p><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto;"><tbody><tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiY4zPNdBcoGYLLxvQi6uOUhvfkYRgvGYshzPK_gDNJzGHZT7pE57AsbClXAf8jKckureNmKjuyyy5Q1MlqpTkj6zebrdQBPG9lBz61Jw9O9JKY5_Xm9fjG4l42d2u3_eBvrfN-R9hgamW7iQ_Y-7Yq9izsswhIrnc_gPq26Pt2M4Rbnm031TwU_oqJo_M/s1503/Unscriptured.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" data-original-height="1002" data-original-width="1503" height="213" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiY4zPNdBcoGYLLxvQi6uOUhvfkYRgvGYshzPK_gDNJzGHZT7pE57AsbClXAf8jKckureNmKjuyyy5Q1MlqpTkj6zebrdQBPG9lBz61Jw9O9JKY5_Xm9fjG4l42d2u3_eBvrfN-R9hgamW7iQ_Y-7Yq9izsswhIrnc_gPq26Pt2M4Rbnm031TwU_oqJo_M/s320/Unscriptured.jpg" width="320" /></a></td></tr><tr><td class="tr-caption" style="text-align: center;">I've had four years of theology classes and I'm pretty sure I don't recall that being in the Gospels.<br /></td></tr></tbody></table>I'm still unconvinced that street protests like this, especially when they are small scale, actually get anything done. But I will give them credit for persisting.<br /><p><br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-64581188139531521432024-03-07T21:07:00.000-08:002024-03-07T21:07:56.280-08:00Romance On Demand<p>Experimenting with generative A.I. reminds me of how much I enjoyed software testing. I find coming up with interesting, and plausible, use cases and seeing what the systems do with them genuinely fun.<br /><br />Back when dinosaurs roamed the Earth, some friends and I took Japanese lessons. For one of the exercises, I wrote a short story about a bank manager named Tom. Because we were only learning spoken Japanese, I wrote it in rōmaji, or Latin script. Fast forward two+ decades, and I barely had any idea of what it says. So I dropped it into Google Translate; which recognized it as Japanese, but its "English" translation was basically just a trimmed version of the original rōmaji text.<br /><br />So I figured I'd see what the generative A.I. systems would make of it. I asked Copilot, Perplexity, Gemini and ChatGPT 3.5 "What does this say:" and then dropped in a snippet of the text. Gemini took the prompt to be a request for information about Tom, and noted "I do not have enough information about that person to help with your request." Perplexity's translation was a bit redundant in places (and somewhat confusing for that), but it was close to the answers that ChatGPT and Copilot gave.<br /><br />Getting Copilot's answer, however, was a bit of work. It initially took the romanized Japanese I provided, and wrote it out in Japanese characters, using Kanji, Hiragana or Katakana as (I presume) appropriate, so I had to then ask it to translate that text into English for me. It seemed to be fairly true to what I sort of remember writing back in the day, so I took it a step further and dropped in the entire story, which had more details of Tom's commute and how he spends his weekends and includes another character, Noriko. At the end of another two-step translation, Copilot presented some interesting choices for follow-up prompts, like: "How did Tomu-san and Noriko-san become friends?" Curious, I clicked on it.<br /><br />It was an interesting exercise in generative pre-training hallucination, as Copilot spun up the plot of a cozy, cheesy romance novel, with Tom and Noriko as the stars. (Forming an interesting contrast with Gemini.) I can see how building an LLM that's programmed to allow it to expansively "infer" things from a short text sample can be useful, especially given that Copilot clearly noted that it had engaged in a "whimsical journey," but I think that I would have built the system to make it clear up-front that the offer is for what's effectively a work of speculative fiction. That would also give the system a chance to ask just what sort of fiction the user wants; as I would have chosen a much different theme than Copilot's derivative romance plot.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-56928920197003712762024-03-06T21:12:00.000-08:002024-03-06T21:12:14.261-08:00Two-Sided<p>It's been interesting how much bandwidth the Israel-Hamas war, and by extension, the conflicts between the Israelis and the Palestinians more broadly, has been taking up here in the United States. I've listened to a number of podcasts on the events, even without doing anything to seek them out. And one thing that's occurred to me from listening to multiple people discuss both the current conflict and its history is the idea that both sides will appeal to other parties when it suits them, while denying their legitimacy when it doesn't.</p><p>One of the arguments that I'd heard from people supporting the Israeli side of the conflict was that Israel agreed to the terms of the United Nations partitioning of the former Mandatory Palestine, while the Palestinians (and other Arabs in the area) did not. The common Palestinian counter to this is that the United Nations had no legitimate right to hand over Palestinian land to the Jewish residents of the area. It makes sense for the Israelis to support the right of the United Nations to give them land, and for the Palestinians to dispute that right, but both positions are fundamentally self-serving.</p><p>Likewise, the Palestinians and their allies in the international community feel that the United Nations should be able to call for a binding cease-fire. For the Israelis, on the other hand, the only international voices that matter are those that support them, most notably the United States, who they rely on to block any criticism of them in the Security Council. Again, this is understandable, even though the positions are self-serving.</p><p>In the end, this will likely be one of those situations which has no good end. Both parties feel that they have the more legitimate grievance, and their historical views of the conflict tend to have starting points carefully chosen to support their specific viewpoints. Just as one man's war crime is another man's smart fighting, one man's justified reprisal attack is another man's atrocity. No one ever sees themselves as abusive or evil, and rarely, if ever, do they understand their reasoning as self-serving. That's left to the people they don't listen to.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-23221117544352203992024-03-05T19:22:00.000-08:002024-03-05T19:22:24.124-08:00Marked<p> </p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjendxmq6Bm3PuNCTJpcAe8kCQqxelbRiMnz0Yr5dFp9SPBwjTdz_6jaSY3xTiEBGyW1Po0B6pH4hZkBTE1ZEPp0ikB9diFSpxueJIHc7F9ebyn4oIduHgsQus_JbkLlZqRYQX7TF6sDxGx8ZMS-6LQWMBGzeBfqJfR5jVo4Oz6AmN9ADQuYJsaXXtRWZw/s1503/Banner.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1002" data-original-width="1503" height="213" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjendxmq6Bm3PuNCTJpcAe8kCQqxelbRiMnz0Yr5dFp9SPBwjTdz_6jaSY3xTiEBGyW1Po0B6pH4hZkBTE1ZEPp0ikB9diFSpxueJIHc7F9ebyn4oIduHgsQus_JbkLlZqRYQX7TF6sDxGx8ZMS-6LQWMBGzeBfqJfR5jVo4Oz6AmN9ADQuYJsaXXtRWZw/s320/Banner.jpg" width="320" /></a></div>Graffiti is, for the most part, impenetrable to me. I can admire the clear artistry that goes into some of it, but the message is generally beyond me. Such is the case here, with the spray-painted scrawl I encountered on the boarded-up front of what had been a Staples office-supply store in the area. I presume that odd mix of Christian and Egyptian motifs means something (other than the tagger being mentally ill) but I have no idea that that something might be.<p></p><p>This sort of weirdness pops up a lot, here in the Seattle area, and more so as one gets into the city proper. It's the sort of thing that local Republicans decry as the beginning of the end of any semblance of Law and Order in the region, while local Democrats often feel it doesn't rise to the level of criminality. Political arguments (and talking points) aside, it does seem as though the area is in something of a tug-of-war over whether the local architecture should be a canvas or not.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-63921970126275458912024-03-03T20:43:00.000-08:002024-03-03T20:43:56.724-08:00Buzzy<p>Artificial Intelligence is any number of things. Two of them are 1) a technology and 2) a buzzword. This is common; it happens over and over and over again; this isn't a new phenomenon. When a technical term becomes a buzzword, it's spoken of as if the things one does with it are somehow fundamentally different than they were previous. Take e-commerce. Placing an order with a vendor in another location is certainly <i>easier and faster</i> via the World Wide Web than ordering from, say, a physical mail-order catalog. But the fundamental processes on the back end are more or less the same. And the goods and services aren't any different. Ordering a book on Amazon has any number of similarities to me going over to Barnes and Noble and requesting a book.</p><p>I mention this because I found a brief article in The Week about "AI"-generated pornography. It quoted Parrots Lab founder Naja Faysal, and I decided to look up the <a href="https://medium.com/@najafaysal/ai-porn-is-here-nobody-cares-517c1b51b53c">Medium post</a> that the article referenced, to read it for myself. One of the points that Mr. Faysal makes is that recent advances in generative AI raise "critical ethical questions about the representation of consent, the portrayal of healthy sexual relationships, and the potential impact on human empathy and connection." Fair enough. But doesn't pornography created by other methods raise those same "critical ethical questions?" I'm not seeing what this one specific technology is doing that changes the ethical landscape around depictions of human sexuality for the viewer's pleasure. The invocation of AI here seems to be more a means of getting people's attention than a genuinely salient factor.<br /></p><p><br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-15435739670110387062024-03-01T22:12:00.000-08:002024-03-01T22:12:18.202-08:00Or Not To Be<p>For this week's random act of large language model experimentation, I wanted to know how the systems would react to a request for a model of a thing, rather than the thing itself.<br /><br />To this end, I asked Copilot, Perplexity, Gemini and ChatGPT 3.5 two questions:<br /></p><blockquote>1. Give me a model of a joke.<br />2. Give me some text in the form of a joke, that is not actually a joke.</blockquote>(Bing and Google both did their basic Search Engine thing, and so aren't included here.)<br /><br />Copilot and Perplexity gave me a bog-standard "dad joke" for both questions. What was interesting was that they gave the SAME joke, word for word, as the answer to Question 1.<br /><blockquote>Why did the scarecrow win an award?<br /><br />Because he was outstanding in his field!</blockquote>The only difference was that Copilot tacked a "😄" onto the end. It also told me that this was a "light-hearted joke." In case I missed it, I suppose.<br /><br />Gemini and ChatGPT both offered a simple "Setup and Punchline" model for Question 1 with the setup being a question, and the punchline being a statement. This was, in fact, the format that all of the systems used in their jokes. (I had been expecting at least one "knock-knock" joke to make it in.) While ChatGPT offered up a vegetable-related dad joke as an example, Gemini followed up with a setup, but left the punchline blank.<br /><blockquote>Here's an empty model you can fill in:<br /><br />Setup: A man walks into a library and asks the librarian for books about paranoia.<br /><br />Punchline: ____________________</blockquote>To be sure, it seemed more like a test than an example. I suspect that coming up with a good punchline to that would prove difficult.<br /><br />The two systems' answers to Question 2 were interesting, since they seemed to presume that pretty much any two sentences in "question then statement" format qualified as being "in the form of a joke." Gemini asked if I know that 771 million people lacked access to clean drinking water, then told me they were mostly marginalized communities. ChatGPT offered up wordplay. Both of them then explained why the text wasn't a joke; Gemini explaining that it was to leave "space for reflection instead of laughter," and ChatGPT informing me of its wordplay.<br /><blockquote>Why did the computer go to the doctor?<br /><br />To get a byte checked out!<br /><br />(Note: This is not actually a joke, but rather a play on words that uses computer terminology.)</blockquote>But the scarecrow joke also strikes me as wordplay, rather than "providing a clever or unexpected resolution," which was part of ChatGPT's definition of a joke.<br /><br />Gemini is the clear winner of this round, being the LLM that seemed to have the best ability to stick with the idea of not actually giving me a joke. All of the systems surprised me with the very narrow view of humor that they offered; I don't think that any of my favorite jokes from television or stand-up qualify. But this is the thing about LLMs; since they're operating on probability, and the format they offer is one that many child-friendly jokes use, it makes sense that it's the most common, and hence most probable type in the data sets.<br /><p></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-16107318004900325982024-02-28T21:53:00.000-08:002024-02-28T21:53:43.461-08:00Awaiting<p>I was reading an article on Christian Nationalism, and this lead me to an interesting paper on <i><a href="https://digitalcommons.andrews.edu/cgi/viewcontent.cgi?article=1024&context=church-history-pubs">The Rise of the New Apostolic Reformation and Its Implications for Adventist Eschatology</a></i>. The paper is some thirty pages, so I didn't read it carefully all the way through, but it was interesting nevertheless. Reading it, I was reminded of an article that I'd read in <i>Slate</i>, back in the day, <i>Full House at King of Kings Assembly</i>. There was a quote in it, that reads as follows:</p><p></p><blockquote>There's something in the bloodstream of American Christianity that looks for, and reacts to, signs of the apocalypse. To me, it's not a great thing to herald the end of the world while I'm living here. I have kids. I want to see them grow up.</blockquote><p></p><p>While <i>Full House at King of Kings Assembly</i> was about Christian Zionism, that desire to see the end of the world (or at least the end of the present world) come as quickly as possible is still prevalent in Christian Nationalism, which also wants to bring about the return of the messiah. And this seems driven, at least in part, by a notable hostility to the world as it currently is. Which makes sense once you learn a little bit about their theology. This is, after all, a worldview that holds that "much of the world is under demon control, which includes many individuals, people groups, nations, territories, false religions and ideologies." I suspect that if I believed that literal demons were leading most of the world around me, I would be pretty hostile to it, too. Especially if I saw a number of post-Enlightenment changes, such as acceptance of homosexuality, religious pluralism, and separation of church and state, as being the result of supernatural "evil influences."</p><p>But the thought that I kept coming back to while reading the paper was: "Man. That has to suck." Because it simply can't be a pleasant experience to look at the world around oneself and understand that a good deal of what one sees as being the result of the enemies of what is just and right. The world can sometimes come across as a pretty miserable place even when one grants that everyone has good intentions and is doing their best. Seeing people as in thrall to literal forces of evil, of their own doing or not, has to make the world into a much darker and more hostile place that it would be otherwise.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-22846127680088865722024-02-25T19:36:00.000-08:002024-02-25T19:36:36.383-08:00Untimed<p>The problem with the idea that the moral failures of people in the past were the result of their negligence in understanding the self-evident moral principles that were present (although often not widely held) in their world is that it presumes that today is the end of moral history; or at least a history with any clarity. Either today's moral viewpoints become the indisputable, universal, and eternal strictures that somehow eluded the whole of humanity before us, despite having been in plain sight, or we somehow have the misfortune of encountering hidden moral truths that won't be uncovered until later generations discover them. Both of these options strike me as unlikely; accordingly, judging yesterday by the moral standards of today seems hold past generations to a standard many deny the present should meet, all in the service of the pretense that 50 years from now, the 2020s won't be found as morally wanting as many people today find the 1970s.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-74583479869426182562024-02-24T07:35:00.000-08:002024-02-24T07:35:42.814-08:00Accounting<p>The other day, I did another experiment with Generative A.I., this time around a question that I'd been looking to answer for some time.<br /><br />I asked "A.I.": "How many grocery stores are there in Washington State?"<br /><br />In keeping with the original experiment, I asked the Usual Suspects: Bing, Copilot, ChatGPT and Gemini. Alexa had been included in the original experiment, but I don't own an Alexa-enabled device and don't want the application on my phone, so it was excluded. In it's place, I added Perplexity, which I'd just heard about recently. I also asked Google, which was part of the original experiment, but had been excluded from my trials, as it's primarily a search engine. (This time, as expected, Google simply give me a list of web links. I was curious if it would present any inline answers, but it didn't.)<br /><br />When the question first occurred to me, several months ago, I'd asked Bing, and it didn't really give me an answer, but it did offer some not-particularly helpful links. This time both Bing and Copilot appeared to dodge the question. Rather than even attempt to answer the question as asked, they focused on "notable" local stores, providing what seemed like ad copy for Trader Joe's, Grocery Outlet and Haggen. The wording was different between the two, with Copilot being more verbose, but they offered (unsurprisingly) fundamentally the same answer.<br /><br />ChatGPT flat-out admitted that it didn't have the information, and offered places where I might be able to find current statistics. (But, since it's not a search engine, no direct links.)<br /><br />Perplexity gave an interesting answer, informing me of how many independent grocers and food co-ops were in the State in 2019. It also gave the number of Kroger and Albertson's stores in the state that would be sold to C&S Wholesale Grocers if the merger were to go through. But it said "The state is also home to various top grocery chains like Amazon Go Grocery, Safeway, and Haggen," so its clarity and timeliness may need to be updated.<br /><br />Gemini offered a verbose and nuanced answer, noting that how many stores qualify depends on the definition of "grocery store," and offered a range "between several hundred and a few thousand." But it was able to find a source that gave a definitive number (2,313, if you're curious), explained how they arrived at that number, and noted the potential shortcomings of that methodology. (It also took a moment to plug Google Maps and search {natch}, but that can be forgiven.)<br /><br />In the end, finding specific information that someone else hasn't already packaged up and put out there is still outside the capabilities of "A.I." The various systems mostly fell back on behaving like search engines. The information they gave me was well-presented, but there was no attempt to problem solve. I can think of a few ways that I'd come up with a count, if I could access information as quickly as modern computerized systems do.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-9051370792130455912024-02-20T19:35:00.000-08:002024-02-20T19:35:18.368-08:00Deleveraged<p></p><blockquote>"We're eager to continue working with the [Security] Council on this proposal, one that would see a temporary cease-fire as soon as practicable, based on the formula of all hostages being released," [U.S. Ambassador to the U.N. Linda Thomas-Greenfield] said after the vote. The proposal "would get aid into the hands of those Palestinians who so desperately need it," she added.<br /><a href="https://www.npr.org/2024/02/20/1232636543/un-security-council-gaza-cease-fire-vote">The U.S. has again vetoed a U.N. resolution demanding an immediate cease-fire in Gaza</a></blockquote><a href="https://www.npr.org/2024/02/20/1232636543/un-security-council-gaza-cease-fire-vote"></a>So... let me get this straight...<br /><br /><i>Israel to Hamas:</i> You release all of the hostages that you still hold, and we'll <i>temporarily</i> stop trying to kill all of you.<br /><br /><i>United States to Hamas:</i> It's a good deal. You should take it.<br /><br />There's clearly something here that I don't understand, because I'm having a hard time seeing why anyone thinks that Hamas would agree to this, as Ambassador Thomas-Greenfield laid it out. When the United States vetoed Algeria's proposed resolution calling for a humanitarian cease fire, the Ambassador argued that in order for their to be a durable peace, Hamas had to release the remaining hostages. So if the plan expects Hamas to release hostages, why not offer a durable peace, rather than a temporary cease-fire?<br /><br />The more the United States wants to be seen as instrumental in creating "peace in the Middle East," the more I suspect that the process would do better were the United States to stay out of it. Because in order to be able to work with both sides, the United States has to be seen as an honest broker, and no-one outside of the State Department or the government of Israel sees the nation that way. And with good reason. One can debate the cause, but it's pretty clear that the effect is that the United States is pushing for a settlement that is acceptable to Israel, without regard for what anyone else wants. The White House and other elements within the government make occasional noises about a "two-state solution," but that seems to be the one option that American administrations are unwilling to put any political capital behind.<br /><br />In the end, the Palestinians have no good options. They can't fight their way into a position that would allow them to deal with Israel as equals, but it seems unlikely that even if they completely gave up the fight, that they'd be granted equal status, given that current Israeli policy insists that any independent Palestine be effectively a client-state of Israel. Palestine can't win a war of attrition; attempting to do so will likely only end in their gradual extermination and/or expulsion from the land they currently hold. If this is really a state of affairs the United States wants to avoid, something needs to change. And Washington is as good a place to start changing things as any.<br /><p></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-91975286618966797392024-02-18T18:43:00.000-08:002024-02-18T18:43:36.706-08:00Snuck<p><a href="https://www.axios.com/2024/02/18/trump-sneakers-never-surrender-launch">Trump's high-top sneakers sell out hours after launch</a><br /><br />Personally, this seems more like an advertisement for Brand Trump than something that would be filed under "Politics & Policy," but maybe that's just me. Presuming that this is a legitimate news story, what I would like to know is: What are the sources? As near as I can tell, the only way the Axios reporter knows that the shoes are sold out is that they went to the website, and saw the "Sold Out" stamp on the picture of the shoes.<br /><br />But that's not proof that the shoes are sold out. Only that they've been marked as such. Granted, with the ability to buy three pairs at a time, only a few hundred people needed to make purchases to buy the 1,000 pairs reportedly available for sale.<br /><br />Still, anyone can go to a website, see a "Sold Out" stamp, and then come up with two hundred words. Axios didn't even seem to bother looking into the company that's reportedly selling the shoes, "CIC Ventures LLC." (Which, according to <a href="https://twitter.com/solusnan1/status/1759088321727271374?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Etweet">one X sleuth</a>, is headquartered at Trump International Golf Club in West Palm Beach.) Instead, they simply copy and paste a statement from the website seeking to put some distance the business and Donald Trump.<br /><br />This isn't a story, it's journalistic filler. Which I get. After all, filler content is cheaper than material that actually takes work to put together. But it seems like part of the death spiral that news outlets find themselves in.</p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-6902644131507171832024-02-15T19:55:00.000-08:002024-02-15T19:55:32.726-08:00On the Seventh DayLast week, I <a href="https://aaronmclin.blogspot.com/2024/02/days-without-number.html">replicated an experiment</a> with "A.I." conducted by a user on LinkedIn. The question was: "How many days between 2000-03-09 and 2002-09-30?"<br /><br />Basically, ChatGPT 3.5 got the answer all kinds of wrong. It had come up with workable steps to find the number, but made errors in each one, resulting in it being off by 100 days. But it showed its work, so one could see where it went off the rails.<br /><br />This morning, I asked again, and this time, ChatGPT's answer was only slightly off from the ones originally provided by Bing, Alexa, Copilot, and Gemini, answering 934 days. (By my math, it's off by 1.) Again, it gave the steps, but this time, it didn't show its work.<br /><br />I then decided to test a theory of mine about generative A.I., and asked: "How many numbers are there between 1 and 100?" This is a trick question. For starters, it's ambiguous, because I don't define "number." So the answer could be "an infinite number." I also suspected that ChatGPT wasn't going to leave out 1 and 100, because this isn't the way people normally interact with numbers like this. Sure enough: "Therefore, there are 100 numbers between 1 and 100, inclusive."<br /><br />(Just to check, I asked it: "Is 1.5 a number?" To which it responded "Yes, 1.5 is indeed a number. It is a decimal number representing the value one and a half. In mathematical terms, it's a rational number and can be expressed as the fraction 3/2.")<br /><br />Curious as to how ChatGPT parsed the word "between" I asked: "How many days are there between Sunday and Saturday?" The answer came back as: "There are 6 days between Sunday and Saturday, inclusive." So it seems that the use of "inclusive" changed, from counting both 1 <i>and</i> 100 in the number question to counting Sunday <i>or</i> Saturday in the days of the week question.<br /><br />To be sure, these are simple questions. But they point to the idea that ChatGPT, at least, because of the data that it's been trained on, tends to use language in the same inconsistent ways that people commonly use language. And this is why I suspect that prompt engineering is such an important piece of the puzzle. Getting to the "right answer" is likely going to mean constraining the system, especially if the "Temperature" and/or "Top_p" settings of a system are set to allow for more variability in the answers given.Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-20861916263264174262024-02-13T17:31:00.000-08:002024-02-13T17:31:34.297-08:00Wingman<p>I watched Microsoft's Super Bowl <a href="https://www.youtube.com/watch?v=SaCVSUbYpVc">commercial</a> for Copilot. It's a good spot, well put together with a clear message. But I'm not sure that the message conveyed is really the one that Microsoft wants to send.<br /><br />The basic format of the spot is this: it shows a number of individuals, recounting the doubts that other have about them. It focuses on one person, and her succinct response: "Watch me." Cue the inspirational music as we're shown how Copilot can empower people to achieve more.<br /><br />For the people who dream of doing big things that are hard to achieve by oneself, Copilot positions itself as something that can help. But it does that, not by making the people in the spot better at what they're already doing, but by allowing them to do things that they'd otherwise ask of others. Commercials operate by showing a need, and offering a solution, and this one is no exception. But the need it appears to be solving is that of human collaboration. At no point in the ad does the viewer see two or more people collaborating on anything. In the world that the commercial shows, there are no teams... only individuals and Copilot. I can see the appeal to people who enjoy working alone, but I suspect it will heighten anxiety for others. If one's fear is that technology will drive isolation, or that AI will render one's skills obsolete, the message of this commercial is not reassuring.<br /><br />It's the difference between speaking to the target audience and the total audience.<br /><br />In one scene, a young man asks for sign ideas for his classic truck repair shop. Copilot takes on the role of graphic designer. That scene would have been just as powerful showing Copilot helping a graphic designer to create even more ideas. Or helping "Mike" source specific information on a model of truck that he's repairing. It's A.I. as complement, rather than competitor. By showing Copilot as helping people be productive with the skills they already have, a commercial can avoid the appearance of pitting people against one another, or devaluing the knowledge and talents that they bring to the table. It speaks to the total audience.<br /><br />The primary fear that many people have of A.I. technology is that it's going to obsolesce their skills and damage their employment and/or income prospects. And, to be sure, it's going to take more than well-crafted commercials to assuage those concerns. Be that as it may, there's value in not feeding those concerns. Creating an advertisement that implies that Copilot can replace people, and expecting those people not to notice, was a misstep, if for no other reason than it implies that destroying people's livelihoods will be a known side effect of rolling out the technology.<br /><br />The disruption to labor markets of Generative A.I. systems, and possibly Artificial General Intelligence is not a matter of "if," but of "when" and "how intensely." If A.I. companies don't position the technology as something that benefits broad sections of society, as opposed to further concentrating wealth at the top of the income distribution, people will push back against the threat they perceive it to pose. Messaging that speaks to a target audience, but leaves out others, plays into that sense of threat,<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-89492007574760864632024-02-12T20:41:00.000-08:002024-02-12T20:41:00.897-08:00Missed Messaging<p>A few hours ahead of Sunday's Super Bowl game, Donald Trump posted the following to his "Truth Social" platform:<br /></p><blockquote>There's no way [Taylor Swift] could endorse Crooked Joe Biden, the worst and most corrupt President in the History of our Country, and be disloyal to the man who made her so much money.</blockquote>So far, so Donald Trump. I've already <a href="https://aaronmclin.blogspot.com/2024/02/plotted-out.html">opined</a> about the fact that I don't think that Taylor Swift is as much of a political kingmaker as Right-wing/Republican paranoia makes her out to be, so I won't belabor that point here. The point of interest this time is National Public Radio, and <a href="https://www.npr.org/2024/02/11/1230747538/donald-trump-taylor-swift-biden-endorsement-disloyal">their impression</a> of Mr. Trump's words, which they first describe as "Trump's appeal to Swift" and then later as "Trump's plea."<br /><br />It had caught my attention because it seemed like they were soft-pedaling Mr. Trump's tone. I saw nothing in Mr. Trump's words that came across as an appeal, let alone pleading. So I conjectured that maybe NPR didn't want to arouse the ire of the Make America Great Again crowd by calling him out for being aggressive towards Ms. Swift. But on re-reading, it occurred to me that maybe they just didn't know what to make of it all. NPR was framing the post as if Taylor Swift were the <i>audience</i>, rather than simply the <i>subject</i>, even though there is nothing about it that would indicate this. Donald Trump is perfectly capable of speaking to people when that is his intent. And that doesn't seem to be the intent here. If anything, given the fact that it refers to Ms. Swift in the third person, it's priming his followers to be angry with her in the even that she does decide to endorse President Biden. It's simply Donald Trump publicly reinforcing the idea that many of his supporters already have, that Taylor Swift has no business in politics.<br /><br />To be sure, he's being circumspect about it; he avoids saying any directly negative about Ms. Swift. Given his penchant for name calling, he could have easily come up with some playground derogation and attached it to her, then railed against her. But he doesn't. Rather, his tone is one of assuring his followers that Taylor Swift owes him, and therefore, their concerns that she'll turn her legions of followers against the Republicans this fall are misplaced. In other words, Donald Trump's not appealing to Ms. Swift, he's speaking to pretty much the only people he <i>ever</i> speaks to; the people who want him to be President again. It's odd that NPR seems to have missed that.<br /><p></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-38561835271223601342024-02-11T09:35:00.000-08:002024-02-11T09:35:43.224-08:00Full Page<p></p><blockquote>But Little’s decades of impunity underscore a troubling truth about the U.S. criminal justice system: It is possible to get away with murder if you kill people whose lives are already devalued by society.<br /><a href="https://getpocket.com/explore/item/the-perfect-victim"><i>The Perfect Victim</i></a></blockquote><a href="https://getpocket.com/explore/item/the-perfect-victim"></a>This is the sort of “troubling truth” that I, personally, am not troubled by. Not in the sense that I support the idea that law enforcement should be unequal, but because I understand why it is unequal, and being troubled by basic facts of life is bad for one’s mental health.<br /><br />For all that advocates for poor and minority communities, and other groups commonly referred to as “marginalized,” claim that they don’t want less attention and fewer resources to be devoted to investigating the disappearances and deaths of young, attractive, white, and upper middle class women, maintaining the levels of public interest (and thus, media coverage) and law enforcement engagement for every disappearance and death would require a level of resourcing that I suspect would quickly become unsustainable. Accordingly trade-offs have to be made. And, as with pretty much all trade-offs, those made by the criminal justice system respond to the incentives that are placed before it.<br /><br />But open talk of trade-offs tends to be avoided. The <i>Washington Post</i>’s choice of words speaks to this in the formulation of: “people whose lives are already devalued by society.” This characterizes “value” as something that starts high, and is, intentionally or not, lowered. If all people are going to be valued equally, than everyone should be valued as much as those people who are now considered most valuable. While I’m sure that there have been instances where someone has been described as “overvalued” by society, I haven’t come across one.<br /><br />Criminal justice requires trade-offs because resources are limited. There simply aren’t the personnel, nor the money to pay for them, to apply a consistently high level of effort to every disappearance or death in which foul play is known, let alone suspected. And so law-enforcement agencies put their resources where there will be the loudest complaints of under-resourcing. And that tends to be those cases with the most attention from the public and the media. But it’s not only a one-way street. Part of the reason for the differential in how society values people is a justification of who resources are expended on. When the public focuses on women who are seen as mothers, daughters, students, and contributors to their communities, that’s done in part to justify granting investigations into what happened to them a greater slice of the resource pie.<br /><br />Samuel Little, was, quite simply, bright enough to understand that, if he was going to indulge his recurring desires to choke girls and women to death, he had to be careful in who he selected.<br /><blockquote>For example, he said, “I’m not going to go over there into the White neighborhood and pick out a little teenage girl.”</blockquote>Because murdering a little teenage white girl would have created incentives for more resources to be applied to both finding her, and finding who had victimized her. I’m not a fan of that incentive structure, but being actively troubled by it would mean not sleeping at night. After all, it’s a very common phenomenon.<br /><br />In the end, the <i>Post</i> article (which, by the way, is from late 2020, before Samuel Little died in prison) is about casting law enforcement and the criminal justice system as indifferent, sometimes apparently deliberately. It’s the sort of thing that’s designed to highlight an injustice, while making it clear that the reader is absolved of any responsibility for it. That’s why it’s a “troubling truth about the U.S. criminal justice system” rather than a illustrative example of how societies choose to expend resources when hemmed in by constraints. If the United States, as a nation, wishes to avoid future instances of a serial murderer using the incentives that society creates to cover their tracks in the service of committing crime after crime, the populace as a whole is going to have to change those incentives. And to the degree that being troubled by them tend to disincentivize examining and understanding them, it acts to hasten, rather than prevent, the next series of killings.<br /><p></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-25949386868368731502024-02-10T08:36:00.000-08:002024-02-10T08:36:07.191-08:00And Statistics<p>One of the side effects of regularly listening to the 538 Politics Podcast, especially their "Good or Bad Use of Polling" segments, is that I've become better able to understand data when it's being presented in the media. Because few things are really what they're cracked up to be. Take the following story from Axios:<br /><br /><a href="https://www.axios.com/2024/02/09/marriage-wellbeing-happiness-survey">Married people are happier than single adults: survey</a><br /><br />Brought to you by the wedding-industrial complex. Not really. It's actually brought to you by a <a href="https://ifstudies.org/about/our-mission">conservative think-tank</a>, that's looking "to strengthen marriage and family life." Because headlines need to be punchy, Axios somewhat misstates things.<br /><br />Gallup, the organization that conducted the surveys, doesn't set out to directly measure happiness. What they're measuring are people's self-reported positioning on the "<a href="https://news.gallup.com/poll/122453/understanding-gallup-uses-cantril-scale.aspx">Cantril Self-Anchoring Scale</a>," which asks respondents to place themselves on a 0-10 scale of worst to best possible lives for themselves 1) in the present and 2) in five years time. Using these scores, Gallup places people into one of three buckets: Thriving, Struggling or Suffering.<br /><br />What the Institute for Family Studies is <a href="https://ifstudies.org/blog/married-people-are-living-their-best-lives">touting</a>, is the fact that the Gallup data show a correlation between being married, and attaining a "Thriving" rating in the Gallup groupings. Their point, which is to be expected, given their mission is that "In America at least, it looks like well-being ebbs and flows with marriage." The column does, however make an important caveat: "Marital status is not randomly assigned." And this is important, because the article is at pains to point out "that marital status is a stronger predictor of well-being for American adults than education, race, age, and gender." While it notes that the data can't prove a causal relationship, the piece maintains that marriage is an overall good. Which makes sense for an organization that's looking to boost marriage.<br /><br />Looking at the <a href="https://www.datawrapper.de/_/GTzhQ/">actual data</a>, what stood out for me was 2017. With the exception of divorcées, this is the high-water mark for all groups' thriving percentages. And 2018 was a bit worse for all groups. 2020 was a disaster for everyone, and only the never married cohort continued their recovery into 2023. Divorcées were on track to take over the second-place spot before the pandemic derailed everything, and they've barely recovered any ground since 2020. All groups are much lower in the rankings as of last year than they were in 2017, and married couples are on a downward trend. The Institute for Family Studies article speaks to none of this. Which is unfortunate, because there is likely to be insight there.<br /><br />Instead, it shifts, and focuses on the fact that "People living in metropolitan areas with higher rates of marriage enjoy higher subjective wellbeing." To demonstrate this, it take the metropolitan areas with the ten highest and ten lowest rates of marriage, and plots their Gallup "Thriving" rates, along with their rates of "deaths of despair," defined here as "deaths from suicide, drug or alcohol poisoning, or overdose." (As an aside, "deaths of despair" has become something of a buzzword, and like most buzzwords, it's not strictly accurate. After all, drug and alcohol poisoning/overdose deaths may also happen at parties.) But to do this, it relies on an <a href="https://www.brookings.edu/articles/we-need-to-talk-about-well-being-why-the-study-of-well-being-is-crucial-for-race-relations-and-advancing-prosperity/">analysis</a> done for The Brookings Institution. That analysis shows something that the Institute for Family Studies alludes to; the deaths of despair rates for the top and bottom ten metro areas by "Thriving" rates. Provo-Orem, Utah holds the number one spot on both lists, having the highest percentage of married households and of people who say that they are thriving. But, notably, other than that, none of the top ten metro areas for "Thriving" show up on the top ten list for marriage rates, and there is no overlap at all in the lists for the bottom 10 metros. Brookings, rather than focusing on rates of marriage when evaluating whether people are likely to self-report as thriving, points to demographics and location, noting that different ethnic groups have different "Top Ten" lists. It doesn't mention marriage at all as a factor.<br /><br />In the end, I found myself looking for a data point that wasn't there. Mainly because I don't believe that Gallup asks about it in their survey. Is there a correlation between whether people <i>want</i> to to be married or not, and whether they report themselves to be thriving? Policies designed to help people who want to be married become married would look significantly different than policies designed with marriage for everyone in mind.<br /><br />Coverage of the benefits of marriage, especially when promoted by marriage boosters, tend to view the institution as something that's good for everyone, regardless of other concerns. "Married people are happier than single adults" and "Married People Are Living Their Best Lives" being just two examples. But "People who have the relationships they want are happier than other adults" and "People Who Have the Relationships They Want Are Living Their Best Lives" are just as likely to be true.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-37541149207447017022024-02-09T20:17:00.000-08:002024-02-09T20:17:00.303-08:00Year of the Dragon<div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiS0rAfXVNH41Z7tP4Lq-MGYPuX9sGHbOWHH_WXTub6q72vnb2DxNNorZ3haxvxzYsK58IZ62TJ1uTpSMV0IpEALsVSVNkdR8_fCdhnO-_d9pM2R8A2dgHOs_wK_AT1AWFFR_ucuzUxyXcPFeoMMI287vrJcn83OZt9_-IOxElVh2QDnr4TGZtTttBNBEM/s1503/Dragons.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1002" data-original-width="1503" height="213" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiS0rAfXVNH41Z7tP4Lq-MGYPuX9sGHbOWHH_WXTub6q72vnb2DxNNorZ3haxvxzYsK58IZ62TJ1uTpSMV0IpEALsVSVNkdR8_fCdhnO-_d9pM2R8A2dgHOs_wK_AT1AWFFR_ucuzUxyXcPFeoMMI287vrJcn83OZt9_-IOxElVh2QDnr4TGZtTttBNBEM/s320/Dragons.jpg" width="320" /></a></div><p>It's been a while since I've done any figure photography, and these Dragons seemed on theme for the upcoming Yin lunisolar calendar New Year. I'm going to have to work on finding a good backdrop in my current place, especially given that my cameras default to opening the aperture, even in reasonably bright light, resulting in very shallow depth of field. To get around that, I'll need to have a place where I can set up my tripods properly. Either that, or get some reflectors to raise the light levels.</p><p><br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-4618804963644416072024-02-08T20:45:00.000-08:002024-02-08T20:45:29.468-08:00Days Without Number<p>A Product Director from Amazon <a href="https://www.linkedin.com/feed/update/urn:li:activity:7161386944814338050/">posted on LinkedIn</a> 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."<br /><br />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 <i>is</i> a search engine, after all.)<br /><br />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."<br /><br />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.<br /><br />ChatGPT (3.5) went to town, giving me the entire process by which it arrived at its answer:<br /></p><blockquote>To calculate the number of days between March 9, 2000, and September 30, 2002, we can use the following steps:<br /><br />1. Calculate the number of days in full years between the two dates.<br />2. Add the number of days from March 9, 2000, to the end of 2000.<br />3. Add the number of days from January 1, 2002, to September 30, 2002.<br /><br />Here's the calculation:<br /><br />Full years between 2000 and 2002: 2001 and 2002 (2 years).<br />Total days for full years: 2 years * 365 days/year = 730 days.<br /><br />Days from March 9, 2000, to the end of 2000:<br />March 9 to March 31 = 31 - 9 + 1 = 23 days.<br /><br />Days from January 1, 2002, to September 30, 2002:<br />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.<br /><br />Now, summing up all the days:<br /><br />730 (full years) + 23 (days in 2000) + 284 (days in 2002) = 1037 days.<br /><br />So, there are 1037 days between March 9, 2000, and September 30, 2002.</blockquote>One of these answers is not like the others... ChatGPT was <i>way</i> off.<br /><br />Which kind of gets to the thing with generative "A.I." The artifice is there, and fairly impressive. But <a href="https://aaronmclin.blogspot.com/2023/10/nothing-new.html">they aren't intelligent</a> 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 <i>that</i> 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 <i>why</i> 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.<br /><br />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.<br /><p></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-30782806257619878252024-02-07T19:15:00.000-08:002024-02-07T19:15:07.991-08:00EscalationsCharles 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.<p>Aside:</p><p>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 if context thrown in for added clarity.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-25441220409409802672024-02-06T20:42:00.000-08:002024-02-06T20:42:09.997-08:00Booted<p><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto;"><tbody><tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjNCOY1uqY5HLH1kvxt5_u8qq5DI6x2xVdak7G78PvB2dy0RjVizVNMHek5TjfSEmMMp85hib2SdsvoN9-nE0akcKOxsue5cEdtqbtO0CIgRIayEuX1WW8bldoaTtWyaJkdWveswQt430fWmz3rOItCUCK5e-6RvsN5ZS56Sz4PurknjnleZ8dSKxz70gM/s1503/DasBoot.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" data-original-height="1002" data-original-width="1503" height="213" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjNCOY1uqY5HLH1kvxt5_u8qq5DI6x2xVdak7G78PvB2dy0RjVizVNMHek5TjfSEmMMp85hib2SdsvoN9-nE0akcKOxsue5cEdtqbtO0CIgRIayEuX1WW8bldoaTtWyaJkdWveswQt430fWmz3rOItCUCK5e-6RvsN5ZS56Sz4PurknjnleZ8dSKxz70gM/s320/DasBoot.jpg" width="320" /></a></td></tr><tr><td class="tr-caption" style="text-align: center;">A child's boot, abandoned in a parking lot.<br /></td></tr></tbody></table>The thing about found object photography is the knowledge that behind every object there is a story; one that will likely never be known.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-5282352491752866662024-02-03T18:41:00.000-08:002024-02-12T17:51:02.561-08:00Plotted Out<p>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.<br /><br />But it's also worth noting that alleged conspiracies offer insight into the anxieties of the people who hold them. Consider the (relatively) <a href="https://www.axios.com/2024/01/30/taylor-swift-maga-biden-trump-2024">new</a> conspiracy <a href="https://theweek.com/politics/right-obsessed-taylor-swift-election-conspiracy-theory">theory</a> concerning singing star <a href="https://www.npr.org/2024/02/01/1228373511/heres-why-conspiracy-theories-about-taylor-swift-and-the-super-bowl-are-spreadin">Taylor Swift</a>. You've likely <a href="https://www.politico.com/news/magazine/2024/01/31/taylor-swift-conspiracy-theories-gop-00138817">heard of it</a>, it's <a href="https://www.cbsnews.com/news/taylor-swift-travis-kelce-conspiracy-theories-chiefs-super-bowl/">being covered</a> all <a href="https://www.bostonglobe.com/2024/01/31/nation/taylor-swift-super-bowl-election-conspiracies-right-wing-media/">over</a> the <a href="https://www.wired.com/story/taylor-swift-psyop-conspiracy/">place</a>. 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 <a href="https://twitter.com/RonFilipkowski/status/1752314183700316305?s=20">at it again</a>, in order to ensure that President Biden wins another illegitimate term in the White House.<br /><br />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.<br /><br />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.<br /><br />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 <a href="https://twitter.com/VivekGRamaswamy/status/1751944730114498838">talk the talk</a>. 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.<br /><br />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.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0tag:blogger.com,1999:blog-1396682199178122530.post-48457726418095600412024-02-02T21:05:00.000-08:002024-02-02T21:05:39.142-08:00But Wait, There's More<p>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 <a href="https://aaronmclin.blogspot.com/2023/06/most-of-people-most-of-time.html">monetizing scaring people</a>, 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.</p><p>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 "<a href="https://komonews.com/news/local/seattle-shooting-udistrict-university-uw-north-shot-gun-homeless-crisis-homelessness-man-bicycle-outreach-wsdot-camp-encampment-i5-off-ramp-police-washington-tent-bike-kcrha-transportation#">Man shot after walking past homeless encampment</a>," multiple times in an evening (it seems that they run these stories on repeat), it starts to make a certain level of sense.</p><p>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.</p><p>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.<br /></p>Aaronhttp://www.blogger.com/profile/03128359303310478673noreply@blogger.com0