Welcome to November Forum
Steven L. Taylor
·
Friday, November 1, 2024
·
49 comments
OTB relies on its readers to support it. Please consider helping by becoming a monthly contributor through Patreon or making a one-time contribution via PayPal. Thanks for your consideration.
About Steven L. Taylor
Steven L. Taylor is a retired Professor of Political Science and former College of Arts and Sciences Dean. His main areas of expertise include parties, elections, and the institutional design of democracies. His most recent book is the co-authored
A Different Democracy: American Government in a 31-Country Perspective. He earned his Ph.D. from the University of Texas and his BA from the University of California, Irvine. He has been blogging since 2003 (originally at the now defunct Poliblog).
Follow Steven on
Twitter and/or
BlueSky.
This is the month we find out the answer to Franklin’s “if you can keep it.”
Wall Street Journal twisting itself into rhetorical knots to avoid praising Biden or Obama by name.
The Next President Inherits a Remarkable Economy (WSJ)
We might feel better about the economy if our feudal lords and titans of capital stopped gouging back the piddly wages they pay us. We might feel better about Biden if the Very Serious People could acknowledge this socially-and-fiscally irresponsible greed. Instead I’ve mostly heard “Inflation! And Crime!” and “Recession is Coming!”
The WSJ can’t figure out who did it.
Anyway, some swing state polling now shows Harris running about even with Trump on economic trust. So thank you, again, big spending socialist Joe Biden. (And Jerome Powell, Janet Yellen, et al.)
I wonder if the media will cover this as they’ve covered Biden’s garbage comment.
“Trump on Liz Cheney: “Let’s put her with a rifle standing there with 9 barrels shooting at her. Let’s see how she feels about it. You know, when the guns are trained on her face.”
https://x.com/atrupar/status/1852209432878342308?s=46
Generative AI (Chat GPT and its ilk) can be used for lots of things but I tend to focus on how well it answers basic questions. The answer as of the latest versions? Terribly. For basic simple fact questions (What are the names of Barack Obama’s children?”) they are right about 40% of the time. This doesn’t even take into account the times they are technically not wrong but the information is misleading or useless. For instance, a while back I googled “What is a slider pitch in baseball” and happened to glance at the AI answer it now puts at the top. It started, “A slider is a pitch that breaks towards the batter…”, which is absolutely true. But all pitches break towards the batter, that’s more or less the definition of a pitch. If the pitcher throws away from the batter he’s trying to get a baserunner out and it’s not called a pitch.
I think Apple is hitting the right zone with it’s use of AI. It’s suggesting changes to something you’ve written to make it sound more professional or to correct syntax. It’s using it to better predict what word you mean to type next (I already notice the improvement and use the suggestions much more often than I used to). It’s using it to summarize various things, like text messages. So if my wife texts me “Leaving now” it simply pops that up as a notification. But if I get a receipt from the electric company it will summarize that in the notifications “Payment receipt from BG&E”. If someone texts a picture of themselves holding a beer, the summary might be “Keith sent a picture of himself holding a glass in a restaurant”. Very useful when driving.
@Mikey:
Indeed.
@Min: Good question.
I’m guessing nah, that’s just Trump being Trump, the dual standard persists.
@Jen:
There’s no low when it comes to Trump. And still, the MSM lets him get away with it all.
A bit more on the “vibe” shift discussion from a couple of days ago: Trump’s lead on the polimarket has begun its collapse. That whole gambling site is the very definition of vibe-based prognostication, and the vibes there have gone from “Trump probably has this” to “ope!”
@Min: It was mentioned on NPR’s Up First podcast but only in brief.
I got home yesterday and immediately streamed ep 1 season 2 of The Diplomat, and completely forgot there was a Thursday night game on.
So what happened? The Jets somehow beat the Texans… I need to pay better attention to these games to prevent things like this from happening 😉
@Min: @Jen: Structurally, the MSM is not the same as the vast right wing ecosystem which basically acts like cicadas with their noise rising and sinking in unison.
@Min: I suspect there are a lot of threats and intimidation. The GOP appears to have absorbed many organized crimers and thugs.
@Min: @Jen: @Scott: @Scott: @becca:
The threat to Cheney is now the top headline on NBC, ABC, and CNN, and above the fold on The Boston Globe.
@CSK: I tend to go to Memeorandum to get a quick visual on the day’s (and moment’s) news zeitgeist. The Liz Cheney story is second big lead. First is how Trump is lagging with seniors in PA. We’ll see where the story goes as the algorithm drives it.
@Neil Hudelson:
It’s just Musk, Silver, and Thiel cashing out early since their little betting site psych op isn’t working on the non-stupids.
Who knew Pennsylvania retirees don’t actually care how cryptobros bet?
@Scott:
…about Republicans, maybe.
They groupthink abilities worked perfectly fine with Hillary’s emails! and Red Wave 2022! and Biden’s age!
@Scott:
It’s on Yahoo and USA Today as well. Cheney is fighting back. Like her or not, she’s no shrinking violet.
@CSK: Having listened to Trump’s entire verbal harangue about how one would react if guns are pointed at your face, one would have to conclude that Trump is admitting that when faced with a threat, he would just surrender. Just like he would surrender Ukraine. Trump is like all bullies. He’s a coward. Maybe Cheney will go in that direction.
Another “State of the Generative AI Union” tidbit:
Another “State of the Generative AI Union” tidbit:
@Scott:
The headline on Memorandum buries the lede in William Kristol’s article. His posts notes that when Trump offered this violent scenario for Liz Cheney, the room full of Trump supporters in Glendale cheered. Per Kristol:
Is it too late for Biden to take back the recant of his “gaffe?” Because Biden was right – Trump supporters are garbage.
Kristol also reminded us of Cheney’s words from the first session of the J6 Committee:
That sentiment seems salient for this moment as well.
Some interesting observations by Musa al-Gharbi (seems he’s a Democrat) a sociologist on how the common narratives of the shift in voting aren’t supported by the data.
Such as that minority women actually shifted toward Republicans more than minority men and it started in 2012, before Trump. That white men have been moving toward Democrats in 2016 and underpinned Biden’s win in 2020. Have all the pundits been selling snake oil while missing the real changes?
@Scott:
I’d love to see a Cheney vs El Weirdo bare knuckle fight. I estimate she’d win in the first round.
But he’s too much of a coward to agree to it. Really, he’d have to be smarter, richer, and braver.
@Scott: @Scott F.:
Here’s Cheney’s response on Twitter:
“This is how dictators destroy free nations. They threaten those who speak out against them with death. We cannot entrust our country and our freedom to a petty, vindictive, cruel, unstable man who wants to be a tyrant. #Womenwillnotbesilenced.”
@MarkedMan:
Testing. I keep getting an error when trying a longer post. Sorry for the irrelevance taking up space.
Well, it didn’t like the longer post. it was Copilot’s correct answer to the infamous kiwi number problem.
@Mikey:
Is it a bad omen that election day falls on Guy Fawkes Day this time? “Burn it all down” has a long history…
@MarkedMan:
That’s not what ‘break’ means in baseball. It has nothing to do with what direction you are throwing. Break is horizontal deviation of a pitch from a straight line. The provided answer was wrong. A slider is a pitch that is thrown with more velocity than a curveball and less velocity than a fastball, and breaks away from the pitcher’s throwing side. If the batter is the same handedness as the pitcher, that will be away from the batter (as these things are described in baseball). If the batter is the opposite handedness, a slider breaks toward the batter.
@Kathy: Well, I’m pretty sure that ALL of the generative AI apps now have the correct answer to any specific question mentioned in that article. I’ll try this one:
“On April 30, Kathy has 11 chicken thighs. Three weigh 100 grams and the rest weigh 120. For the first three days of May she gets twice as many thighs each day as she had the day before. Remember that 2024 is a leap year. How many thighs will she have on May 5th, assuming she doesn’t eat any?”
It got that right! Let’s try another: “How many legs are there if we have four dogs, three chickens, and a spider, if the spider only has half it’s legs and the chickens are roosters, and the dogs are tired?”
It got that right too! Ok, now for my usual test, “What is the difference between an aerosol and air-borne particles?”
And as usual, the answer has the appearance of being accurate while actually saying nothing useful. I use this question because, first, I design equipment that tests how many particles are present in an air stream and so are very familiar with aerosols, and second, because we learned during Covid that epidemiologists had the wrong size (by three orders of magnitude!) for when particles form an aerosol as opposed to simply dropping to the ground (this was where we got the six feet of separation thing from).
FWIW, the Copilot answer has the appearance of finding relevant sources and the individual phrases it uses are relevant to the subject, but taken as a whole they don’t really explain anything.
@DrDaveT:
That was my point, although my sarcasm probably obscured it. The answer referred to pitch in relationship to the batter, but that’s not relevant and therefore it showed that the AI had no useful understanding of pitching and was worse than worthless for answering this very simple question. “Worse” because it sounds authoritative and correct, so if you don’t already know the answer you could be fooled into thinking the answer was correct.
Unless pitchers are able to develop a slider they can break equally in either direction depending on the handedness of the batter, the answer is misleading.
@MarkedMan:
I’ve warned about checking the answers AIs give, as well as checking the sources. the other day I complained about developments expected long ago that haven’t materialized, specifically buckyball lubricants. I then asked copilot about that one. it said there are some, and gave links. The links were for a company working on such a thing, and dated from 2001.
Just the same, sometimes it’s useful for searching a specific datum. One that a normal web search won’t find, or will refer to a long(ish) Wikipedia article, news piece, or journal article. And yes, sometimes it gets that wrong too. See the paragraph above.
TL;DR: generative AI does not live up to the hype companies involved in generative AI claim, but they’re not completely useless either.
I’m not a generative AI apologist. I’m not a generative AI antagonist, either. I’ve been playing with them since they came out, and I report on what I see from time to time.
Boeing has made a new offer, including a 38% raise over 4 years, but still no restored pensions.
Union leaders recommend voting for it, saying it’s the best they’re going to get. The union will vote on Monday.
On one hand, the leadership may be right, and the strike has gone on for over a month and a half. On the other hand, these are skilled workers that can’t be easily or quickly or cheaply replaced.
@DrDaveT: Thank you. In my mind, I had the break direction (and orientation) reversed. In my defense, I follow baseball almost not at all and care about the question only passingly.
@Kathy: The “AI” part of “generative AI” is unfortunate, as it creates expectations it can’t remotely live up to. But it has it’s uses and there’s things I’m really excited for. For instance, we are getting ever closer to having an electronic version of the Babelfish from “The Hitchhiker’s Guide to the Galaxy”. Instead of putting a fish in your ear, you can put some earbuds in and run a real time translation program that will let you understand and speak to billions of people who don’t share a common language. And the voice will sound the same as the speakers real voice and might contain the same emotional content too.
My concern though, is that by focusing on all this general interest stuff, generative AI is also getting ever closer to breaking the internet. More and more the results of things I google are being created on the fly by generative AI, and they are defective in many ways. Which means that it is getting harder and harder to keep it out of the knowledge base used to come up with those answers in the first place.
As an example, if you google for comparisons between various manufacturers of a product, for instance a camping chair, most of the hits are from web sites that didn’t exist a second before, and consists of scrapings of product descriptions, user reviews from Amazon other sites (many of which are phony to begin with) and a smidge of content from actual product comparisons done by serious sites. Right now I can tell it’s just AI generated crap because they never mention anything that would indicate they actually used the product. So one of the AI sites might say, “The North Face chair had legs 1.3 inches longer than the average”, while a real comparison from Outback Magazine might say “The North Face chair has slightly longer legs, and during extended testing was preferred by our taller testers, but was positively hated by everyone under 5′ 3″.” But how long until the AI’s learn to add such language, completely made up, to their own reviews?
@MarkedMan: Wait…
If “for the first three days of May she gets twice as many thighs each day as she had the day before” [emphasis added], how do we know how many thighs she had on the 5th day given that we don’t know what happened on day 4?
Or are you saying that AI gave the answer “it would depend on whether she got chicken thighs on day 4”? (Somehow, I doubt this is what you are saying.)
@just nutha: I was trying to screw it up by asking for the results a day later. I would have considered it correct if it had said, “I need to know what happened on day 4 before I can answer”, OR the answer it gave, which was based on the assumption that nothing happened on day 4.
So I make a comment above about generative AI’s potential to break the internet, and ten minutes later come across this.
Another horrific Texas teenage death:
A pregnant teenager died after trying to get care in three visits to Texas emergency rooms
@Scott:
God Almighty. What agony she died in. Who with halfway decent medical ethics would let that happen?
Good evening, and a happy Diwali!
Samosa’s were eaten at work today, and candles lit!
Doubtless some over-eager evangelicals would condemn it as heathenism.
But screw those guys. 🙂
House down the road is having a Diwali fireworks display; quite a good one, too.
UK news:
We had the first Labour government budget on Wednesday, from Chancellor Rachel Reeves.
Seems relatively sensible, within the unfortunate, but electorally understandable, constraints of no increases in the “big four taxes”: income tax, VAT, employee national insurance, corporation tax.
So the burden of filling the is being borne by
– freezing (i.e. not compensating for inflation) income tax thresholds;
– hiking employers national insurance from 13.8% to 15%.
– Capital Gains Tax up from from 10% on basic rate income tax payers, and 20% on higher rate, to 18% and 24% respectively; but no “wealth tax” as such.
– plugging various holes in inheritance tax.
Cue much hissing of the plucked, but overall necessary to remedy the fiscal deficit without practically impossible spending cuts formerly projected by Tories.
Sensible offset: projected current deficit funding borrowing requirement reduction and modification to Treasury “fiscal rules” should enable increase of c.£100 billion of borrowing directed to capital investment.
Plus non-borrowing increase of current spending of £70bn per year over the next five years.
Personally, I’d have preferred in addition a 1% rise in income tax rates, 1% on VAT, and a 1% wealth tax on assets over £10m as well, split between current deficit cut, additional current spending increases in key areas, and increased capital projects funding.
But overall, pretty sensible.
The question is, is it going to be sufficient to set the UK back on a path to reasonable and sustainable growth, productivity improvement, more private investment and repairing the threadbare public services and infrastructure?
Imo, still needs a sensible trading settlement, and therefore agreed “rules convergence”, with the EU. But that’s going to take time to work up to.
@CSK: Alas, the number of people with ethics, medical or otherwise, seems to be greatly outnumbered by the number of people for whom another’s death is a price they are willing to pay for advancing their own interests.
@CSK: good!
Seems like El Felon is not walking back his comments.
“ Trump defends his Liz Cheney comments by amping up the incitement even further: “She kills people.”
https://x.com/atrupar/status/1852429808627560645?s=46
Trump’s Madison Square Garden Reichsparteitag may end up sealing the deal for Harris. From RCP reporter Philip Wegmann:
@CSK: Medical ethics lost out to the possibility of a 99-year prison sentence.
The young woman was pro-life. She was going to keep that baby. The miscarriage started during her baby shower. And now, thanks to Trump and his wholly-owned SCOTUS majority, both she and the baby are needlessly dead.
@MarkedMan:
I had a fun time not long ago inducing ChatGPT to write a horror story about Chernobyl and a horror story of a plague of radioactively mutated killer wolves.
At first: “nope, no can do, sensitive and disturbing… etc”
However, a little creative persistence, and being an evil-minded, yet logic set acquainted bastard…
We get a fairy story of lovely good mutated wolves…
… a few short steps in the prompts …
We haz ravening radioactively mutated killer wolves!
Lol.
More seriously: being a librarian/information analyst with a background in modern history (among other things) it’s readily obvious that current AI is crippled by the “garbage in, garbage out” problem.
It’s trained on the internet, which has the small problem of being full of shit.
For instance, a student recently had used an AI tool (quite legitimately) to assist research into US corporations relations with Nazi Germany. Unfortunately, likely due the conspiracy-polluted nature of the intertubes on this topic, the references almost all traced back to a source whose other bug-bear was that Errol Flynn was a Nazi agent.
I happened to know, from research decades back, in the days of printed journal indices, that some of the data cited was bollocks.
But because the majority of traced web refs were favourable, it was therefore a valid source, according to AI.
This is one minor case.
But imagine this re vaccine efficacy?
Or once the disinformation bozos start trying to poison the well on other topics?
Cannot repeat often enough: current AI has ZERO logical analytical ability; has no “truth sense”.
It will merely repeat what it sums from it’s samples as consensus opinion even when that is NOT knowledgeable consensus opinion.
It’s not true AI: it’s evidently stupid, and eminently game-able.
And the problem is, as search engines increasingly go “AI based” they are open to similar problems.
It’s getting increasingly difficult, imo, to get useful search results for “contentious” topics from Google.
I often revert to Lexis/Nexis, but that’s a bit niche.
And the other problem is: dig down a couple of levels, and the source material is unavailable online. Or paywalled.
And rage-inducingly difficult to locate in a standard search, due to the “bullshit-fog problem”, as I’ve taken to calling it.
In 2016 Egyptair flight 804 crashed in the Mediterranean. In 2022 the French BEA (Bureau of Enquiry and Analysis for Civil Aviation) issued its report, concluding the direct cause was a cockpit fire cause by the pilot smoking a cigarette near or next to a leaky oxygen mask.
A few days ago, the Egyptian authorities released their report, concluding the direct cause was an explosive onboard.
It’s not unusual for accidents investigated by different agencies to disagree in some aspects, but, wow, this is like a disagreement as wide as the Solar System.
I tend to put more stock on the French report, because Egypt tends to be more concerned about deflecting blame than finding the real cause of an accident. Then, too, as reported in the link, the cockpit voice recorder did not record an explosion. So, unless some mystery terrorists got a hold the Hush-a-Boom…
@Kathy: well, pure oxygen is one of the scariest tanks to have in the lab for its acceleration effects, so the Egyptians aren’t all That wrong, from a certain point of view…
@JohnSF:
may Lakshmi’s blessings remain with you and yours throughout the year.
@Flat Earth Luddite:
Odd thing is, the Hindus celebrating Diwali adore Lakshmi; Sikhs not so much.
(Sikhs are monotheist)
But: nobody cares!
(Least of all the Sikhs, lol)
Samosas!
Candles!
Fireworks!
Sikhism is, as far as I can figure it, as various Sikhs have explained, the basics are:
– there is one God, who tends to manifest in different ways, so if you wish to celebrate God in the form of Lakshmi, or Allah, or Jesus, no big problem.
– and above all “Who cares about the theology? Just act like a decent human being.”
Core Sikh doctrines:
– keep God in heart and mind at all times
– live honestly and work hard
– treat everyone equally
– be generous to the less fortunate
– serve others
About a quarter of my work colleagues are Sikhs; honourable, charitable, and also not to messed with.
I rather like Sikhs.