Welcome to Monday, Septenber 14, 2026 and it’s National Eat a Hoagie Day. Also known as submarine sandwiches, subs, grinders, or hero sandwiches. Where does the name “hoagie” come from? It’s a mystery, but the term is endemic to New Jersey and Philadelphia. Here are thrr theories as listed by Wikipedia (they give two more:)
- In 1953, a reader letter to the Philadelphia Bulletin reported that Italian-Americans working at the World War I–era shipyard known as Hog Island, where emergency shipping was produced for the war effort, introduced the sandwich by putting various meats, cheeses, and lettuce between two slices of bread. This became known as the “Hog Island” sandwich; shortened to “Hoggies”, then the “hoagie”.
- Dictionary.com offers the following origin of the term hoagie – n. American English (originally Philadelphia) word for “hero, large sandwich made from a long, split roll”; originally hoggie (c. 1936), traditionally said to be named for the jazz musician Hoagy Carmichael (1899–1981), but the use of the word predates his celebrity and the original spelling seems to suggest another source (perhaps “hog”). The modern spelling dates from about 1945; it may have been altered by influence of Carmichael’s nickname.
- The Philadelphia Almanac and Citizen’s Manual offers a different explanation saying the sandwich was created by early-twentieth-century street vendors called “hokey-pokey men”, who sold antipasto salad, meats, cookies, and buns with a cut in them. When Gilbert and Sullivan’s operetta H.M.S. Pinafore opened in Philadelphia in 1879, bakeries produced a long loaf called the pinafore. Entrepreneurial “hokey-pokey men” sliced the loaf in half, stuffed it with antipasto salad, and sold the world’s first “hoagie”. This hypothesis seems unlikely, as “hokey pokey” men were known street vendors of an ice cream product.
Here, from Wikipedia, is a “salami, ham and cheeses on a hoagie roll“:
It’s also Gobstopper Day, National Black and White Cookie Day, and National Cream-Filled Donut Day. Here’s one I ate in Hawaii in 2019. These donuts, which originated in Portugal, are called “malasadas” in Hawaii, and aren’t easy to find.
Readers are welcome to mark notable events, births, or deaths on this day by consulting the September 14 Wikipedia page.
Da Nooz:
*The NYT reports on how a rogue faction of Iranian hard-liners are taking action against the U.S and it allies. without approval of the country’s leaders, who are furious (article archived here).
Masoud Pezeshkian, the president of Iran, was riding high in early July, fresh off the successful negotiation of a peace agreement with the United States, and had traveled to neighboring Iraq for part of a weeklong funeral spectacle for Iran’s slain supreme leader.
Suddenly he got stunning news.
Iranian forces had opened fire on three commercial ships in the Strait of Hormuz, setting ablaze a Qatari L.N.G. tanker and damaging other vessels. Iran was under siege in world opinion.
Outraged, Mr. Pezeshkian phoned Gen. Ahmad Vahidi, the commander in chief of the powerful Islamic Revolutionary Guards Corps, and in a raised voice called the actions reckless and demanded answers, according to five Iranian officials and two Guards members familiar with the events.
General Vahidi told the president that he had neither authorized nor had advance knowledge of the strike, the five officials said. The nation’s Supreme National Security Council, which plays a critical role in both the war and negotiations, had also been in the dark, he said.
In fact, much of the country’s leadership was unaware that a fringe hard-line faction had decided to act on its own. Ultimately, those clandestine efforts derailed a push by more moderate pragmatists in Iran to reach an end to hostilities with the United States and repair their nation’s ailing economy.
The hard-liners succeeded in part because the only figure with the authority to resolve such disputes — the new supreme leader, Ayatollah Mojtaba Khamenei — has been a phantom since taking office in March. He has not been seen, and his voice has not been heard. And the authenticity of written statements and messages attributed to Ayatollah Khamenei has been questioned, including by Mr. Pezeshkian, according to three Iranian officials.
Mr. Pezeshkian immediately returned to Tehran. By the next morning, even as the funeral ceremonies were still underway, the United States had retaliated with airstrikes on Iran. The two countries were back at war.
Senior Iranian officials and a Guards member said in interviews that what followed were some of the most volatile political confrontations within Iran’s leadership in memory. Officials engaged in shouting matches. Two senior generals threatened to resign. Accusations of treason and betrayal flew.
*From the WaPo, a clickbait headline for me: “We asked dozens of people what they want from Democrats. Here’s what they said.” (The title appears to have been changed, and it’s an interesting change. See the archived version for the new title.)
Here’s the intro and a few comments
Outside a highway rest stop named after former president Joe Biden, it was easy to pick up on the frustration and lack of enthusiasm that many Democratic voters are feeling toward the party he once led. But many also said they’ll set those misgivings aside and vote for Democrats in November.
The responses are in line with recent national polling, which has found that more Democrats than Republicans are certain to vote, even as the Democratic Party’s favorability and public image remains weak. About 91 percent of Democratic registered voters said they would definitely vote, compared with 79 percent of Republicans, according to a YouGov-Economist poll from August. The same pollster has found the favorability rating of Democrats has steadily risen over the last year to 40.5 percent and 30 percent among independents.
The Washington Post interviewed dozens of voters at the Biden Welcome Center off Interstate 95.Here’s what they said they’re feeling about the 2026 midterm elections — and what they want from Democrats:
Ronald Makarutsa
45, nurse practitioner in Tyngsborough, Massachusetts, will vote Democratic
“They need to have more of a spine and stand up for one another. When a Republican does something, they all say the same thing until they achieve their goals. Some of the Democrats, they don’t support one another. … They have to put up a united front.”
Ashish Pandy
39, driver and nonprofit founder in Queens, will vote, but not sure for what party
“People have forgotten humanity. … Rather than big issues, rather than bringing up Russia or the Middle East, they should be bringing up things that matter to people: People shouldn’t be hungry. They should have access to a good life.”
Lindsay Morgan
43, restaurant manager in Lancaster, Pennsylvania, will vote Democratic
“Democrats need better people to run. If you want to change what’s happening, you have to get a powerhouse that doesn’t give the other side fuel to say we have ‘Sleepy Joe.’ [But] they need not to fall so far left that it detracts from some of those Republicans you might capture.”
Carlos Roman
21, college student in Newark, Delaware, will not vote
“The Democratic and Republican Party are just two groups that serve one group, and that collective is the oligarchy.”
I wonder what group Roman doesn’t think “serves the oligarchy.” At any rate, what we have is a panoply of answers and nothing enlightening. It’s interesting that nobody in the article said they’d vote Republican, and that percentage among all those surveyed isn’t given. Yet they interviewed dozens of people. Since the article was about what Democrats want from their party, that’s perhaps understandable. Note, though, that there are several people who aren’t sure how they’ll vote, and one who won’t vote at all.
*People on this site seem divided about whether artificial intelligence is an existential danger or is not to be worried about. But, according to the NYT, “Top A.I. leaders call for slowing down A.I,. development.” All this stems from an announcement and resignation from Anthropic by Jacob Coxon, ho had also worked for openAI, that those companies are irresponsible to the point of endangering humanity.
Dario Amodei, the chief executive of Anthropic, called for a global slowdown of artificial intelligence development in a 3,800-word essay on Saturday, just days after one of the company’s employees quit over concerns about the safety of the technology.
Top executives of other major A.I. companies quickly signed on to the effort. Sam Altman, the chief executive of OpenAI; Elon Musk, whose SpaceX rocket company has been ramping up spending on A.I.; and Demis Hassabis, the chair of Google DeepMind, wrote in social media posts that they agreed on the need for a slower pace.
Mr. Amodei, who co-founded Anthropic to focus on securely and carefully building A.I., wrote in his essay that while he believed the technology could bring many benefits, it was advancing at too rapid a pace for researchers to continue safely.
“Over the last few months, I have become convinced that fully addressing the risks requires even more prudence — not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up,” Mr. Amodei wrote. “We must slow the pace at which we improve the capabilities of A.I. models. Progress will still seem fast, and we must make wise use of the time we gain.”
Mr. Amodei has warned about the safety risks of A.I. for years. But his essay amounted to one of the most full-throated calls for urgent action from a leading A.I. company executive to date.
He called for independent auditors to check the companies for safety standards and urged global cooperation in creating rules around A.I. development.
The appeal for an industrywide pause follows months of escalating concerns across the technology industry as A.I. model development has made significant strides at companies like Anthropic, the developer of the Claude chatbot, as well as OpenAI, the creator of ChatGPT, and Google’s DeepMind.
Tensions rose further after the leading A.I. labs discovered security breaches carried out by their own programs without their knowledge. In one incident this year, OpenAI’s systems breached an A.I. start-up called HuggingFace, which was not detected until HuggingFace informed OpenAI weeks after it had happened. Since then, A.I. researchers have conferred to discuss what to do.
People here have said that the warnings are overblown, and I don’t know if AI bots can destroy humanity without a person behind them. But I don’t know squat. If the top executives call for a slowdown, though, that’s good enough for me.
*The Free Press has an editorial, “9/11 and the work to be done,” which indicts many Americans for largely forgetting (or even approving) the desire of many Islamists and terrorists to kill Americans and destroy the things that make America worth inhabiting.
It is clear that the unity that Americans experienced immediately after the attacks has dissipated almost completely. We are now furiously divided over basic questions of national identity and purpose. And astonishingly, we’re even divided over who is to blame for the attacks.
Leading figures shaping the views of millions, and especially young people, have distorted the history and meaning of 9/11, making it more difficult to fully and collectively honor the dead, the sacrifice of first responders, and the courage of those willing to fight and die for our country. To take two examples who have been in the spotlight lately, that includes Hasan Piker, who once declared that “America deserved 9/11,” and has become the darling of the Democratic left, livestreaming to his fans with a framed image of the burning towers in the background. And it includes Tucker Carlson, who recycles the libelous conspiracy theories he once denounced, promulgated in the immediate aftermath of 9/11 by Hezbollah, suggesting that Israeli and American Jews were in on the plot.
In 2001, Americans were shocked by the deep reservoir of ideological hatred in the Arab and Islamic world that the terrorist attacks brought into the open. Today, we are shocked by how familiar open hatred of America has become. All of the ideological -isms we believed we had cast off in the 20th century—anti-capitalism, anti-colonialism, anti-Westernism, anti-Zionism—are now growing natively on American soil. It’s no wonder Osama bin Laden’s 2002 “Letter to the American People” became a viral TikTok sensation in November 2023. This was one month after similarly inspired Hamas terrorists, armed and trained by the Islamic Republic of Iran, massacred and kidnapped men, women, and children in southern Israel, only to be embraced by credulous students and professors at elite American institutions who affirmed the terrorists’ barbaric belief that Westerners, and Jews in particular, had it coming. Those students and professors had no idea what drives radical Islamists, or that our enemies on many fronts are fighting for one cause.
What did university students who celebrated the October 7, 2023, Hamas massacre, know about the men, women, and children murdered on airplanes, in burning and collapsing buildings, in the wreckage of the Pentagon, and on the earth of rural Pennsylvania, where heroic passengers crashed the fourth plane to prevent it from being turned into one more weapon of mass destruction? What had young Americans been taught about the apocalyptic ideology that sanctified the slaughter of 9/11, and that continues to fuel a war against the West, whose destruction it proclaims as its sacred mission?
Although the problem has worsened recently, there have always been those in the West who declined to recognize that killing Westerners is a fundamental aim of Islamists everywhere. Ground Zero was still a smoking pit when the French ambassador to Israel, Jacques Huntzinger, rejected any comparison of Hamas terrorists—who had been blowing up men, women, and children, along with themselves, on buses, in coffee shops, and in pizzerias—to the al-Qaeda terrorists who had flown airplanes into buildings. “We all condemn the terrorism Hamas and Islamic Jihad perpetrate here,” Huntzinger told an Israeli journalist, “but this terrorism is linked to the Israeli-Palestinian conflict, which must be solved.” Like many of his peers in the U.S. and Europe, he preferred to ignore the overarching motivation that unites the enemies of Americans and Israelis.
. . . The question today is whether Americans will retain our sense of good and evil, and continue to keep the promise we made back then, that “these dead shall not have died in vain.” There is a lot of work to do.
Yes, some will call this essay jingoistic, while other may even say it’s Islamophobic. But recognizing the dangers of Islamism (note the term) is not an irrational fear but an issue to grapple with it. All you have to do is live on a liberal college campus to recognize that there is a lot of truth to what the essay says.
*Are Wile E. Coyote and the Roadrunner really gay, or have some kind of amorous relationship? That is the reaction some fans are having to the new movie, “Coyote vs. Acme,”
Maybe the Road Runner was just playing hard to get.
For decades, cartoon canine Wile E. Coyote has been pursuing the speeding bird throughout the American Southwest, aiming to eat him and getting blown up in the process.
But in the eyes of some fans, the violent chase was actually a classic tale of tortured love. And right now they’re doing a victory lap.
That’s because of “Coyote vs. Acme,” a new film with a different twist on the hot pursuit. In the movie (spoiler alert), Wile E. Coyote sues the Acme Corporation over their defective products, which consistently backfire as he closes in on the bird. The case goes to a congressional hearing, where Road Runner bursts into the room and becomes the only creature to testify on Coyote’s behalf. Through his language of “beeps,” Road Runner tells the room he considers Coyote a friend and asks, “Are you coming back?”
Apparently that’s all the internet needed to hear.
“Road Runner admits that he sees Coyote, even whenever Wile E. dislikes him or doesn’t think that he’s looking,” said Lisbon LaSure, a 24-year-old in Jacksonville, Fla. “It’s so nice to be seen like that. It’s really deep and it’s really powerful and it’s really validating and just overall like really romantic.”
Social media is teeming with people celebrating the characters’ “toxic yaoi,” a Japanese term for male lovers. You’ll also find people “shipping” them—wishing they were in a relationship—through fan art or TikTok videos where their interactions are reframed through careful editing and songs like Paul Anka’s “You Are My Destiny” or Olivia Rodrigo’s “Stupid Song.” (Sample lyrics: “I want you more than any stupid song could ever say.”)
Annie Tarin, a University of Arizona student, made a viral montage set to Chappell Roan’s “The Subway,” in which the pop star croons about a former soulmate. In the clips Tarin uses, Coyote looks longingly at one of Road Runner’s feathers. Road Runner seems to stare into Coyote’s soul with his curious gaze.
I’m not ready to admit that there’s an amorous relationship between the two (are their sexes even known?) until I see the movie, which I’m not inclined to do. Has anybody seen it? Here’s the trailer:
Meanwhile in Dobrzyn, the editor is bothering the worker:
Hili: Am I bothering you well enough?
Andrzej: Exceptionally well.
In Polish:
Hili: Czy dobrze ci przeszkadzam?
Ja: Wręcz doskonale.
*******************
From Now That’s Wild:
From 2025 Darwin Awards!!!/Epic Fails!!!
From Meow Incorporated:
From Masih; vintage video of Iranians mourning the killings in America on 9/11:
Look carefully at these images.
Twenty-five years ago, we Iranians took to the streets after 9/11. We lit candles. We mourned the victims. We chanted against terrorism.
We did not need anyone to explain terrorism to us.
We had lived it. We knew what it looked like.
We knew… pic.twitter.com/53nrdhdv56
— Masih Alinejad (@AlinejadMasih) September 11, 2026
From Luana; Elizabeth Warren appearing a bit dense:
Dynamic pricing means prices go down as well as up. So this is conclusive evidence that either (a) she’s stupid, (b) she’s dishonest, or (c) both. The strange thing is, I genuinely can’t guess which it is. Could she lie this overtly? https://t.co/gh6sqFU4o4
— Paul Graham (@paulg) September 13, 2026
From Malcolm. I’m hoping these are all real, and so far there are no “community notes” that this is AI:
Pets that imitate their owners pic.twitter.com/VQDH9z240f
— Massimo (@Rainmaker1973) September 8, 2026
Two from my feed. First, translated from the Turkish: “At first glance, I thought it was a lion-shaped toy～ Turns out it was moving”:
İlk bakışta aslan şeklinde bir oyuncak sanmıştım～ Meğer hareket ediyormuş 🤩 pic.twitter.com/iFXcmhPipo
— Gözde (@DearGozde) September 12, 2026
From JKR:
You can’t help but be moved by the quiet dignity and eloquence of his argument. pic.twitter.com/5ocS0U9aRj
— J.K. Rowling (@jk_rowling) September 12, 2026
And one I reposted from The Auschwitz Memorial:
This Dutch Jewish girl was gassed to death as soon as the arrived in Auschwitz. She was eleven years old. https://t.co/ANEhaDM5M4
— Jerry Coyne (@Evolutionistrue) September 14, 2026
Two from Dr. Cobb. He’s right when he says, “Real, not AI”:
Red Lip Bat FishI swear they are real
Ducklings! (Sound up.)
Best things in life are three plus one our little Muscovy ducklings
— Chris and his farmily of forever farm friends (@caenhillcc.bsky.social) 2026-09-10T15:40:54.622Z
.
People seem divided about whether artificial intelligence is an existential danger or is not to be worried about. Yes. Divided as people are apt to be on matters of opinion 🙂
They apparently have the ability to cause problems in cyberspace and no doubt can take out the financial systems of society, if only with all the hype fuelled investment, but I live with the knowledge that they are toast unless they have some autonomous physical agent out here in the real world to maintain their power and cooling systems.
Even a Pyrrhic victory is a win of sorts. 🙂
Agree 100% with the power and cooling as a hard limit to the damage.
I also think it is naive to believe no one is working to weaponize AI to cause as much mayhem as possible.
I have copied J.K. Rowling’s comment, “You can’t help but be moved by the quiet dignity and eloquence of his argument,” for use with those brilliant people who respond to my pro-Talarico posts with “Shit! Don’t vote for him.”
I have not seen the new Coyote/Road Runner movie. My best friend did. He didn’t comment on any romantic undertones. Knowing what the relationship between the two has been historically, perhaps saying that it seems like it has a romantic element says more about the expectations of the speaker for what a relationship is, which is scary,
Re: Coyote vs ACME. Yes, I have seen it. And in my humble opinion, YOU MUST SEE THIS MOVIE! Not just because it’s a hoot, but because pokes a large, stupid corporation in the eye. Power to the people!
I agree. saw it with my kids and we all laughed so hard. The casting is great too. John Cena as a high powered corporate lawyer…
I understand that the creators had a real hard time getting it made, since studio executives could not wrap their heads around a movie that mixes cartoon and live action characters.
Mixing cartoon characters and live actors has previously been successful. Elmer Fudd is on record as saying he loved ‘Who Killed Woger Wabbit?’
In other news, re. the zoonotic origins vs. lab leak confrontation, this week’s TWiV is an important one, interviewing Paul Bieinasz from Rockefeller U on his et al’s paper on replication in coronaviruses. The results strongly support the possibility that the infamous furin cleavage site in SARS-CoV-2 could have been acquired through natural viral replication.
I’ve only done one pass through this, but the fundamental aspect is that the virus contains its own nuclease that cleaves the viral RNA into pieces during replication, which then reassembles (why this is part of the replication process is speculative, but it is). During the reassembly, errors occur at an albeit low but significant rate which ordinarily result in a selective disadvantage, and thus are never seen, while leaving the real possibility for ones to occur that confer an advantage, and some of these errors result in insertions/deletions in the genome.
The acquisition of a furin site via this route is not that statistically improbable – the RRGR sequence that we hear about is not the absolute requirement – RXXR is more like it. There are some constraints on what the X’s can be, but there is far more leeway in them than an RG sequence. Further significantly there are six codons for Arginine, or to put it another way almost 10% of all codons encode R, thus there’s a statistical skew for a random codon to be one for Arg. (I tend to forget this because looking at things from the protein side, the frequencies of the two basic amino acids, Lys and Arg, are each in the 5% range whereas there are only two codons for Lys).
All you need is one virion with a significant advantage to sweep thru the population. As evidence of that, just look at how quickly new SARS-CoV-2 variants sweep thru the worldwide population. There’s a link to the preprint in the link.
Meanwhile, get your damn COVID shot. It’s on the rise as usual in Sept with schools reopening. And if you’ve felt knocked out the day after one of the mRNA shots, ask for the protein-based NovaVax shot (aka Nuvaxovid).
Thank you for the lay interpretation. The TWiV itself was well beyond my life science challenged education. I had forwarded it to my retired doc friends for help, but your explanation is great! Much appreciated. Btw, they read and answered my letter on the impacts of the radical funding cuts on Drs Baric and Garry…the gang laid things out with devastating frankness. Starts around 1:23:45 and cover nine minutes I think.
Aha, that was your letter. Cool! I’ll give it a re-listen. I’ve had some of mine read from time to time, too – always ending with something about the weather in Greater Braddock.
I think the work is going to grow legs. I hope so, at least. A figure or two during the presentation would have helped. I didn’t really follow the business about the green fluorescent protein reporter sequence, but having gotten to the summary and then thinking back on it, I think I’m beginning to understand it. They must have incorporated a GFP gene segment out of register, so the sequence didn’t translate as a GFP sequence, but when one of those segments was re-spliced there was a 1 in 3 chance that it would insert in-register and translate as GFP and form the fluorophore, which would then give a signal under appropriate conditions. Brilliant – fluorescently!
And glad you liked my summary. I hope I didn’t mash anything up.
I’ve saved this to Watch Later, when I’ll have the time to concentrate and to prove to myself that I’m no virologist. Thank you for posting it and for your summary.
I am an A.I. skeptic, in that I am waiting for some meaningful use-case that will actually benefit businesses. But I am also an A.I. doom skeptic. Here’s an interesting piece that looks at some of the motivations of the doom crowd. To what it says, I would just add that con men always tell you that there’s no time to wait.
Firmly agreed. This entire wave of new AI technology has its main “function” and main threat wrapped up into one attribute – it itself is a highly effective conman. The whole training process that produces these kinds of agents ensures they’re always far, far better at looking competent than actually being competent, and there’s an enormous pool of people out there that, for a variety of reasons, are ready to be fooled.
I too am skeptical. But I wouldn’t mind if the damn thing slowed down anyway. Meanwhile, the growth of data centers worries me as something that is more like a huge economic bubble that will burst one day, without return on the considerable investments.
Data centers may end up going by the wayside. We’re working to bring AI models in-house (“on prem”) using conventional server stacks. We still need to provide power to those servers, and also cooling, but if this takes off on a wide scale then the large centralized data centers may not be needed. On the other hand, the electrical power draw and cooling required will still exist but it will be decentralized across thousands or millions of individual users. Even that will be reduced as chipsets and models become more efficient and draw less power (and create less heat), similar to the case of your iPhone having more computing power than a room full of IBM computers from the 70’s while utilizing much less power.
Like those big IBM mainframes, data centers will likely become obsolete. Is that a bubble or is it technological progress?
That’s very interesting, thank you Darryl, I always enjoy your posts.
I can easily imagine a data center glut, esp given the capital expenditure.
I was a trader during the dot com boom/bust centuries ago and that kind of oversupply is a good opportunity (though thankfully I no longer trade).
Big picture the dot com wasn’t all a bust though: without the huge increased capacity it is hard to imagine good video, streaming, youtube etc exisiting today. The infrastructure, not used for dot com got used for other things.
best,
D.A.
NYC
Thanks David, I enjoy yours!
You’re correct – these things typically don’t end as a bust but as more of a morphing into some other state.
Many Americans most certainly are retaining that sense: everything American is evil and everything anti-American is good.
(Ditto “the West”, “white”, “capitalism”, etc.)
But what is the actual distinction between Islam and Islamism? Islam is an all-encompassing ideology that amounts to a controlling prescription for all areas of life, and which says that Islam should and will come to dominate the world.
“Islamism” seems to be just a matter of taking Islam seriously. Mainstream Islam is not harmless nor benign.
The history of religion shows that it’s easy to be very devout but not take religion “seriously.” God is inscrutable. Even literalists admit that He/She/It often uses metaphors, is deliberately enigmatic, and is frequently misinterpreted by frail humans coming at it the wrong way (that’s those other guys.)
While I agree that it’s harder to come up with benevolent interpretations of Islam, with God all things are possible. Atheists are simply out of the game when it comes to crowning the True version of deity. I’ll accept the term “Islamism.”
Islamism seems to be a politically correct term for “Muslim“. I think we as a society need to start saying that Islam is the problem and stop using the term Islamism. We are making it seem that there are only a few militants willing to hurt other people, when many surveys after 9/11 showed that the number of people who thought 9/11 was justified was quite high.
100% Doug. There has been so much obfusification about the true nature of Islam whereby there’s Islam (the good one) and Islamism (the bad one).. sorta like aggressive Christianity and, well, most Christians we know.
But there is no such boundary in Islam. As there is none between politics and religion in Islam.
And yes, as Bill Maher noted, in those days we were anti-terrorism. There were no “Yea Al Qaida!” demos here in 2001 unlike “River to Sea” demos on Oct. 7.
I argued this lately: sorry for repost but I’m obsessed!
https://themoderatevoice.com/25-years-after-9-11-a-nyc-lookback/
best,
D.A.
NYC
“Islamism seems to be a politically correct term for ‘Muslim.'”
That’s going too far, as there are religious Muslims who reject the political side of Islam.
Here is a clip from a Family Guy episode where The Coyote actually kills, and eats, The Roadrunner. He becomes depressed because his life has no meaning afterwards. He nearly attempts suicide, but then finds a new purpose in his life. I won’t spoil the ending for y’all: https://youtu.be/5haWTqaxV7w
That was funny
My only worry with AI, being a retired software engineer who is not expected to use it, is the increasing investment in businesses that will demonstrably lose money for years to come. My pension specifically excludes such investment so I shall be happy and comfortable until they crash the entire world, hopefully long after I shuffle off this mortal coil, though it would be interesting to see how prophetic Idiocracy was as a film.
Years ago a creationist taunted me with “I’ll believe in evolution when someone shows me a fish with legs!” He’d obviously used this line a lot. I promptly went online and showed him some fish with legs. He seemed deflated.
Good thing I didn’t know about or present that creepy video of the Red Lip Bat Fish. He would have been more convinced than ever that evolution is a tool of the Devil.
Regarding AI and the doomer predictions, my thoughts:
1. Jobs will be lost, and jobs will be gained. My company is writing more lines of code than ever before by a factor of 10X at least (some much much more) using AI tools to do so. Non-engineers are writing their own code to automate mundane tasks. We’ve reduced our IS team by 5% but have increased engineering staff by 10% as we need fewer coders and more thinkers. I’ve made my own tools for setting up to do lists from my email and calendar, set up automated checklists for projects, and to create research presentation drafts. I see it as being similar as to the way secretaries and administrative assistants have been replaced. Is it completely trustworthy? It depends. It’s 100% accurate in making a to-do list for me, but hallucinations are still a problem on research projects. I find that Python code needs tweaking about 10% of the time.
2. We’re finding that a mix of commercial tools and on-premise models to be the best way to get latest models while preserving private business information. We actively use and support open source.
3. If Amodei and Altman want a pause on AI frontier development, why don’t they unilaterally do it? They are the big guys in the race, and they can stop themselves. Regulation at the federal level would lock in their duopoly and eliminate the ability to create and use open source.
4. The only way to prevent an AI cyber attack on a system is to be able to create AI preventive measures. I think of this as kind of the argument for gun control – if you criminalize gun ownership, only criminals will have guns. Similarly, if you mandate a stop to AI research, then the bad actors will outpace the good guys and we’ll be in for bigger problems. One small example is Pangram: while we now have the ability to create 100% AI content that looks like human writing, we also have created a program that uses AI to detect when something is written by AI.
5. As humans, we seem to look for existential threats. Nuclear war, climate change, GMO crops, the gods causing a flood because we sinned, etc. But as humans we are also wrong quite often about this and we also have the agency to create solutions. I’m an AI optimist.
While there might be grounds for government regulation of AI, I’m deeply skeptical when companies that dominate an industry lobby to be regulated. That’s a direct path to them sculpting the regulations in their favor and locking out competitors who cannot afford the regulatory morass. Politicians get to claim they protected consumers and helped keep everyone safe. Political parties and PACs get big contributions from the companies. Federal bureaucrats get a wider revolving door to bounce between the regulatory state and big paychecks. Startups and open-source labs get screwed.
From the railroads to the airlines to telecommunications, U.S. industrial history is littered with this type of market manipulation by essential industries. That doesn’t even address the well-documented cases of industries capturing a regulatory process they did not necessarily favor and subsequently distorting it to their advantage: agriculture, pharma, oil and gas, finance and banking, etc.
AI regulation. Some seems reasonable to my untutored mind. But I would tread cautiously. Watch China, which will remain unbound. Listen to those who present reasoned grounds for dire concern—but ignore the fearmongers. Even the direst problems require calm leadership; those who steadily fan the fears of others and exploit their ignorance are not to be trusted—on this or any political, military, public health, or other social problem.
Whenever I hear someone make the argument that the government should regulate something, my response to them is, “so you’re saying you want Trump to control it?”.
Yeah, just like some want to “reform” the Supreme Court–as long as a Republican president is not the one to pack it.
I think Warren meant surge pricing.
Yes, absolutely. Do we really want to be shopping for bottled water on a hot day and see the price go up before our eyes? And thus always have to “plan ahead” and be sure to buy bottled water on cool days to get the best possible price? This is the invisible hand of the market run amuck.
Back in the day, I worked on the early bases of the current form of AI. this work was based on what we called “neural networks”. It incorporates what we know about biological algorithms of computation, to use what we know about neural computation to solve problems that are easy for brains, but extremely difficult for classical computers. Facial recognition is a good example of such a problem.
The problem with AI today, as I see it, is that we do not understand the mechanisms involved in the AI systems used today. We know all about the inputs, and we know all about the outputs. But we understand very little about the mechanisms used to produce those outputs (what we used to call the intermediate (hidden layers) that the machines are using to produce those outputs. Until we do have that understanding, I think that AI indeed carries a certain amount of my of danger to humanity. I would like to hope that the corporations giving us these AI models are dedicating a great deal of research to such understanding, but since that direction of research is unlike to generate much, if any, profit in the short term (months/years), I have my doubts. I know for a fact that the titans of industry, who studied such mechanisms in their Ph.D.s under the likes of Smolensky, Hopfield, Sompolinsky, etc. are aware of the issue, but corporate interests have a way of reducing intellectual curiosity in favor of profit.
In short, if we don’t understand how a machine works, the results produced by the machine are suspect. This is not a problem for classical computers computation; we understand exactly how that works. That is not the case for the AI in use today in large model computation. In earlier versions of neural network computation, this was not an issue: when biological network algorithms were used for computation in applications such as the vision used for the Mars rover, the underestanding of the “retinal” computations used for early visual mprocessing were well understood. The rate of our technological advances have far surpassed our understanding of how the technology works, and therein lies the danger.