by liberal japonicus
I guess the big news is from the Michiganders and Abdul El-Sayed. I like Jamelle Bouie’s encapsulation of this, which is Vote Blue No Matter Who. If anyone has a more granular view of the race, please drop it in the thread and feel free to write about anything else.
There is a strong desire to cast the DSA/Democratic primaries as some sort of epic power struggle, because it feels like there ought to be one. All I see is an upstart organization exploting the free-floating anger in the Democratic base with generic lefist complaints about Big Money and the pinata of the year, Israel, for the 1 issue leftists. AIPAC hits both buttons with 1 push. In a campaign you push the buttons you think work best. Everyone complains about Trump so that’s no differentiator.
El-Sayed’s narrow win is a sign of how much anger is free-floating out there in the Democratic base. A version of Trump’s campaign strategy. It will be interesting to see if he continues or moderates, but in running against a Republican moderation offers no advantage. Any gain in moderate voters who want Mom and Dad to stop fighting will be more than offset by the loss of the angry base. Hell of a needle to thread.
The AIPAC thing is driving the algorithm, but what gets buried by the algorithm are all the core Michigan-centric issues around which the voters were making their wishes known: AI data centers, Medicare for All, solar farms, affordability. The data center issue in particular hurt both Stevens and Mills (in ME). Solar farm support is hurting El-Sayed some. In both cases I’d say that those issues matter more to many voters than the AIPAC thing, and which made choosing hard for many others.
The thing is, stuff that’s critical in the minds of local voters is of minimal interest to someone writing (or editorally selecting) news stories from a thousand miles away. So they focus on what the candidates said about issues which are of interest to them.
The rest of us can all to easily accept that those are the electorally important issues. Thus feeding the narrative that the voters aren’t being listened to. Because, in a very real sense, they aren’t.
Tony’s comment
I have heard liberal/progressive voices complain about Talarico beating Crockett in the Texas senatorial primary — to the point of saying “Screw it, I don’t care if Paxton wins the general.”
was interesting to me, because it points to an element of time. The primary between Talarico and Crockett was reported by some as a fight for the soul of the Democratic part, but it seems to have settled down and I have this sense that the primary battles was ages ago, which suggests that a possible alleviation would be to have the primaries earlier so all of the ill will would have time to dissipate. I know we have a scrambled basket of rules for these elections, but getting them further away from the election gives Trump time to piss everyone off and forget that they were oppposed in some way.
That’s how editorial boards work, but it ignores how the social media platforms themselves then select which of those topics to boost in their lists. I’m not at all surprised that stories about AIPAC generate more traffic than stories about opposition to data centers. The media is owned and controlled by the AI oligarchs, who slso happen to be major political donors, and also happen to then receive major federal contracts.
It takes work both to be informed, and to get outside of the enclosed information environment that the platforms build around us.
Editorial boards are coerced to play by the platform rules if they want the traffic.
Matt Yglesias coined the term “Pundit’s Fallacy”:
“the belief that what a politician needs to do to improve his or her political standing is do what the pundit wants substantively.”
Okay, I’ve had the urge to comment on this post for a while, but kept putting it off.
To wit: “inshallah is like the spanish mañana, but without the sense of urgency.”
Meanwhile, in another part of the forest:
https://www.theatlantic.com/magazine/2026/08/charismatic-christian-church-tennessee/687624/?gift=cx0iluuWx4Cg7JjlT8ugCbHOkkfFdV2HJU1M3fMoqLk&utm_source=copy-link&utm_medium=social&utm_campaign=share
The statistics! Sometimes I think really scary numbers of Americans are going stark, staring mad. Then I remind myself not to feel smug; as they say “when America sneezes the world catches cold.”
This stuff is the kind of thing we kept expecting at the end of the millennium, I guess it was just getting started…
By 2024, roughly 61 percent of American Christians agreed with the statement that “there are modern-day apostles and prophets,” and roughly half agreed that “there are demonic ‘principalities’ and ‘powers’ who control physical territory,” according to a survey conducted by Paul Djupe, a Denison University political scientist who is among the few scholars attempting to track the ways that NAR ideas are transforming Christianity. By December 2025, roughly 59 percent of evangelical Christians and 22 percent of non-evangelicals agreed that “the church should organize campaigns of spiritual warfare and prayer to displace high-level demons,” Djupe found in a follow-up survey.
Since this is now our open thread, I am hopping about all over the place. This, in today’s Grauniad by Katherine Rundell, a writer I admire (although I have not yet read her book about John Donne) is about education, kids and AI. She also writes children’s books, and I may have mentioned before that the first time I heard of her was when my sister (knowing I have read children’s books all my life and still do so) gave me a book by her called Why You Should Read Children’s Books Even Though You Are So Old and Wise. I found the piece below interesting, since we talk so much here about AI and education, I thought some of you might too.
https://www.theguardian.com/books/ng-interactive/2026/aug/08/i-hate-what-ai-is-doing-to-the-minds-and-happiness-of-the-young-katherine-rundell-on-the-view-from-the-classroom
I asked a contact who is the president of a billion-dollar tech enterprise, what would you teach your sons? He said he wouldn’t teach his children to use AI, because it will prepare them for nothing. He would teach them the humanities. Teach them who we are, who we have been, what we can make. The FT’s Gillian Tett reported that a senior financier’s New York firm now seeks out humanities students, because “AI-native” Stem graduates are entering the job market with “alarmingly shallow ideas”.
…..
I talk a great deal about reading for pleasure; I’ve written elsewhere about the necessity of pleasure in the classroom, and why it is not an optional extra but the hallmark of true education. We must find ways to show children that reading gives you access to an inexhaustible source of delight. But I worry, too, that we are limiting the meaning of the word pleasure. Learning is a strange and discomfiting experience; it’s a struggle and a frustration, and epiphanies are rare. It is not always pleasure in its easiest sense.
So I think we also need to talk to young people about reading to outrun and outfight injustice. Reading to get faster, swifter and tougher; to develop powers of scrutiny and criticism and huge creative imagination. Reading human work to insist that we are human. Reading for liberation and emancipation; reading for intellectual freedom, and to have the strength to fight in the arenas of the mind, because when we let a computer think for us, we put a gun to the head of civilisation.
There are dark days ahead. If we are to fight against those with untold financial wealth, we will need untold intellectual wealth. We will need tough-minded people; people with stamina, a huge range of knowledge, and an ability to see wide and far enough to understand in real time what is being done to us. Reading for pleasure and delight are vital – but alongside that and even more urgently: reading for power.
Rundell says nothing with which I disagree. The only bright side that we have in all of this is that the AI models we are developing are not sustainable, and a civilization built upon them will collapse in upon itself because it is not compatible with the planetary limits for life at large. What I worry about, though, is the stunted humanity that the collapse will leave in its wake.
We are failing an entire generation. We have already sold the birthright of at least a couple of previous generations even before this, but this current tragedy is a deeper betrayal.
…the AI models we are developing are not sustainable…
I (unfortunately) have faith in the applied mathematicians**. Now that we know that hardware improvements have made 1970s (and earlier) neural net math useful***, it’s a viable research subject again. I expect orders of magnitude improvements in the resources required.
** Long ago I was at some sort of social function, standing on the edge of a conversation, not really paying attention, when one of the people suddenly turned to me and asked, “What about you, Mike? Who are your people?” In hindsight I assume they were looking for something like Baptists or Germans. What popped out of my mouth before I had thought at all was, “The applied mathematicians.”
*** Eg, trained neural net OCR is a huge improvement over all previous approaches. There is a growing body of evidence that trained neural net global weather forecasting is superior in some ways to the big physics-based models. This is a couple of years old, but there are probably millions of people who would pursue such productions as a hobby if the price were right. I admit that I was working on something the other day and asked Google “c language subroutine bradley adaptive image thresholding” and it handed back code that worked and produced expected sorts of results.
Michael-
I would agree with you if that were the bet that the AI Oligarchs were making. It’s not. What they are pouring all of their resources into is the hope of scaling up LLMs to a magical point at which those LLMs spontaneously become capable of true creativity and insight, and will transcend their datasets to create new knowledge. Then the creators of the winning system(s) will live in the miracle of the singularity, and the rest of the world (whatever that is) will be sustained at an acceptably humane level just outside in their second-tier utopia.
If the AI bubble pops sooner, rather than later, we may get your neural net future once things hit equilibrium again, but the path they are on diverges from the one that you wish for.
Every friend I have who has been working for one of the big AI firms in the news is disillusioned, and looking to get out of that gig.
The models are not sustainable and will collapse much sooner than that. Their overwhelming weakness is actually well known. Consider: one of the problems AI users have is the tendency of AI to just make stuff up. Legal briefs and academic papers are among the high profile examples, but there are numerous others. Anybody using AI needs to have a human expert to review the output and fix the fantasy bits.
The problem is, we are raising a generation of people who use AI extensively, without acquiring the ability to give the output a reality check. They are like people who grew up always using crutches. In an emergency, they are simply unable to toss the crutches aside and run. Similarly, the AI generation aren’t learning the things that they would need to check to output of AI.
That pretty much nails it: prepared for nothing. Although I take exception to the thesis that that the humanities are somehow exempt. Humanities students are subject to the same mental crutches.
An actual Artificial Intelligence might be able to do that. But the LLMs that are currently branded as AI? They are inherently unable to do that. They can write new sentences which emulate what was written in the documents that they were trained on. But that’s it. They might happen to create something which constitutes new knowledge. But it would take a human being, a knowledgeable human being, to establish that.
“I was working on something the other day and asked Google “c language subroutine bradley adaptive image thresholding” and it handed back code that worked and produced expected sorts of results.”
But – you knew the results to expect.
It might be enlightening to as the most junior, fresh out of school, employee to evaluate the results. With details of how the evaluation was arrived at.
Any bets on how that plays out?
Masha Gessen drawing parallels between Trump/Heritage Foundation’s US and Soviet/Putin’s Russia – unfortunately quite on point.
https://www.nytimes.com/2026/08/03/opinion/trump-smithsonian-history-censorship.html?unlocked_article_code=1.4FA.3FJA.AtfCq5fDz7eI&smid=url-share
I would agree with you if that were the bet that the AI Oligarchs were making. It’s not.
I don’t expect order-of-magnitude breakthroughs to come from the AI oligarchs. I expect them to come from people who think “Doing this on-the-cheap is an interesting research question,” because they don’t have the massive resources of the oligarchs.
Every friend I have who has been working for one of the big AI firms in the news is disillusioned, and looking to get out of that gig.
I certainly sympathize with them. I retired from telecommunications (and related) work because consolidation of the industry resulted in a situation where none of the people in charge wanted to pay for anyone to ask interesting questions.
Solving math.
“We provide new results for the following problems. The results were achieved by an internal version of Astra, our next major model. The total number of tokens needed to find solutions to these problems would cost roughly $2,000 at Sol API rates. These arguments were then prepared into manuscripts by humans with the same model. Afterward, the model formalized each argument in a Lean certificate for each solution a model’s narration of its thinking process.”
1. High-dimensional sphere packing…
2. Binary and spherical codes…
3. Non-sofic groups…
4. Connes’s rigidity conjecture…
5. Arithmetic circuit complexity…
6. Quantum parallel repetition…
7. Closest vector problem…
8. Ehrhart’s volume conjecture…
9. Multicolor Ramsey numbers…
10. Extremal number conjectures…
Ten advances in mathematics and theoretical computer science
As y’all know, my wife has been in memory care due to early onset dementia. She died early this morning. It’s sinking in, and there’s stuff that has to be done, so I’ll probably be offline for a while.
so sorry to hear that Michael.
take care.
Take care Michael. So sorry for you.
Michael, please take care.
Echoing what everybody else has said, Michael. Take care.
So sorry to hear this, Michael.
The ones we love leave a big hole when they leave us. I’m so sorry for her, and for you, for the difficulty of her last years.
Well hell. I don’t have any words appropriate to this.
I am very sorry, Michael.
Sincere condolences. Take the time you need.
I am sorry to hear about your loss, Michael.
That needed to be separate from my rant.
The comments here on the internal politics of the Democratoc Party are why I left this blog.
Start with the unfair comments or lies or whatever about Stevens. Welcome to politics. Nobody should be excused for lying. Thank goodness we live in a universe where the decent workhorse Democratic politicians and their supporters never lie about their opponents, never stereotype the supporters of the other side, never hippy punch, and oh, if they support a massive atrocity overseas, well, that’s not great and it involves telling some pretty massive lies but at least they aren’t lies about Democratic politicians. I understand that wonkie wouldn’t have voted for Stevens because of Gaza, but if lying matters, take a look at the lies Stevens had to implicitly or explicitly endorse to justify her position. Anyone familiar with this or similar issues is familiar with the massive, gigantic cold- blooded lies told about it.
Gaza is typical of what politicians support. It is the dreaded social media which in the modern age, literally the past few years, finally made many Americans realize the sort of crap we support overseas. The “ moderates” were caught flatfooted, Some of them, like Clinton, blamed TikTok. Atrocities supported by our politicians aren’t supposed to matter in Democratic primaries, which means, let us be clear, they are not supposed to matter at all. If a politician can win a bit of support for supporting an atrocity and suffers no negative consequences, that is what they will do. In this particular case, as opposed to something like Yemen, where maybe one person in a hundred knew about it, politicians had a standard playbook answer, “ Israel has a right to defend itself”. To their shock, this no longer settled the issue.
I suspect that Gaza was not the main issue for many Michigan voters or I hope not— I would hate to think the election would be so close if it were. But it did matter to some, probably including the people who had lost family members to the bombs Stevens wanted to ship over. They didn’t even need a Twitter or TikTok account to know about it.
For decades the left was told that it shouldn’t go third party but fight things out in the Democratic Party primaries. It turns out that this was wasn’t what was wanted—they were supposed to accept being insulted as crackpots and to vote for the centrist and shut up.
Regarding various DSA types and fellow travelers, Chevalier is unable to give a straight answer to whether a murderer should go to prison but the rest seem fine. Lander is great and the fact that he beat Goldman is fantastic. I wish we had had a someone like him in my district running against that jackass Latimer. I do not assume that leftist candidates are the best in purple states- I hope it works out in Michigan because if he loses, it is bad for the country and also, the “ atrocities don’t matter” crowd will have a field day.
So for now we live in an age where Democratic politicians have to at least pretend to care more about atrocities we support in Gaza. This is long overdue. It may not be the central issue for most, but fortunately it is no longer an issue that people can dismiss with the condescension one typically sees in the Cult of the Savvy.