XPENG had very minor sales growth in July year over year, rising 4% from 36,717 in July 2025 to 38,027 in July 2026. Across the first seven months of the year, the company had 204,004 sales. That’s still down a bit from the first seven months of 2025 — 233,906 ... [continued]
The post XPENG’s Sales Rise 4% in July appeared first on CleanTechnica.
NIO logged 35,934 vehicle deliveries in July. That’s 71% more than in July 2025, when the company had 21,017 deliveries worldwide. However, it’s slightly down from the 40,597 deliveries of June 2026. Across the first six months of the year, NIO had 227,057 vehicle deliveries, and the company’s cumulative vehicle ... [continued]
The post NIO’s Sales Rise 71% in July appeared first on CleanTechnica.
I’ve gotten on an AI kick this past week, and that means writing about the AI bubble that has been forming — see here, here, here, and here — as well as some severe risks of AI and this fun idea to replace CEOs with AI. But the fact of ... [continued]
The post Magical Thinking & Modern Society appeared first on CleanTechnica.
I truly cannot comprehend how someone with so much money and ability to hire people to help him understand things can be so completely detached from reality. Or perhaps that’s what a trillion dollars does to a person. In actuality, though, there is no excuse for Elon Musk — his ... [continued]
The post Elon Musk Is Still Saying Crazy Things About “Era of Abundance,” Among Other Things appeared first on CleanTechnica.
The largest conservation areas in the U.S. are in the Pacific Ocean, where tiny islands are epicenters of biodiversity. The Trump administration is now opening their waters to commercial fishing, sparking a battle over some of the country's most pristine ecosystems.
Last week we wrote about how Elon Musk’s xAI had filed a lawsuit to attempt to block Minnesota’s anti-nudify app law. As we tried to explain, even if you (reasonably, understandably) dislike both Elon and “nudify” apps, there were real problems with the Minnesota law. In particular, it was not narrowly tailored to just target truly harmful image edits. Indeed, it wasn’t even limited to the non-consensual use of the tech. The state’s Supreme Court had already handed the Minnesota legislature a clear roadmap for drafting a law like this that would pass strict scrutiny. The legislature ignored it.
However, the lawsuit was filed just days before the law was set to go into effect, which was called out by Minnesota’s Attorney General in arguing against a temporary restraining order:
First, X.AI’s lack of diligence confirms that a TRO is unnecessary. The company waited until the last minute to sue…
And that seemed to influence Judge Donovan Frank, who denied the motion for the TRO mainly because Elon waited until the last minute to file.
The Court respectfully denies the request for a temporary restraining order before tomorrow. xAI filed the motion on July 29, 2026, nearly three months after the law was signed, and only three days before the law is set to take effect. Such a delay in bringing the action and the motion suggests that harm is not immediate.
And, sure, it was kind of silly for Musk to wait right up until the law was set to go into effect, but that’s not all that rare with challenges to these kinds of laws. I also find the court’s suggestion that the harm isn’t immediate a bit odd, given that (as xAI had rightly pointed out) the potential liability under this law is massive: $500,000 per “access, download, or use.” That means any single use of Grok to edit an image that violates this law (which, as we discussed, goes way beyond nonconsensual sexual imagery, and could even cover someone editing a photo of themselves in a way they endorse) could lead to a huge bill for the company. And it’s now in effect, meaning in theory Minnesota’s AG, Keith Ellison, could already seek fines against the company — though there’s no indication that his office has done so yet. And while it may be politically appealing to try to enforce immediately, that may play badly before the court when there are hearings coming up in a few weeks on a preliminary injunction.
It’s also unclear if xAI actually changed anything on its end. In its filing, the company said that if the law went into effect, it would need to restrict access to certain features:
Confronted with $500,000-per-image strict liability and no safe harbor, xAI has no practical choice but to restrict Grok Imagine’s image-editing features in various ways when the statute takes effect on August 1, 2026
But the law has gone into effect, and as far as I can tell, there’s been no announcement of any changes. It’s possible such changes have been made already and just not announced. But it does come off as a bit weak to file a lawsuit on Monday saying that “if this law isn’t fixed by Saturday we’ll make big changes” and then have the law go into effect… and those changes are not publicly announced anywhere.
This ruling may not mean very much at all. The court has ordered both sides to brief a preliminary injunction over the next couple weeks, with a hearing on August 19th that can get into the actual First Amendment problems with the law. That also means xAI will likely have to explain, in those filings, whatever restrictions it has or hasn’t added to its systems.
xAI’s initial filing was not bad, but I hope they lean more heavily on the case I discussed in my last post, in which Minnesota’s Supreme Court spelled out exactly what the state’s non-consensual intimate imagery law needed to survive strict scrutiny. Because that case walks through, in great detail, the steps a (somewhat similar) law had to take to pass strict scrutiny and survive the First Amendment — even as the court acknowledged the law was punishing a form of protected speech.
There remain some oddities around this law, starting with the fact that the ACLU of Minnesota backed it in the first place — a surprising stance for the ACLU — before reversing course once Musk sued:
xAI’s criticism of the law is drawing support from some free-speech advocates including the American Civil Liberties Union of Minnesota, which supported passage of the law in an earlier form in February but has since turned against it.
“While we believe that creating the technology to alter or ‘nudify’ photos of identifiable people is protected by the First Amendment, we also recognize that the non-consensual creation and dissemination of such material can inflict damage on people appearing in those images,” the ACLU of Minnesota said in a statement Friday.
“In engaging with lawmakers on this issue, we hoped to strike an appropriate balance between First Amendment rights and the ability of people harmed to seek remedial measures, not unlike the remedies available to people harmed by defamation. The final version of the bill does not strike that balance,” the organization said.
Seems like the sort of thing you should have worked out before supporting the law, but fine.
Also, I had mentioned in my last article that some believe the law was written so badly on purpose, to convince Musk to sue in order for Democrats to use it as a political tool and… they are certainly making political hay of it on X, where they seem to be overjoyed that they can mock Elon.
And, sure, mocking Elon is fun. But if you’re going to mock him, it should be over the things he’s actually doing that are problematic.
The issue with this law isn’t that it’s trying to deal with the issue of nudify apps. Or that it’s trying to stop Elon from doing terrible things. It’s that it did so in such a ham-fisted, damaging, obviously unconstitutional manner that bans way more than it claims, is not narrowly targeted, and pretty clearly cannot survive strict scrutiny. Minnesota had the roadmap to pass a legitimate version of this law. It chose not to do so. That Musk didn’t receive the immediate TRO due to the late filing doesn’t make the law a good law. The proper thing for Minnesota’s legislature to do would be to write a law that actually abides by the First Amendment, but perhaps that wouldn’t get them the kinds of headlines they’re getting now.
Wildfires near Spokane, Wash., have burned at least 7,000 acres, destroyed hundreds of homes and forced the evacuation of tens of thousands of people.
Changes in the Earth’s climate can be measured by satellites and sensors in the ocean, but they are hard for ordinary people to visualize — which makes them easy to ignore. Wildfires, on the other hand, do a splendid job of focusing our attention on the situation. Here is a ... [continued]
The post Wildfires Are The Earth’s Way Of Telling Us To Change Our Behavior — Or Else appeared first on CleanTechnica.
VinFast expands retail footprint with new e-scooters and plans for 30,000 exchange hubs MANILA, Philippines — Vietnamese battery-electric vehicle maker VinFast has formally introduced its electric two-wheeler division to the Philippines, signaling a major push to decarbonize one of Southeast Asia’s most congested commuter landscapes. Marking its initial product rollout ... [continued]
The post VinFast Debuts Electric Scooter Lineup and Battery-Swapping Network in the Philippines appeared first on CleanTechnica.
Wildfires in Spokane have destroyed hundreds of homes and led to evacuation orders for 60,000 people.
(Image credit: Young Kwak/AP)
Ten years ago, I wrote a piece on how the internet was sucking journalistic resources out of the middle of the country and concentrating it on the coasts — primarily New York City. While digital publishing theoretically could have distributed reporters away from high-cost cities, in reality, it reduced the financial returns on local reporting and increased the physical distances between journalists and their audiences.
A lot’s happened to media in the ensuing decade — perhaps most notably, the pandemic-fueled rise in work-from-home arrangements that, again, could have reduced the number of journalists overpaying for a Brooklyn walkup. But Bloomberg’s Justin Fox has an interesting datapoint today that shows the concentration of media keeps on growing.
The nation’s media capital is in many ways far less central to the national conversation than it used to be, with TikTok, YouTube, Substack, artificial-intelligence chatbots and other new modes of communication sucking up so much of Americans’ time and attention. Yet in the production of certain kinds of media — journalism in particular — New York is more dominant than ever.My measure of dominance is employment, which is incomplete and flawed but better than anything else available. Having looked into the centralization of media employment a decade ago, I was inspired to revisit the data by an online complaint (from Minneapolis) about “the New York-ification of Democratic politics.” What I found is that New York-ification of media has for the most part continued to grow even as media industries have continued to struggle.
Fox pulls labor data for a number of media professions, including magazines, book publishing, and broadcasting. But of most interest for our purposes is newspapers (a category which now also include digital-only publishers). In 2016, when I was writing about coastal concentration, 1 out of every 30 U.S. newspaper jobs (3.3%) were based in Manhattan. That was already up markedly from 1-in-48 jobs (2.1%) in 2000 — which was, roughly speaking, the final peak of the print newspaper business.
But that trend hasn’t retrenched in the years since — it’s accelerated. By 2025, 1 out of every 12 U.S. newspaper jobs (8.3%) was based in Manhattan. What was once one of America’s least centralized industries is increasingly squeezed into a few square miles of Midtown.
Check out the full post to see how the distribution of “news analysts, reporters and journalists” has become increasingly Acela-adjacent, along with other slices of contemporary media employment.
A translation of a mass shooter’s cryptic journal in the days after an attack. A public records review that revealed failures to install flood warning systems in Central Texas. An exposé of American technology companies’ complicity in building the Chinese surveillance state. An audit of the SEC’s crypto lawsuits that showed weakening enforcement under the second Trump administration.
On May 4, the Pulitzer Prizes recognized these stories among the winners and finalists across 15 journalism categories. The reporters behind each of these stories also disclosed using AI to the judging committee. Ultimately, five award winners and three finalists this year disclosed AI adoption in their submissions — the most since the disclosure requirement was added in 2024.
For the past two years, I’ve spoken to Pulitzer-recognized reporters about how they used AI in their reporting. Both years, generative AI took a back seat to more conventional machine learning technologies, like using embedding models to produce complex data visualizations and pattern recognition models to analyze satellite imagery in conflict zones. This year, though, generative AI tools and commercial large language models (LLMs) were more commonly used, largely to speed up the process of combing through document dumps.
“To state the obvious, perhaps, AI is here to stay,” Marjorie Miller, the administrator of the Pulitzer Prizes, told me. “The industry [used to be] far more apprehensive about AI tools than it is today, with a clearer understanding now of what uses might be appropriate — data collection and analysis, for example — and when it might not, such as in writing and editing stories in any format that might be considered for a Pulitzer Prize.”
Miller cautioned that as AI evolves, reporters will need to “ensure and reassure” the Pulitzers that submissions are ultimately produced by human beings, even when AI is used as an assistive tool. Given recent controversies over AI-generated text allegedly appearing in prize-winning literary works, Miller also said next year the Pulitzers will include an AI disclosure question in their book entry forms.
In the days after deadly floods hit Kerr County, Texas in the summer of 2025, reporters at The Wall Street Journal had a clear reporting question: Had this area ever dealt with dangerous flooding before?
To find an answer, the reporters turned to public records. The team built a custom scraper that pulled every public meeting minute, agenda, and transcript from the Kerr County website.
Journalists are trained to find the needle in a haystack, but doing so on a breaking news timeline can be challenging. To speed up the document review, the reporters leaned on a pre-built internal tool called WSJPT (a play on ChatGPT). The tool standardizes basic LLM requests across reporting projects, including prompts for summarization, classification, and image description. In this case, the reporters used the tool to summarize every page of every document scraped from the county portal.
The team combed through these summaries using a combination of LLMs and more old-school natural language processing (NLP) techniques (e.g. stemming, lemmatization) to find sections that referenced past flooding events.
“We aren’t obviating the need for human investigation of a pile of documents — and I don’t think we would if we could,” said John West, a computational journalist at the Journal. “Instead, we’re trying to sort the pile so the most relevant stuff is right at the top.” West clarified that every section flagged as possibly relevant by these tools was read by a reporter, and then every document deemed relevant was read in full.
Based on this analysis, the Journal identified former Sheriff Rusty Hierholzer, who had pushed county commissioners to install a stronger flood-warning system a decade ago. In 2016, Hierholzer called for the installation of outdoor sirens, recounting an experience flying in helicopters and “pulling kids out of trees here (in) our summer camps” when floods hit nearby Kendall County in 1987, killing 10 campers. Hierholzer’s recommendations were not implemented at the time, the Journal found.
The findings were foundational to one news story and several follow-ups on the floods. The Pulitzers named the Journal’s overall coverage a finalist in the Breaking News category. West says this playbook — using a “mix of off-the-shelf and custom software” to summarize and parse documents — was also central to the Journal’s reporting on the Epstein Files this year. That coverage was a finalist in the Public Service category.
Like many of the investigations this year that disclosed AI adoption to the Pulitzer judges, no AI disclosure (such as a label, footnote, or accompanying methodology) appeared in the Journal’s own stories on the Central Texas floods.
“We did not disclose the use of AI. It functioned as a sophisticated way of searching through the documents, but we read the docs, and ran the findings down,” said West.
He contrasted that choice with a recent Journal investigation about toxic fume incidents on U.S. commercial aircraft. That story used LLMs to read more than one million FAA documents and to generate incident rates per airline and aircraft. For that story, which did disclose AI usage, West said he “got to write the longest methodology statement I’ve ever written.”
On August 27, 2025, a 23-year-old woman killed two children and wounded 27 others during a mass shooting at the Annunciation Catholic Church in Minneapolis, Minnesota. The shooting rocked the local community, but in the hours that followed, there were few answers about the shooter’s motivations.
When news of the shooting first broke, reporters at The Minnesota Star Tribune gathered in a Slack channel to coordinate their coverage. They identified the shooter’s YouTube account and videos showing her turning the pages of a journal written in a language the team didn’t recognize.
Dana Chiueh, an engineer in the Star Tribune’s AI Lab (now a fellow at ProPublica), took screenshots of the videos and entered them into an enterprise ChatGPT account. The chatbot recognized the text as Faux Cyrillic, a variant of Russian typography that can be used to spell out English words.
“[Faux Cyrillic] is not a real language. It can be thought of more as a type of code that one might use if they were trying to conceal what they were writing,” said Chiueh, explaining that ChatGPT allowed them to quickly see if there was relevant background information buried in the code.
After hours of tedious manual screenshotting, Chiueh wrote a custom script to pull the screenshots from the YouTube videos automatically. Ultimately, ChatGPT was able to produce an initial translation pass on hundreds of journal pages, over 600,000 words.
“At first we were doing something really scrappy. That spirit of being able to quickly prototype and iterate is something that is really useful for a breaking news situation,” she said.
A team of journalists then put the translations into Google’s NotebookLM, a Gemini-powered notetaking tool they used to search for keywords and pull out themes. They found mentions of past jobs, relationships, and locations the shooter had visited, like pawn shops and shooting ranges — all information that informed the outlet’s shoe-leather reporting.
“We were very conscious that AI hallucinates, so we made sure that any quotes, context, and anecdotes were reviewed by a human translator,” said Tom Scheck, investigations editor at the Star Tribune.
Rather than sending the entire AI-generated translation to a professional, reporters flagged important passages for review by two Russian language academics at the nearby St. Olaf College. For the most part, the AI-generated translations were correct, but the academics found a few errors, including a passage that misrepresented the shooter’s potential motivation.
If it weren’t for the help of AI translation, Scheck says the Tribune would probably have hired a translator to go through the documents from the beginning, slowing down their turnaround time. Instead, the triaged translations informed an initial story on the manifesto the night of the attack and contributed significantly to a profile of the shooter published four days later. Both stories were a part of the coverage that won the Star Tribune a Pulitzer in the Breaking News category.
“AI allowed us to take a first pass on the content and then prioritize what we [might] use,” he said. “We know we have to run the marathon, but AI helped us start at mile marker five instead of at the traditional starting line.”
Over the past 25 years, the Chinese government has built up a sophisticated mass surveillance program. A series of investigations published by the Associated Press last year exposed just how many of the technologies fueling this surveillance apparatus were sold to China by American companies.
Reporters mapped the supply chains for state-of-the-art surveillance tools, tracking their development in Silicon Valley and deployment in China, implicating companies like Nvidia, Intel, IBM, Dell, HP, Cisco, Oracle and Microsoft. The reporting earned the AP a Pulitzer win in the International Reporting category.
Key to these stories were tens of thousands of leaked emails and databases from a Chinese surveillance company, as well as thousands of government records and procurement documents (like vendor bids, signed contracts, and invoices). AI was essential to sifting through these documents and making them searchable, according to Garance Burke, a global investigative journalist at the AP who worked on the project. The AP used LLMs to identify specific company contracts, summarize government records, flag specific people or technologies for further investigation, and organize all the information collected into more easily managed databases.
In other words, AI was an assistant in the early reporting and research stages of the investigation, helping to make sense of a massive pile of documents. The team unearthed evidence that IBM had worked with the Chinese defense contractor Huadi to design a national fingerprint database, and evidence that Intel and Nvidia helped enable AI capabilities on surveillance cameras used in Xinjiang and Tibet. The documents also showed that HP sold the Chinese police “digital fencing” products, which have been used to track when Uyghurs and other surveilled populations try to travel outside their home towns and provinces, among many other findings.
“The AI tools helped reporters search and review large volumes of public records more efficiently, but they did not replace the reporting or verification process,” said Burke. “Reporters manually reviewed documents surfaced through AI, independently assessed the accuracy of AI-generated summaries, and did not quote from those summaries.”
Donald Trump is the self-declared “crypto president” — an industry booster who earned over $1.4 billion in personal crypto business dealings during his first year back in office. A team of reporters at The New York Times wondered if the administration’s pro-crypto stance had influenced the work of agencies that regulate the industry, namely the SEC.
To try to answer that question, the Times reviewed all of the SEC’s crypto-related enforcement actions dating back to 2017. The analysis surfaced a troubling trend. Since Trump took office again in 2025, the SEC had pulled back on more than 60% of its ongoing crypto cases, lessening penalties, freezing suits, and even dismissing cases entirely.
“It is unheard of for the agency to retreat from a swath of lawsuits against a single industry,” wrote the reporters in their investigation published last December. “Although the particulars of the crypto lawsuits differed, many of these firms had something in common: financial ties to Mr. Trump.”
The investigation is one of several stories that exposed Trump’s ongoing conflicts of interest with the crypto industry and earned the Times a Pulitzer win in the Investigative Reporting category.
Many of the Pulitzer awardees that used AI this year disclosed using LLMs to speed up and prioritize document review. The Times investigation stands apart. For their investigation, reporters downloaded more than 10,000 documents, including thousands of SEC news releases and over 700 federal court cases, according to Miller, the Pulitzers administrator. Reporters read and classified each of these documents manually over the course of several months. LLMs were only brought into the reporting process after this full human review was completed.
Reporters used OpenAI’s GPT-5 model to conduct a secondary review of the documents and check their work. They fed the model the documents from each of the lawsuits, as well as a detailed set of instructions on how to classify them. These classifications labeled “whether a case was crypto-related, whether it was inherited by the next administration and how liability was decided,” according to a methodology published by the reporting team.
The team compared the GPT-5 classifications with the ones assigned manually. When there were discrepancies, reporters went back and read the documents again to double-check their work. The Times declined to provide further details on this review process.
Hallucinations and other errors produced by LLMs are often cited as reasons not to use generative AI in investigative reporting. In this case, Times reporters turned the tables. They used LLMs as a tool to help keep human error in check.
Thinkings out loud
Overheard: "In tech, every bit of progress makes something easy obsolete."
Did AI make me believe this?
Charley Johnson: AI Can Change Your Mind Without You Noticing
Why just brands?
WordPress VIP: Future of the Web 2026: Chapter 1: The brand that feels human in the AI era earns the decade ahead.
A way for your stuff
Phil Windley: Using Home Assistant with Manifold.
Further evidence that advertising corrupts and digital advertising corrupts absolutely
Joe Mandese in MediaPost: Why Malvertising's New Vector May Leave A Sour Taste In Your Mouth. Excerpt: "Advertising — or more explicitly, 'malvertising' — has long been a vector for distributing malicious software, but a new, insidious exploit utilizes an innovative model for distributing it. Instead of utilizing a single malvertisement to install malicious code on a user's device — usually by redirecting the user to a landing page that automatically download pre-compiled malicious code — SourTrade does it in clandestine installments, treating the user's browser as a passive file receiver that functions as an active local compiler or assembler launching the code when it is completed. You can read more about it in this post by the team at cybersecurity firm Confiant…" Which is here.
Sometimes it helps to just write out everything that happened leading up to a situation so we don’t lose sight of just how ridiculous it is.
I mean, would you believe that a President would sue his own IRS for something that happened under his own watch and did him no harm, demand $10 billion, and when a judge calls into question how that’s possible, come up with a “settlement” between the two “parties” that grants himself, his family, and his far-flung businesses (already under investigation for tax fraud) a complete get-out-of-jail-free card from tax audits along with a special $1.776 billion slush fund which he can hand out to insurrectionists who tried to overturn an election he had lost?
And then, his Acting Attorney General seeking the permanent job (who had been the President’s personal lawyer in early criminal cases, including one where the President was convicted of 34 felonies), would tell inquiring Senators that the slush fund wasn’t moving forward, but refuse to put that in writing?
When you write it all out that way, it sounds kinda crazy.
Last month Judge Kathleen Williams nuked the fund and alerted various state bars to the ethical lapses of all the government lawyers involved, including Todd Blanche, the President’s personal criminal defense lawyer-turned Acting Attorney General-turned nominee for the permanent Attorney General position.
Blanche’s nomination had stalled out, in part because two Republican Senators (Cornyn and Tillis, both of whom are leaving office after Trump refused to support their attempts to stay) have questioned the slush fund. Blanche refused to put it in writing that the fund was going away, leading to this strained exchange with Cornyn:
Republican Sen. John Cornyn of Texas, reading from the settlement on Wednesday, noted that it says the terms of the deal “may be modified only upon the written agreement of the parties.”
“Has there been a written agreement of the parties to modify the settlement fund?” he asked.
“No, the settlement fund is just not moving forward,” Blanche responded. “There’s no modification. It never started. No money went from the Treasury to any other account.”
When pressed, however, as to whether the settlement agreement may be enforceable by the parties, Blanche conceded that the president could challenge it down the road.
“Yes, it’s an enforceable document, so I suppose if President Trump’s counsel sought to enforce it, that they potentially could … try to enforce the contract. They can’t force the Department of Justice to move forward with the weaponization fund. They could potentially say that we breached by not moving forward,” Blanche said. “They haven’t done that, and I’m not aware that they’re planning on doing that.”
Given all of that, you would think that, perhaps, Trump wouldn’t make any public effort to say that the fund should still exist.
But, again, we live in the upside down world, where if Donald Trump wants something, he just throws a temper tantrum until he gets it. Three new developments have happened regarding all this in the past few days.
First, on Friday, Donald Trump appealed Williams’ order and filed a long, rambling motion about how unfair the order was. That doesn’t sound like someone who is willing to accept that the fund is dead. Second, on Saturday, Trump posted a rant to his personal social media site about how mad he was at Cornyn and Tillis, and how he still needed the fund, and even hoped Congress would pass a separate law giving him the slush fund. Finally, despite these two things making it abundantly clear that Trump still plans to create a slush fund for his cosplay militia, on Sunday, Blanche finally signed a letter claiming the slush fund was dead.
Let’s take these each in order.
First the appeal and the whiny motion about it. The most incredible thing about it is that, even though it was filed by lawyers hired by Donald Trump (the person), as I was reading it, it read like it was filed by the Justice Department. Remember, the entire reason that Williams killed the “settlement” of Trump’s lawsuit against his own IRS was that there appeared to be no adversarial parties in the case, and it was just Donald Trump suing himself in order to give the “settlement” the stamp of authenticity that it was part of a federal case.
While, to their credit, this new filing at least brings in real lawyers from an actual serious law firm, DLA Piper, alongside Trump’s original two-bit lawyer Alejandro Brito, the argument itself is still a mess. Again, you could totally picture either Donald Trump or his DOJ/IRS filing this since the original ruling applies to both. Because they’re not adverse parties. They’re all on the same side.
The ruling rests on a grievous legal error: that, because President Trump supervises the Executive Branch, he and the United States necessarily possess the same legal interest and cannot be adverse to each other. That is wrong. President Trump asserted a personal claim arising from the theft of his own tax information. The IRS and the U.S. Department of the Treasury (“Treasury”) represented the sovereign’s interests in public funds and federal law. Donald Trump Jr., Eric Trump, and The Trump Organization asserted their own independent claims. Presidential supervision did not erase those rights or merge those interests.
The Sanctions Order nevertheless erroneously treated its novel theory of adversity as the starting point and then refracted the facts through that mistaken lens. Settlement became proof of collusion, even though there was none, potential (weak) defenses became proof that the claims were fictitious, which they were not, and ordinary professional relationships became proof of coordination, which did not exist. But the Sanctions Order identified no pre-filing sub rosa agreement, no false allegation, and no specific direction by President Trump controlling Defendants’ litigation decisions. The predicate legal error thus supplied the conclusion, causing the Sanctions Order to incorrectly recast entirely innocuous facts as misconduct.
The danger posed by the Sanctions Order is profound. The Court called the Article III question “unprecedented,” appointed six amici, and devoted nearly thirty pages to resolving it— yet it then declared the contrary position “so obvious and so insurmountable” that advancing it warranted career-altering sanctions. Id. at 9-38, 53 n.69. That mistake converts disagreement into professional punishment, and strikes at the integrity of the adjudicative process itself. A system that sanctions lawyers for advancing positions on constitutional questions cannot sustain fearless advocacy or principled legal development.
Note the tell in that last bit: the “professional punishment” landed on both sides — Trump’s lawyers and the government’s — which is exactly the point Williams was making. And, yes, courts are historically reluctant to sanction lawyers at all. But that reluctance is the problem, not evidence that the one judge willing to do it got it wrong.
As for the claims that the DOJ was actually representing the IRS’s best interests, and not Donald Trump’s, that’s belied by quotes from Donald Trump himself — who publicly admitted he was negotiating with himself — and the terms of the “settlement” which make no sense and do nothing to support the interests of the IRS, the public, or the American taxpayer. If this case was actually about Trump’s leaked tax returns, how does giving the family a release from tax audits or handing over nearly $2 billion to insurrectionists (wholly unrelated to this suit) have any of the American people’s interests in mind?
The strongest argument this filing has is that one part of the ruling is arguably prior restraint. To make sure that the settlement agreement for the slush fund was not used for improper enrichment, Judge Williams barred the parties from invoking the settlement in other proceedings. Trump’s new lawyers say that this part is prior restraint:
In addition, Rule 11 and inherent authority cannot support an order forbidding the President, private Plaintiffs, federal agencies, and sweeping categories of associated persons from “referring to” the Settlement Agreement or invoking it in any future judicial, administrative, regulatory, arbitral, or other official proceeding. D.E.106 at 47. That unprecedented injunction regulates future speech and advocacy, is unconstitutional, and is patently unlawful.
The Sanctions Order’s command is a content-based prior restraint. It singles out one subject—the Settlement Agreement—and suppresses one message, i.e., that the Settlement Agreement exists and may carry legal effect. Prior restraints carry a “heavy presumption” of constitutional invalidity, Bantam Books, Inc. v. Sullivan, 372 U.S. 58, 70 (1963), and constitute “the most serious and the least tolerable infringement on First Amendment rights.” Nebraska Press Ass’n v. Stuart, 427 U.S. 539, 559 (1976). They require findings that the threatened harm is “both great and certain and cannot be mitigated by less intrusive measures.” CBS, Inc. v. Davis, 510 U.S. 1315, 1317 (1994) (Blackmun, J., in chambers). The Sanctions Order identified no threatened unlawful speech, no resulting harm, and no reason narrower relief would be inadequate. Indeed, it did not address the First Amendment at all.
And while this is the strongest argument, that doesn’t make it actually strong. The order does not stop Trump or his lawyers from talking about the settlement. They can post about it on Truth Social, complain about it on Fox, write a book about it. What they can’t do is walk into another court, agency, or arbitration and invoke a settlement that a federal judge just voided due to the lack of adverse parties. The only “speech” being restrained is further fraud on the court.
But the fact that Trump bothered to make this argument at all is the real tell. You don’t fight for the right to invoke a settlement in future proceedings unless you’re planning to invoke it in future proceedings — which is precisely what Blanche has spent months assuring senators would never happen.
Then, even as Blanche was negotiating with Cornyn and Tillis to get them to vote in favor of his AG nomination by promising not to create the fund, on Saturday, Trump said hell yes he intends to make use of the fund — and added that he’d like Congress to pass a law establishing the same fund as a backup, in case the courts kill the one he negotiated with himself:
That’s a Trump post on Truth Social saying:
If Senators Cornyn and Tillis, both upset because I wouldn’t Endorse them (they lost, and quit, respectively!), aren’t going to approve Todd Blanche, one of the most respected professionals, according to everybody, in the Country, to be the United States Attorney General, then I will keep Todd as Acting A.G., and push hard to get the Anti-Weaponization Bill, which takes care of those who have been so badly treated by the Crooked Joe Biden (and Obuma!) Administration (I get nothing, although I was treated horribly!), PASSED. Todd Blanche was a voice of reason! It will immediately be back on the table, and I will get it done. Thank you for your attention to this matter! President DJT
Once again, Trump’s ability to shoot his allies in the back remains undefeated. Cornyn (especially) and Tillis have said that their hesitancy in approving Blanche had a lot to do with the weaponization fund. Blanche’s approval appeared contingent on a theatrical claim that the fund wasn’t going forward. To have Trump then come out and say “hell yes it’s going forward, even over the objections of Cornyn & Tillis” seems unlikely to appease those two Senators.
Tillis, for one, was not impressed:
That’s Tillis tweeting:
Despite comments as late as yesterday that the fund is dead, President Trump clearly intends to resurrect the payout pot for punks either by inappropriately establishing another bogus fund or pushing Congress to vote for a bill that the majority of Republicans in the Senate would be against.
It’s unfortunate that Todd Blanche, who I consider qualified for the job, will not be confirmed because of this reversal. Hopefully, we can resolve this by Tuesday.
It’s been quite clear that Trump has no intention of ever letting go of this slush fund, but it’s classic Trump that he couldn’t even keep his mouth shut long enough to pretend the fund wasn’t going forward and to get Blanche approved.
Which then brings us to the last bit that came out late Sunday: Blanche tweeting out that the DOJ had finally decided to “rescind” the weaponization fund in writing. Something he had refused to do over the past few weeks.
There are two separate letters there. The first “rescinds” the fund, while making an obnoxious snarky comment about “frivolous lawsuits” filed to challenge it:
A. The Attomey General’s May 18, 2026 Order establishing the Anti-Weaponization Fund (“Fund”) is rescinded and shall have no force or effect.
B. For the avoidance of doubt, nothing in this Order is intended to contradict or otherwise be contrary to prior representations by the Department of Justice that the Fund is not operative. No Members were appointed; no funds were transferred; no process for receiving claims was established; no claims were paid. Nevertheless, several frivolous lawsuits have been filed challenging the Fund, and at least one court has declined to dismiss those claims as moot. This Order establishes, beyond any doubt, that there is no Fund…
If the cases were “frivolous” then why would you need to rescind the agreement?
The second, with snide shots at Congress for not believing his statements, says a bit more on the matter:
Department of Justice Rescinds Anti-Weaponization Order and Addresses the May 19 Release
The Department of Justice today announced that the Acting Attorney General rescinded the May 18, 2026 Order that established “The Anti-Weaponization Fund.” Although the Acting Attorney General has repeatedly advised Congress through testimony, including under oath, as well as in written responses, that the Fund is not moving forward, and the Department has repeatedly represented to district courts that the Fund is not moving forward, today’s Order officially rescinds the May 18, 2026 Order.
Further, the May 19, 2026 Order regarding a mutual release of claims applies by its terms only retroactively. Additionally, the terms “Lawfare and/or Weaponization” in that Order were already defined in the Settlement Agreement as the use of government power “to target individuals, groups, and entities for improper and unlawful political, personal, and/or ideological reasons.”
The Acting Attorney General stands by all of his July 15, 2026 Senate Judiciary Committee testimony, including answers to questions asked by Senator Cornyn related to the scope and applicability of the May 19, 2026 Order. A transcript of the Acting Attorney General’s testimony is attached. He incorporates that testimony by reference, and restates that it is DOJ’s interpretation that the Order only has any effect, including on any release of claims, on the named parties in the lawsuit referenced in the Order.
This all seems pretty obviously designed to get Cornyn and Tillis off his back by saying “see?!? I’ve now put it in writing. Now shut up and vote to make me Attorney General.” And, who knows, it might work.
That’s even though nothing in there actually answers Cornyn’s actual question discussed above. The DOJ cannot single-handedly go back on this agreement. It was, we are told (including in the appeal from Trump on Friday), a private agreement between two distinct parties: Donald Trump and the Justice Department. As such “rescinding” it from just the DOJ side is meaningless since, as Blanche himself admitted to Cornyn, Trump can still try to claim in court that the settlement is valid.
And given Trump’s public statements about why he needs the fund, as well as his lawyers’ statement in the appeal that it tramples his rights to order him not to claim the settlement is valid in court… you’d have to be pretty slow not to pick up on what Trump intends here.
Also, notice what didn’t get rescinded. The May 18th order (creating the slush fund) is what he pretends is gone, but the May 19th order, which clears Trump, his family, and their businesses from IRS audits over past business practices remains. Blanche pretends that saying it “applies by its terms only retroactively” is a meaningful limitation, but that was always obviously the case. By all accounts, the Trump family books are a mess, and they likely owe the IRS a ton of money. And the Blanche announcement makes it clear that remains as is.
So, to sum it all up, Blanche has rescinded an order based on a settlement he can’t single-handedly rescind, to appease Senators whose votes he needs, while the President has made it clear he’s going to move forward with his fund no matter what.
The simple reality here is that we have a crooked President with his hand-picked henchman trying to do the bare minimum to appease a compliant GOP-run Senate, while the President himself can’t keep his own mouth shut long enough to even put up the pretend facade that there’s no corruption going on here.
Under any other President, this would be an impeachable offense. Here it was just another weekend under the Trump Presidency.
The US startup Harbinger has joined American Rheinmetall and other partners to crack the Defense Department's EV code with an autonomous, flat bed hybrid designed for crew-free resupply in danger zones.
The post It’s Official: US Army Dives Into The EV Tank appeared first on CleanTechnica.
The Public Safety Scanner Tool, or PSST, sends automated alerts of urgent scanner traffic in real time and will be free for newsrooms Open Vallejo, our nonprofit newsroom in the Bay Area, has kept a police scanner running for most of its six years, as many local newsrooms do across the country. It is how our reporters catch a shooting before the press release goes out, a chemical leak before…
As Professor Bebop, Rogers shared a unique mix of rhythm & blues, jazz and soul while delighting listeners with his inimitable jive announcing style.
The post Dave Rogers, longtime volunteer WTJU host, dies at 79 appeared first on Current.
This week’s action: Identify a topic, experiment or challenge that could benefit from collaborating with researchers.
Start here: Identify a topic or challenge that you could use an outside perspective on. This could be an area of coverage that’s lacking, an audience you’re struggling to engage, an impact metric you’d like to study or an experiment you’ve been meaning to try.
Take action: List three benefits the collaboration would bring to your organization. Perhaps a collaboration outside of journalism will help you surface new information, offer new frameworks or approaches, or provide extra bandwidth and support.
Reflect: Collaborating with non-news experts or researchers is a two-way street — both parties should inform each other’s work. What insights or benefits would a collaborator gain from exploring this topic or challenge with you?
The post Local news doesn’t have to address community challenges alone appeared first on American Press Institute.
Welcome to our August Need to Know Special Edition series on partnering with non-news experts.
The challenges local news faces — news deserts, burnout, community trust and connection — don’t exist in a vacuum. They are influenced by and interlock with trends in communities and the country, such as economic and technological change, demographic and educational trends, and political polarization. And for each challenge, there is knowledge from other sectors that can help:
Bringing ideas from outside of journalism into the conversation about how journalism evolves organizationally and how it serves communities has become a mark of our programming at the American Press Institute. We bring non-news experts to our Local News Summits to provide new perspectives on the same challenges, inspiring news leaders with new frameworks, approaches and knowledge.
This month, we’re outlining three frameworks for how journalists can engage with non-news experts and researchers to help solve a problem, rethink a beat or collaborate on an experiment. This week, we’ll start with you thinking about how partnering with a non-news expert might help you.
The post Collaborate with non-news experts appeared first on American Press Institute.
Chances are you’ve already heard about what happened in Twin Falls on Saturday. A man went on a shooting spree in Twin Falls, Idaho in the parking lots of an In-N-Out Burger and a visitor center. Among the wounded was a cab driver who was sitting in his Tesla at ... [continued]
The post Staying Safe While Charging appeared first on CleanTechnica.
I have a neighbor who often wears a T-shirt that says, “We The People Are Pissed Off!” I suspect it is meant to indicate his support for MAWA (Make America White Again), but it could just as easily apply to many in the US electorate today, as grocery prices, medical ... [continued]
The post Climate Hushing Strategy May Cost Democrats At The Polls appeared first on CleanTechnica.
It’s not unusual for things to get tense on a Sunday morning news program. Politicians are invited on to discuss critical issues and are often pressed to explain divisive positions. […]
The post Dana Bash did her job — and RFK Jr. didn’t handle it well appeared first on Poynter.
Europe’s public hydrogen refuelling network is not growing, but shrinking slightly. It is, however, changing shape. The part built for passenger cars is shrinking, while a new generation of dual-pressure stations is carrying 700-bar capability forward into the bus and truck layer. Germany closed 36 first-generation hydrogen refuelling stations during ... [continued]
The post Europe’s 700-Bar Hydrogen Network Is Closing And Reappearing appeared first on CleanTechnica.
What a vibrant marketplace of ideas
Sign up here to get an email whenever First Dog cartoons are published
Get all your needs met at the First Dog shop if what you need is First Dog merchandise and prints
Hydrogen fuel cell buses and trucks are zero emission electric vehicles, but they need green hydrogen to complete the sustainability loop.
The post Fuel Cell Trucks And Buses Are Still Coming, And They Still Need More Green Hydrogen appeared first on CleanTechnica.
Authorities issue alert saying indoor spaces without airconditioning are ‘dangerous’ as temperatures pass 40C for fifth straight day
South Korea has registered its highest temperature since records began more than a century ago, as authorities warned people to “immediately” stop all outdoor activities in affected areas.
The south-eastern city of Yangsan saw the temperature reach 42.5C on Sunday afternoon, the Korea Meteorological Administration (KMA) said, the highest temperature reported since modern weather observation began in 1904.
Continue reading..._Castello_Plan_1660.jpg){kind=link}