Two school days. That’s all it took.

In 2024, I chaperoned field trips two days in a row, for two different grade levels, and came back to roughly 450 ungraded assignments.

I knew what to do, I’ve done it before, mark them credit or no credit and move on. Students get something out of that. They did the practice. But if any of them were practicing it wrong, nobody catches it, nobody tells them, and the misunderstanding rides along into the next unit.

That pile of work led me to build an AI grading assistant. And this past April, I removed its most automated feature: the one that could return an AI-generated grade and comment to a student before I had reviewed it.

Building that feature was easy to justify. Removing it taught me which part of grading a teacher can’t hand off.

Most of what students turn in to me isn’t a clean essay. I teach engineering, and my students submit designs, schematics, code, and photos of physical work. That’s part of why many teachers I know still don’t grade with AI. They’ll use it to scaffold a unit or soften an email to a parent, but grading with it usually means pasting work into a chatbot one assignment at a time, which is so slow I can grade it faster myself. So, I built my own tool.

I teach mechatronics, and if mechatronics teaches you anything, it’s that efficiency matters. You optimize the system and eliminate friction. I brought that mindset to the product I built, and the logical endpoint was auto-return. The AI could evaluate the work, draft the grade and comment, and send it back to the student without another click from me, late submissions included. I had spent hours tuning it to grade against my assignment, handouts, instructions, and rubric.

Then a student came up to me one day, happy about the encouraging comment on an assignment. The comment had motivated him to redo the work and resubmit it.

When AI Takes Control

The problem was that I didn’t write the comment. I hadn’t even seen it.

If it had passed by my eyes and I’d confirmed it, edited it, or decided it belonged there, this would be a different story. But in that moment, the student thought the encouragement came from me, and I wasn’t actually in the exchange.

Nothing about the feedback was inaccurate. That almost made it harder to explain. After more than two decades in a classroom, I couldn’t put words to what felt wrong. I just knew it did. The issue wasn’t whether AI could draft useful feedback. It could. The issue was whether a student should receive a teacher’s judgment when the teacher hadn’t made one.

So, I removed auto-return, and the automatic grading of late work went with it. What replaced it is a review dashboard: the AI drafts every grade and comment against my rubric and lays it out in front of me. I can edit, override, reject, or return the feedback in one pass. It’s still fast. But now my eyes and my judgment touch every grade before a student sees it.

That changed how I think about human review. It can’t mean glancing at a score and clicking approve. It must mean checking the student’s work against the rubric and owning the result.

The software can propose a judgment. It cannot own one.

Policy is starting to move the same way. New York City’s public school guidance now says AI must not replace educator decision-making, and other states are weighing rules on human review and student data. The rules will keep changing. The principle shouldn’t; a student’s grade needs a person who is accountable for it.

When I walked one of my administrators through the tool, what he liked most wasn’t the time savings. It was that it requires a rubric. Teachers write rubrics for big projects, but the daily, low stakes work rarely gets one, and that’s exactly the work that gets marked credit or no credit and never comes back with feedback. The trade runs both ways: students get clearer expectations up front and comments on work that used to get a checkmark.

He had two concerns, both fair. Parents and students should know when an AI-assisted tool is grading, so it belongs in the syllabus. And if a student contests a grade, the teacher should re-grade it by hand. We agreed the second should happen anyway, with or without AI. Humans make grading mistakes too.

My students know I built the AI tool. What they care about isn’t the technology. It’s whether the feedback is fast, the rubric is clear, and the grade is fair. A few times the tool has docked points for work it missed, almost always because a screenshot cut off the edge of the page or the writing was too faint to read. Those students came up to me, I looked at the work, and I gave the points back. I want that. A grade should be something a student can question.

What surprised me is that a student will challenge the AI long before he’ll challenge me. A kid will walk right up and say, “The AI got this wrong, I should have full credit.” That same kid won’t tell me, to my face, that I made the mistake and owe him ten points. Both of us can be wrong, but the machine is easier to push back on than the teacher, and that’s good for the student. The grade still passes through me. The draft between us just makes it easier to speak up. If a grading system makes students afraid to challenge the result, the system is wrong.

AI Grading Advice

If your school is wrestling with AI grading, start with the disclosure. Don’t say “AI may be used.” Say what that means: comments and grades are drafted by AI and reviewed by the teacher. Then answer the harder questions. Where does student work go? Is it stored, or used to train models? How secure is the platform, and has anyone independent reviewed it?

We are teachers, not graders. We grade, yes, but we also sit in IEP meetings, call parents, design lessons, and try to notice the student who is quieter than usual. If a grading assistant hands me back the hours I spent marking daily work, and I spend them on better lessons and better feedback, everyone wins.

But when a student asks, “Why did I get this grade?” the answer cannot be, “Because the system said so.”

It has to come from me.

I slipped under the wire and got a demo of “Character Limit” submitted to Steam just in time. Now for the stressful part: It’s live in Steam Next Fest.

In April, the development of Character Limit had reached a point where it could be tested by actual players. It did fantastically well at Dreamhack Birmingham, and I had also started doing testing on iPhone and iPad with Testflight.

However, later that month, registration for something known as Steam Next Fest was coming to a close. It’s a promotional festival held a few times a year, focused on getting players to try out demos for games that have yet to be released.

Since I want the game to be completed long before the busy fall iPhone season, that meant the next Next Fest I could take part in was in the middle of June. I figured there was plenty of time to get the Dreamhack demo to a better and more robust state, so it could be downloaded to players’ computers to try out.

But, since WWDC was also happening the week before Next Fest, that meant I would face a busy time here at AppleInsider, which would impact any development.

So, the decision was made to leave the Testflight build ticking along on iPhone, and to return to it when the game is in a much better place. Instead, I allocated my time to focus just on the Mac and PC demo.

It was both a very good and very stupid decision.

Fixing the demo

Getting the demo working as perfectly as possible for Next Fest wasn’t just to ensure the game got more promotion through a major digital storefront. It also forced me into shoring up the demo so it could be used by the actual public.

So far, the game was made with my hardware and setups in mind. To make it work for practically any player in the world, I had to make sure anything that would break or be an edge case was handled gracefully, since any combination of hardware would be thrown at it.

Cue a few weeks of fixing bugs in number distribution, making an actual tutorial screen, and localization foibles that I had missed before.

It was then sent to some colleagues and friends for testing. And I’m glad that I did.

An edge case I didn’t consider, and the fix. Image credit: William Gallagher

Advertisement

William Gallagher of this very site tried the game demo out. Partly because I had to work out how Xcode notarization worked, which was tricky but not difficult.

It turns out that William has an insane display ratio because he uses an ultrawide screen. Due to the way the game functioned, playing full-screen meant you missed the bottom and top sections of the interface.

For a word game, it helps if you can see the letters you’re turning into words. Or even just see the Quit button on the main menu.

This was fixed quickly by making a floating section of the interface that was the correct aspect ratio, that would be visible in the middle of the screen when viewed by an ultrawide. All the content of the game would then be put into that box, similarly anchored to the center.

This worked surprisingly well and was practically the last stumbling block for the demo itself.

While not everything has been fixed, it’s just got enough “done” to be usable as a proper demo.

As a consequence, I have many changes to transfer from the demo into the full game. It’s work I would’ve had to do anyway before release, so doing it all now rather than later was a good move.

Steam pains

While making the game and working with Apple’s Notarization system is one battle, Steam is quite another.

The tricky process of setting up a Steam page for the game has already been done. While you can set up a separate store page for a demo, you could also just add the demo download button to the main sales page.

There are reasons to go either way, but I went with the one main sales page approach. Partly because it would focus traffic onto that one page, partly from laziness.

However, for some reason, Steam requires you to submit effectively everything you would need for the separate store page for the demo, even if it doesn’t exist. That includes the text for the page, which I had to write.

Other elements make sense, like platform specifications, capsule art images, and the name, since they would be used elsewhere.

Then it came time to upload the demo, which led to a multitude of problems.

The first issue was administrative, as I previously had to transfer the game from one Steamworks account to another, due to having signed up wrongly to begin with.

A personal account and one owned by a limited company differ in many ways, which meant a second account had to be made and the store listing transferred over.

It turns out that you can transfer the game as one listing, but the demo is a separate listing and doesn’t necessarily get transferred over with the main title. I spent so long failing at uploading the demo because of this missing element, and then I had to wait most of a week for it to transfer.

There’s a graphical SteamPipe interface for Windows, but you’re stuck with the Terminal on macOS for some reason…

The second problem is Steam’s uploading system. There is an option to do so from the website, but there’s no documentation explaining how to use it properly, so it was a no-go.

The usual way is a SteamPipe upload, using the Steam SDK. This is a process that involves a ton of work, including installing a command-line version of Steam, constructing a build and an upload script, and then setting the launch options.

I understand fully that this is a system designed more for the big developers with massive teams and people who truly know what they’re doing. For a first-time indie, this is massively intimidating.

That is, until you realize that the SDK download includes a graphical interface that does a lot of the heavy lifting for you. Except it’s Windows-only and not available on the Mac.

I had a Windows PC nearby, so it wasn’t a difficult procedure. But it is disheartening that there’s not a macOS graphical interface available from Valve.

After getting the game demo uploaded, both it and the store details were submitted for review. The demo was apparently fine and dandy, but the store section was not.

Apparently, while you’re encouraged to make the images of the capsule art be different so that players can tell the demo apart from the main game, you also cannot add more words to the art other than your game name.

They objected to me using the words “Demo Edition” in the art. Also, my logo was being covered by a demo corner banner that wasn’t really mentioned previously.

Sure, my mistake for not fully understanding. It meant some tweaks and a few more days of waiting for a re-review and eventually being accepted.

On June 9, the second day of WWDC, Steam’s email confirmed that the store page met its requirements, and that the demo could be published. That evening, Character Limit became available to play in demo form on Steam.

Happiness and trepidation

I have written before about how Steam is a big deal for me. As a gamer with a Steam account value that could buy a reasonably priced car at current prices, I am very familiar with Steam,a nd have used it for many, many years.

Registering and having a Steam listing for something that I made was an emotional experience. Just that hit hard.

I didn’t expect putting the demo live to hit even harder.

I can now open my Steam library on my Mac and see “Character Limit Demo” on the long list of (mostly unplayed) games. I can now click it and open the game on my Mac or my PC.

Seeing it on there did actually make me weep a little bit. A stupid little game that I have been noodling on for most of a year is now in an application I regularly open up.

Seeing the demo in Steam was a big deal.

It’s probably a feeling that authors get when they see their work on Amazon or in their local bookstore. Except it’s a game and I can’t walk into a physical location and hold the game in amazement.

After walking around the block for some air, I checked it ran OK, and it does.

This was an extremely big and happy moment for me. But it was cut short with the daunting realization that this isn’t the finish line.

The rest of the game has to be put into place for the final version. Research into more languages and modes for future updates, too.

Then there’s the ever-continuing slog of marketing and promotion before the game’s release. That’s an inevitability.

But, more immediately, there’s the fear that the demo will not be well received by anyone. The last thing anyone creating anything needs is for people to complain about your baby, even if it’s entirely justified.

Writing on the Internet for over a decade certainly gives you a thick skin for criticism of your words. But this is a level of apprehension and worry that I’ve not had to deal with in a very long time.

That’s not even taking into account the eventual battle to get the finished product in the App Store. That’s next.

Here’s hoping Gabe Newell’s money-generator audience is somewhat kind during this week’s all-important Steam Next Fest.

Xbox is preparing to shut down or sell at least three of its studios — Double Fine, Ninja Theory and Compulsion Games — according to reports from The Verge and Bloomberg on Monday. 

Ninja Theory employees were informed on Monday that the studio would be closing, according to The Verge, but the team is attempting to find a buyer that can keep them operational. Ninja Theory is the studio behind the Hellblade series and it was featured in the recent Xbox Summer Game Fest showcase, revealing a new entry due in 2027.

Double Fine is the legendary studio behind the Psychonauts games, Brütal Legend, Broken Age, Keeper and all manner of LucasArts adventures, founded by Tim Schafer and friends in 2000. Double Fine leaders are in active negotiations to buy themselves back from Xbox rather than be closed altogether, Bloomberg said. 

Compulsion Games is in a similar position, according to the report. Compulsion is the Montreal studio behind the uber-stylish games Contrast, We Happy Few and South of Midnight, the latter of which came out in April 2025.

Bloomberg reported that several other studios under the Xbox Game Studios banner are also negotiating for their futures and are at risk of being shut down. As it officially stands, the Xbox Game Studios umbrella covers dozens of studios, including Arkane, Bethesda, Halo Studios, id Software, Obsidian, Playground Games, ZeniMax and Activision Blizzard King. We have contacted Xbox for clarification on the reported closures and buyout talks.

Microsoft’s modern game-studio acquiring spree kicked off in 2018 with the purchase of Undead Labs, Playground Games, Ninja Theory and Compulsion Games, plus the formation of The Initiative. This momentum continued building at a concerning pace in 2020 and 2021, when Xbox acquired eight more studios under ZeniMax Media, including Arkane, Bethesda and id. Everything came to a head in 2022 with the announcement that Xbox planned to purchase Activision Blizzard for $69 billion, the largest acquisition in video game history. That deal finally went through at the end of 2023, after a lengthy battle with regulators.

Since the acquisitions, Microsoft has enacted multiple massive rounds of layoffs affecting thousands of employees in its gaming division, and also shuttered well-known studios, including The Initiative. Unfortunately, Double Fine, Ninja Theory, Compulsion and the other at-risk Xbox studios are in good company.

Longtime Xbox division head Phil Spencer stepped down this year and was replaced by new CEO Asha Sharma, alongside other executive-level changes. Also on Monday, Xbox Game Studios head Craig Duncan left the company; he first took the role in October 2024. Employees across Xbox are bracing for more layoffs in 2026 after an ominous public memo from Sharma dropped in mid-June, just as the glow of Summer Game Fest fully faded.

from the America-isn’t-your-playset dept

What’s most disturbing about Trump’s “Restoring Truth and Sanity to American History” executive order isn’t its fully-blinkered, jingoistic take on American history where America does no wrong and is almost always white right. I mean, that’s pretty awful on its own, but it’s the flip side of pretending whites do no wrong: pretending any victims of whites or any contributors of other races/colors/creeds simply don’t exist.

It’s the sort of thing dictators do. It’s the sort of thing novelists write about. It’s the sort of thing we Americans used to be able to criticize wholeheartedly because our country would never stoop so low as to rewrite history to please whoever was currently in power.

We no longer have the high ground. But we might be able to start making our way back to the top of the hill. It’s not because Trump et al are getting better or smarter or simply a bit less hateful. It will be because the courts are doing what they’re supposed to be doing: slowing this budding fascist’s roll.

A federal judge in Massachusetts has ordered the Trump administration to reinstall displays it removed from National Parks sites over the past year as part of a crackdown on diversity, equity and inclusion (DEI) content and climate change information.

[…]

Advertisement

Judge Angel Kelley sided with the challengers on Friday, finding that the federal government’s action “sets a dangerous precedent of censorship and sanitization” while undermining the “integrity” of the National Parks system.

The decision [PDF] leads off with the ideal this nation and its national parks are supposed to embody…

Often referred to as “America’s largest classroom,” National Parks serve in that spirit by telling the stories both of those who write history and those who go unheard. The beauty of history is the unvarnished storytelling of a time gone by and the delivery of undeniable truths. The Government’s stewardship of these park sites thus carries a responsibility to present history in full rather than in favored fragments.

… before detailing the hideous destruction being perpetrated by the Trump administration:

Unfortunately, the Government has disregarded these principles. Under the guise of promoting American dignity, this Administration seeks to share a limited history by ordering the removal of all signs, displays, and interpretive exhibits at National Parks that do not align with its preferred narrative, thereby telling half-truths. In recent months, the Government has torn down exhibits in Philadelphia’s Independence National Historical Park memorializing the legacy of people enslaved by the country’s first President; removed signage detailing climate threats at Fort Sumter in South Carolina, one of the most environmentally endangered sites in the country; and wiped away descriptions of history and science at countless National Parks across the United States. Not only does this undermine the integrity of the National Parks; it sets a dangerous precedent of censorship and sanitization.

Dozens of instances of censorship and erased history are listed in the lawsuit. And it’s all the sort of thing you’d expect from this administration:

By the date of the filing of this action, Defendants had removed dozens of signs related to climate change, civil rights, and diverse communities.

[…]

In addition, Defendants have removed multiple signs involving slavery, abolition, immigration,
labor, women’s suffrage, and civil rights…

The court says the order violates the law repeatedly. Not only does it steamroll existing laws governing the National Park Service and its congressional oversight, it fails to justify its own existence with even the briefest nod to serving the public’s interest. Most damningly, the executive order ignores the facts in favor of pushing the administration’s preferred version of US history:

[T]he Order fails to rationally connect any facts to the action taken. It claims that the removals will “restore Federal sites . . . to solemn and uplifting public monuments that remind Americans of our extraordinary heritage.” However, the Order fails to explain how unearthing and displaying the historical contributions of marginalized groups detracts from celebrating “our extraordinary heritage.” Indeed, the NPS’ purpose in installing these materials in the first instance was to attract new audiences to National Parks by celebrating diverse experiences.

In other words, this executive order is basically just a Truth Social rant pretending to be a lawful directive.

[T]he Order fails to rationally connect any facts to the action taken. It claims that the removals will “restore Federal sites . . . to solemn and uplifting public monuments that remind Americans of our extraordinary heritage.” However, the Order fails to explain how unearthing and displaying the historical contributions of marginalized groups detracts from celebrating “our extraordinary heritage.” Indeed, the NPS’ purpose in installing these materials in the first instance was to attract new audiences to National Parks by celebrating diverse experiences.

The order concludes by using Trump’s self-serving rationalizations against him, which is exactly the sort of thing I’d love to see more of in future court orders rejecting this administration’s fascist advances:

Because Defendants deemed it important to strip the parks of these undeniable truths in anticipation of the 250th Anniversary of our great Nation, it is equally important that our shared history be honestly told and fully restored by the 250th Anniversary to properly honor the remarkable achievements of the United States.

LOL. Stick that in your White House lawn MMA fight, you asshat. Restore what’s been destroyed, says the court: that would be “truth and sanity,” rather than this steady stream of atrocities this administration continues to inflict on the US.

Filed Under: bigotry, censorship, erasing history, interior department, national park service, trump administration

Huawei has unveiled HarmonyOS 7 at its developer conference in China.

The update introduces a redesigned interface with glass-inspired visual effects and new AI-powered features, some of which look quite similar to Apple’s often criticised Liquid Glass design that continues in iOS 27.

The company also claims there is a noticeable performance boost over the previous version. The update is set to roll out across Huawei’s ecosystem of smartphones, tablets, PCs, wearables and smart home devices.

The U.S. Department of Justice announced Friday that it has seized the CFAKE.com and SOCFAKE.com websites, which allegedly hosted nonconsensual AI-generated nude images and videos of women, in what appears to be the first publicly announced domain seizure under the TAKE IT DOWN Act.

According to the DOJ, the sites shared sexually explicit digital images, or deepfakes, depicting politicians, celebrities, athletes, musicians, and even royalty from multiple countries.

“According to the probable cause affidavit supporting the seizure warrants, the digital forgeries were made to appear to be sexual images of famous women, including politicians, first ladies of multiple countries, royalty, journalists, television presenters, athletes, entertainers, and others,” reads the DOJ announcement.

A deepfake is AI-generated or AI-manipulated media that depicts a person saying, doing, or appearing in ways that never occurred. Deepfake images and videos can be created from existing photos, videos, or audio recordings and are commonly used to generate nonconsensual nude content, impersonation scams, phishing attacks, and cryptocurrency fraud.

The CFAKE.com and SOCFAKE.com domains were seized on Thursday by the DOJ and Homeland Security Investigations after a federal judge found probable cause that they were being used to violate the TAKE IT DOWN Act.

The domains now display a seizure notice stating they were taken offline pursuant to a seizure warrant as part of an operation involving the US, Italy, and France.

“THIS DOMAIN HAS BEEN SEIZED by the United States Department of Homeland Security, Homeland Security Investigations (HSI) New Jersey Field Office pursuant to a seizure warrant issued by the U.S. District Court for the District of New Jersey as part of coordinated law enforcement actions by HSI, French National Police, the Paris Prosecutor’s Office, Italy’s Polizia di Stato – Postal and Cybersecurity Police, United States Department of Justice’s Computer Crime and Intellectual Property Section and the United States Attorney’s Office for the District of New Jersey for violations of 47 U.S.C. § 223,” reads the seizure banner on the websites.

“The TAKE IT DOWN ACT (47 U.S.C. § 223) prohibits the nonconsensual publication of intimate imagery and digital forgeries (i.e., deepfakes). Violators are subject to fines, imprisonment or both.”

The investigation began after Italy’s Postal and Cybersecurity Police alerted US authorities to the websites.

According to Italian media reports, investigators opened an inquiry in October 2025 after receiving complaints regarding AI-generated sexually explicit images depicting women from politics, sports, entertainment, and other public-facing professions.

Italian authorities later obtained a court order blocking access to the websites within Italy while continuing their investigation. The DOJ says evidence gathered by US law enforcement was later shared with French authorities.

French prosecutors and investigators then conducted an investigation that led to the arrest of a suspect in Nice, France, on June 10, along with the seizure of cryptocurrency allegedly connected to the operation.

The bipartisan TAKE IT DOWN Act was signed into law in May 2025 to combat the spread of nonconsensual imagery, including AI-generated deepfake pornography. The legislation was championed by First Lady Melania Trump as part of her “Be Best” initiative.

The law makes it a federal crime to publish sexually explicit altered images depicting identifiable individuals without their consent. The legislation also requires online platforms to remove reported intimate images and deepfakes within 48 hours of receiving a valid request from a victim.

“These domain seizures mark a significant victory in the fight against deepfake pornography,” Acting Attorney General Todd Blanche said in Friday’s announcement.

“The TAKE IT DOWN Act, championed by First Lady Melania Trump, gives us the tools we need to combat the abuse and exploitation of women and children through these fabricated images.”

The law was previously used against an Ohio man who pleaded guilty to charges related to creating AI-generated sexually explicit images.

However, the seizure of CFAKE.com and SOCFAKE.com appears to be the first publicly announced use of the law to target websites allegedly used to distribute deepfake pornography.

Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.

The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.

Get the whitepaper

Over the weekend, Google CEO Sundar Pichai faced a small revolt when he delivered his commencement speech at Stanford University, where he earned his graduate degree in materials science and engineering. About 200 students from the graduating class reportedly walked out, while others loudly booed the tech executive.

The focus of the protest was Google’s defense ties — including Project Nimbus, the controversial $1.2 billion contract, shared with Amazon, to provide cloud and AI services to the Israeli military, as well as its relationship with the U.S. Immigration and Customs Enforcement agency.

Student signs included phrases like “ICE SPIES WITH GOOGLE AI” and “GENOCIDE RUNS ON GOOGLE,” as well as “FREE FREE PALESTINE,” a press release associated with the protest notes. Students also waved Palestinian flags and shouted “free Palestine,” online video of the protest shows.

“We are walking out because we refuse to glorify the corporations that fuel this violence and exercise our power to choose differently,” a statement associated with the protest reads.

The walkout was organized by a number of campus activist groups, including Stanford Students for Justice in Palestine, No Tech for Apartheid, and Tech for Liberation. TechCrunch reached out to Google for comment.

As the war in Gaza has raged, Google’s participation in Nimbus has drawn protests from both inside and outside of the company. In 2024, Google fired 28 workers for protesting the contract, although it has continued to suffer internal dissent over the issue since then. It was also recently criticized by the Electronic Frontier Foundation, which accused it and other companies of “choosing to look the other way” on Israel’s use of their services.

Project Nimbus also enjoys support from Amazon. Microsoft has also been criticized for its support of the Israeli military, although the company restricted the Israeli government’s use of its technology after an investigation found that its cloud services were being used to mass-surveil Palestinians.

The student protest also drew criticism from business leaders online. Vinod Khosla, the billionaire co-founder of Sun Microsystems and one of Silicon Valley’s most prominent venture capitalists, posted on X that the protest was “biased, idiotic, short-sighted and very selfish,” adding that it was selfish because the students “ignored the bottom 3 billion people on this planet that could benefit from AI and they are worried about their misinformed selfish self-interest.”

Pichai’s appearance at Stanford is part of a broader pattern. Speakers at college graduation ceremonies around the country have faced boos when they have attempted to get outgoing college students excited about AI. But rarely has student animus been as targeted as it was with Pichai, directed not at AI hype, but at the specific business decisions made by the company he leads. In general, young people seem to believe that AI is threatening their employment opportunities and may be ruining other parts of society as well.