By Phil Baker

For the past several months, our developers at Neil Young Archives (NYA) have been working on major revisions to our iOS and Android apps, including a much-improved interface and an entirely new feature, Neil Young Radio.

We know how frustrating it can be for Android users to wait weeks or months for an app that's already available on the iPhone. Often, that's because developers prioritize iOS, developing the two versions sequentially rather than in parallel. With so many Android users among our worldwide members, we were determined to avoid that by developing both versions simultaneously, with the goal of releasing them on the same day.

Unfortunately, Google had other plans. It's now been almost two weeks since our new iOS app became available, while our Android app is still awaiting approval. Both were submitted to Apple and Google on the same day. Apple reviewed and approved our app in less than 24 hours; we're still waiting for Google.

And we're hardly alone. Other Android developers are reporting that approvals that once took hours or a few days are now stretching into weeks, often with little explanation.

In September, Daniel Gultsch, developer of the established Android messaging app Conversations, complained that an update to his 12-year-old application had been sitting in Google's review process for more than a week. Other developers responding on Hacker News described similar experiences, some reporting waits of two weeks or longer.

Earlier this month, one developer reported that an update adding Android Auto functionality remained under review for 21 days before finally receiving approval. Attempts to get help from Google support produced little useful information. Meanwhile, the corresponding Apple application had already been approved.

Android Auto appears to be a particular bottleneck because Google requires an additional review of applications designed to operate in vehicles. That makes sense from a safety standpoint, but it hardly explains why approvals should take weeks or why we developers receive so little information about the delays. We have repeatedly submitted requests for expediting our application with little results. We asked again two days ago and they acknowledged our request, but nothing yet as of today, October 28.

Google controls access to an enormous worldwide audience of Android users. With that control comes a responsibility to provide developers with a timely, predictable, and transparent approval process. Instead, developers are left waiting and customers are left frustrated.

If Google wants developers to continue investing in innovative Android applications, it needs to stop being the obstacle standing between developers and their customers.

Things aren’t looking great.

By Gary Marcus — via Substack

In early March 2023, I had a debate with Keith Teare, hosted by the organization Open to Debate, on the question of whether ChatGPT would do more harm than good. I think about that question a lot, nearly every day.

Right now, the negative argument against generative AI—the ledger of bad things—is quite serious. We have nonconsensual deep fake porn, a rise in phishing scams, wholesale disinformation, cognitive surrender, threats to democracy from the subtle influence of governments on LLMs, not to mention the use of generative AI as a surveillance tool. And the effects of all that IP theft on artists and writers, and potential negative effects more broadly across the employment market (though, mercifully, these have been more modest than some people expected).

Some ominous signs

By Gary Marcus — via Substack

In May, at the peak of tokenmaxxing, people were talking about OpenAI IPO’ing at a trillion dollars, Anthropic at two trillion dollars. Maybe that will still happen, but it probably shouldn’t.

Tokenmaxxing died, Chinese open-weight models forced price wars, there was some hinkiness around some of projections OpenAI’s revenue and the so-called “rogue AI” (really just bad software) fiasco hadn’t yet made the news. With the errant, and possibly criminal, nature of poorly controlled agents now front and center, potential liability is clearly becoming a major new issue. LLM’s might well someday be renamed Large Liability Modules.

A significant new copyright decision by a U.S. appeals court also just undermined the generative AI companies:

The real news here isn’t the result; it’s what we were not told.

By Gary Marcus — via Substack

OpenAI’s massive new math drop:

We’re releasing a broad range of new mathematical results produced by an internal frontier model.

We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and…

— OpenAI (@OpenAI) · on X

[This essay was written in extreme haste before a very long wifi-less flight; please forgive typos.]

Part One: My take

“The most powerful tool is the power to dissolve corporations that engage in repeat lawbreaking.”

By Gary Marcus — via Substack

An interview (my questions in italics) with Zephyr Teachout, attorney, author, and law professor that could not be more timely:

What’s your background, legal and otherwise, and what makes you interested in these issues?

I’m a law professor at Fordham Law School, I’ve drafted tech legislation that has been enacted and plenty that hasn’t, I worked for the New York Attorney General investigating civil violations of other companies, and I have over two decades of experience in studying tech policy and law. I joined the current AI debate because I saw way too many people treating repeat evidence of lawbreaking as wholly unreachable by current law.

Is there any reason to think that OpenAI has broken any laws that (at least in another time) might be enforced? Which and why?

Another Trump enabler pays the price

By Paul Krugman — via Substack

One iron law of current politics is that nobody who associates themselves with Donald Trump emerges with a shred of credibility. For some, that’s not a big loss — it’s not as if anyone respected Pete Hegseth even before he decided that the key to military success in an age of drones and AI was more testosterone. For others, however, the costs are real. I’m old enough to remember when Marco Rubio was widely considered a reasonable, thoughtful politician.

Right now, however, my candidate for the biggest loser is Scott Bessent, the Secretary of the Treasury. Bessent started out with significant reputation to lose: When Trump selected him, this was the headline in the Financial Times:

Jason Furman, Barack Obama’s chief economist, described him as a “credible Treasury secretary who has a real understanding of the global economy”.

But Bessent has now transformed himself into the Baghdad Bob of bonds.

OpenAI’s software didn’t just attack HuggingFace.

By Gary Marcus — via Substack

OpenAI’s software didn’t just attack Hugging Face.

It didn’t just attack a German web server.

It didn’t just attack Australia’s government servers.

And Australia apparently wasn’t the only country attacked by OpenAI’s software.

OPENAI: SOME WEBSITES INVOLVED IN MISALIGNED MODELS INCIDENT ARE OPERATED BY GOVERNMENTS, UNIVERSITIES, PUBLIC AGENCIES, AND OTHERS

— First Squawk (@FirstSquawk) · on X

Worse, in disclosing what happened, OpenAI has dragged their feet every along the way.

An open letter about how you could change the world, tomorrow

By Gary Marcus — via Substack

Dear President Trump,

Tomorrow is a chance for you to make history, when you meet with President Xi and talk about AI. As we both know, almost everybody’s expectations are low.

But you can surprise them! And as I suggested last week in The Economist, if you think about it right, it’s your chance to win a Nobel Prize. (Sam Altman thinks the same!)

Now, here’s the thing: most people are focused on a deal that you don’t want to make, around a slowdown. I get it. I happen to think a slowdown might be a good idea, especially if we made it now while we are still ahead — that could actually lock in our lead!. And I would caution you that if we don’t have a slowdown, really bad things might happen, and you would take a lot of the blame.

But if that’s not for you, it’s not for you.

Instead, what I propose is different. I am not saying that you should go for a slowdown, if your heart isn’t in it.

Instead, you can go for something that will surprise a lot of folks: you can go for collaboration — just like we did back in the Cold War with the Soviet Union, around things like space and smallpox. But this time around cancer and cybersecurity.

More people will tell you China would never want to collaborate.

But that’s just not true.

Sometimes, maybe dreams come true?

By Gary Marcus — via Substack

Earlier today as part of the UNGA (United Nations General Assembly) Digital Cooperation Event, which also featured Yoshua Bengio and the Nobel Laureate Maria Ressa, I gave some “framing” remarks:

Rome is burning and people are fantasizing about Skynet.

By Gary Marcus — via Substack

Minor breaking news (today):

Medium breaking news (yesterday):

Listener beware

By Gary Marcus — via Substack

We absolutely need to hold AI companies liable for the damage they cause. I wrote about this in my book Taming Silicon Valley, pointing to what many others have long written about it. I am in full agreement with Senator Josh Hawley on this tweet from earlier today, too:

So let me get this straight: AI CEOs now say the end of the world is near because of what their AI bots have been doing - but they want an antitrust exemption as a reward? How about instead we make them liable for whatever bad stuff their bots do

— Josh Hawley (@HawleyMO) · on X

But I am here today to call attention to a growing trend: people on the technolibertarian right calling for liability in AI (which is totally reasonable), but then using that as if it were an argument against regulation.

Take this interview with Joe Lonsdale, amplified last night by David Sacks,

Very partly

By Gary Marcus — via Substack

Remember, even as Trump shouts from the rooftops of Truth Social that he is violently opposed to regulating AI …

… he is in fact regulating AI, with a secret AI security framework for evaluating which “frontier” models can be released. Almost nobody knows what’s actually in it.

§

September 24, if played right, could change everything

By Gary Marcus — via Substack

I’ll get to the significance of September 24th in a moment.

But first, this would be wild if true:

I’m told WH aides are advising Trump to give a national TV address in the next 72 hours, declare certain aspects of the AI race an “imminent threat to humanity,” and announce regulations pausing the things Tech companies want paused. 1)

— David Shuster (@DavidShuster) · on X

It’s probably a long shot, but as I write this, AI stocks are falling,

By Karl Bode — Techdirt

I recently noted how California’s being told by the Trump administration that it can’t enforce any of its own telecom consumer protection laws if it wants to receive $1.4 billion dollars Congress already earmarked for broadband expansion under the 2021 infrastructure bill.

That means if California wants money it was already supposed to receive from Congress, it can’t enforce its own net neutrality law, telecom privacy law, affordability metrics, reliability targets, or effectively do anything telecom monopoly giants like AT&T and Comcast don’t like for the next 14 years.

More specifically, some extremely unpopular Trump NTIA revisions to the infrastructure bill’s Broadband, Equity, Access, and Deployment (BEAD) program included a little something called “Condition 50,” which would impose a 14 year moratorium on telecom consumer protection in the state. It’s just grotesque levels of corruption and regulatory capture, and it’s flown largely under the radar.

Tim Wu (Columbia), Lawrence Lessig and Yochai Benkler (Harvard), Jack Balkin (Yale), Zephyr Teachout (Fordham), and Barbara van Schewick (Stanford) have penned a joint letter to Gavin Newsom urging him to refuse to sign the state’s BEAD notice of award agreement while Condition 50 is still attached. Or, better yet, legally challenge the federal government extortion:

“California must decide by November 1 whether to sign the Notice of Award for its $1.42 billion BEAD grant. We urge the State not to sign while Condition 50 remains attached, with or without a side letter. The best course is to challenge Condition 50 in court before the deadline: California has strong arguments that NTIA lacked authority to impose the condition, but it has a realistic chance of winning only if it sues before signing the award. If the State decides not to sue, it should decline the award and make clear that California is ready to accept the funds as soon as the condition is removed.”

(Since the letter’s writing the deadline has shifted to November 2, I’m told). If California signs the letter with Condition 50 attached, the Trump NTIA will argue the challenge belongs in D.C. federal court under a very narrow standard of review, and that California knowingly accepted the condition. Such a case would be extremely difficult to win, the academics note.

Given the total collapse of federal consumer protection and regulatory integrity, AT&T, Comcast, and other giants would be free to rip off consumers — and undermine competitors — with impunity.

Buyer, beware of omissions

By Gary Marcus — via Substack

Gregory (Scotland Yard detective): Is there any other point to which you would wish to draw my attention?
Sherlock Holmes: To the curious incident of the dog in the night-time.
Gregory: The dog did nothing in the night-time.
Sherlock Holmes: That was the curious incident.

– The Adventure of Silver Blaze

Earlier today, I wrote “Keep an eye out, too, for the details of Anthropic’s S-1, and especially on how much they reveal (or not) about what has happened to demand in recent months, after the decline of tokenmaxxing and the increase in the popularity of low-cost alternative models.”

Minutes earlier (unbeknownst to me at the time) they had dropped their draft S-1. And (in some places) literally left out Q3:

Read more

NYA Thanks Gary Marcus!

The Guardian reports a man from western Washington state has been arrested by US prosecutors and charged with conspiracy to commit murder in a foreign country. He is the second suspect in the Canadian school massacre in which six children died. The Department of Justice said James Codey Bryant sent the alleged shooter money and what it calls "technical advice" in the six months before the attack.

Canadian police have identified the woman they believe is the shooter, Jesse Van Rootselaar, who had been telling ChatGPT about her plans. The "safety" people at OpenAI noticed this, and she was reported internally after she spoke with the chatbot about gun violence. But no one informed the police, even though the company supposedly suspended her account. She opened another one and that remained open. Families of the children who died and the province of British Columbia are suing OpenAI and its chief executive, Sam Altman, over what the company knew and kept to itself.

They raced ahead, anyway

By Gary Marcus — via Substack

Major scoop at The New York Times from Sheera Frenkel, Dustin Volz, and Dylan Freedman.

NEW: Employees at OpenAI had raised security alarms months before the Hugging Face incident and related A.I. cyberattacks — their warnings were ignored. From @sheeraf, @dnvolz and me.

nytimes.com/2026/09/29/tec…

— Dylan Freedman (@dylfreed) · on X

I don’t really have words for how awful this is. It’s exactly the nightmare I have been warning about for the last three years. Greedy company makes bad choice that causes chaos; government does nothing to stop them.

Meanwhile, the US government appears to be paralyzed

By Gary Marcus — via Substack

Nope, wasn’t just Hugging Face.

Wasn’t just that and a German website.

Wasn’t even the “dozens” we heard about the other day from OpenAI.

It’s actually (at least) *tens of thousands*, per a new scoop from Madison Mills at Axios. And not just OpenAI, either.

We all agree on what needs to be done. Now let’s do it.

By Gary Marcus — via Substack

BREAKING: Heads of major AI firms warn the UN that without intervention, "AI could be a risk to humanity as a whole".

— The Associated Press (@AP) · on X

This afternoon at the UN Security Council, Yoshua Bengio, Sam Altman, Dario Amodei, and Hugging Face CEO and Cofounder Clement Delangue gave back-to-back speeches.

Can’t wait to see what tomorrow brings

By Gary Marcus — via Substack

Ok, my headline is a joke. (I hope Trump’s riff today at the UN on relabeling AI as Super Intelligence was, too.)

But only barely.

BREAKING: Sam Altman to brief UN Security Council Wednesday on AI

— Kalshi (@Kalshi) · on X

As Dominic Romano, founder of Drainpipe.io put it on X,

An individual who really doesn’t appear to know much about the underlying tech, and who is incompetently leading a company which is creating the perception of a security crisis while being run like an Extortion Racket briefing the UN Security Council is like if Al Capone would have given a brief on security to the FBI, should anything said actually be taken at face value or taken seriously?

I understand why the UN had to give a microphone to Trump today, but I don’t see why they chose to give a microphone to Altman.

NYA Thanks Gary Marcus!

NYA Staff

While Donald Trump sides with tech executives and pushes back against AI regulation, state governors and legislatures are taking matters into their own hands. States are crafting their own laws to address both artificial intelligence and the massive data centers required to run it. The issues range from AI safety and privacy to electricity costs, water consumption, pollution, noise, and who ultimately pays for the infrastructure.

California is the most visible example. Gov. Gavin Newsom has spent years pushing AI legislation covering everything from deepfakes and privacy to AI companions and safety protocols. Recently, he went further by ordering state agencies to accelerate independent oversight of AI developers and draft recommendations for additional safeguards, including independent audits and an emergency "kill switch" for certain high-risk models.

California is also confronting the physical footprint of the AI boom. Lawmakers have passed measures addressing data center electricity costs and demanding transparency around energy and water consumption—with several bills now awaiting executive signature or veto.

Wondering what Sam’s endorsement of Dario is really about?

By Gary Marcus — via Substack

[Dear Reader, I am sorry to flood your email with so many posts lately; there’s simply too much happening that cannot be ignored. This one is short!]

This afternoon is supposed to be the big face-off about the supposed AI slowdown at Salesforce. Sam Altman is supposed to speak at 5p.m. ET, and lots of people are wondering what is he thinking?

An important clue comes from this post yesterday on X. See especially the highlighted bit:

Translation of the highlighted bit?

We will go as fast as we can without going to jail or getting sued out of existence, but for optics we will it call it “pacing”. Also, anything that will reduce regulatory uncertainty is probably great for our IPO, so bring it!

Who knows what piece of infrastructure might be attacked next.

By Gary Marcus — via Substack

"We believe that the world should accept some bad things happening for the benefits of this technology and people having the agency”

— Sam Altman, October 3

Breaking news from Madison Mills at Axios:

Full article here. “Findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond”. Day after scenarios!

Every minute we wait risks disaster

By Gary Marcus — via Substack

Breaking new from The New York Times, the latest of many agent-caused incidents that are happening with frightening regularity, but this time from Anthropic and fairly serious:

I have long felt that OpenAI is handling this inadequately. Seeing the same kind of incidents at Anthropic makes it absolutely clear that the current generation of synthetic agents simply cannot be trusted. Until they can be fixed, they should be removed from the market, just like a car with defective brakes.

New York tries to take care of its own

By Gary Marcus — via Substack

On Monday, starting at 11am Eastern Time, less than a day after Politico reported that Sam Altman said, in words that may last in infamy, “we believe that the world should accept some bad things happening for the benefits of this technology and people having the agency”, New York City’s City Council will hold an important hearing on AI policy. A number of AI-relevant proposals are under consideration.

Because New York City is a large market, I believe that I could help pick up some of the slack where the Federal government is falling short. (A number of states are also trying to fill in the slack).

What was signed, and what wasn’t said

By Gary Marcus — via Substack

Here’s the Accord, which Trump apparently described as a “tremendous” exercise in “self-policing”:

And potentially important news from Nvidia

By Gary Marcus — via Substack

The big news in the media this morning is that Jensen (“What, me worry?”) Huang has announced a platform for “AI Agent Safety”.

OpenAI hacks a foreign government — and the US faces its most critical AI test yet.

By Gary Marcus — via Substack

Before I discuss the latest breaking news, everyone should watch this clip from Jensen’s interview with Ezra Klein (or at least read the transcript below.

Watch Jensen‘s hands, and notice also that he doesn’t know all the facts.

Note the Ezra Klein, ever the attentive and well-prepared listener, picks up on it. The two are implicitly talking about the OpenAI Hugging Face incident, and Jensen doesn’t realize that it was caused by a product that was not, at the time, yet on the market. Ezra gently points that out.

Most importantly, take heed of the solution Jensen recommends.

The new secret is the old secret

By Gary Marcus — via Substack

For the last few days, all the buzz has been about Meta’s new agent, Muse:

Actions speak louder than words

By Gary Marcus — via Substack

I hate to say it, but President Trump is right: Dario has tried to position himself as a “perfect little angel”.

But actions speak louder than words. In the week since Dario dramatically called on the AI industry to “pace the frontier”…

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.

Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our…

— Dario Amodei (@DarioAmodei) · on X

1. Dario has pointed to two organizations for what might one to call independent monitoring of his own company. One is METR, deeply enmeshed in the same Bay Area ideology and social circles as Anthropic, and the other is Accenture, which Anthropic is already in business with.

   

    

   Choosing them as external voices speaks volumes about Amodei’s insincerity.

2. It also came out (not mentioned if I recall in Dario’s essay) that Anthropic is gearing up to build a wet biology lab, presumably without the usual institutional review boards overseeing.
   

   ANTHROPIC BUILDS BIOLOGY LAB TO PUSH AI INTO DRUG DEVELOPMENT

   Anthropic has quietly established a Bay Area wet lab, moving its life-sciences ambitions beyond computer simulations and into physical biological experiments.

   The company wants Claude to direct laboratory robots with…

   — *Walter Bloomberg (@DeItaone) · on X

   A tone-deaf and arrogant company building a wet lab with unreliable AI with too little oversight isn’t just a horror movie, it’s Jurassic Park waiting to happen.

   To refresh your memory, here is Jeff Goldblum shredding the bad judgment of a company that may as well be Anthropic. Relisten to the whole clip; the last line is especially relevant here.

   “Your scientists were so preoccupied with whether they could, they didn’t stop to think if they should”, indeed.

3. Slowdown? What slowdown? Dario ain’t gonna pace anything. In the end, it’s all about the IPO.

More disconcerting updates 😱

By Gary Marcus — via Substack

I have no time to write more, but these two things just crossed my desk at the same moment.

1. NYT just reported Trump is downplaying AI fears (and presumably resisting regulation) for economic reasons:

Here’s who we should be listening to instead

By Gary Marcus — via Substack

Went on BBC last night and nearly vomited in the electronic green room, as they played back-to-back clips from Altman, Huang, and Sanders, not one of which seemed credible to me.

I’m not blaming BBC; what they played was all newsworthy: prominent people making claims that the public might want to know about. (And to their credit, they gave me a chance to rebut them.)

But all of it was in one way or another nonsensical, uttered with perfect conviction.

Let’s take them in turn. Here’s Sam Altman: