granzymes

If this significantly cannibalizes Apple’s App Store revenue I would actually expect that they come up with a different way to monetize (maybe based on installs or number of users).

They could also implement that independent of the injunction, which applies to steering rules.

From least to most speculative:

* The nonprofit is staying the same, and will continue to control the for-profit entity OpenAI created to raise capital

* The for-profit is changing from a capped-profit LLC to a PBC like Anthropic and Xai

* These changes have been at least tacitly agreed to by the attorneys general of California and Delaware

* The non-profit won’t be the largest shareholder in the PBC (likely Microsoft) but will retain control (super voting shares?)

* OpenAI thinks there will be multiple labs that achieve AGI, although possibly on different timelines

Title seems misleading after reading the article.

If that’s how you’re going to count, why stop there? They also have 5-6 versions of the gpt-4o model, several versions of gpt-4, plus gpt-3.5, ect.

They were clear on the livestream that o3 and o4-mini are replacements for o1 and o3-mini just like gpt-4 replaced gpt-3.5.

>deal was where Facebook agreed not to import IG into FB

What deal was this?

FTC does in fact need to show (directly or through indirect evidence) that Meta has monopoly power in a relevant market and that it abused that power in order to win a Section 2 case.

If the relevant market ends up including TikTok or YouTube, FTC will be unable to make that showing.

A court didn’t buy that: the district court and 9th Circuit both held that iOS and Android compete in the Epic v. Apple case.

A jury however found that the relevant market in the Epic v. Google case was just Android. Google is understandably appealing that to the 9th Circuit.

No, that is saying that some of the improvements that went into 4.1 have also gone into ChatGPT, including chatgpt-4o-latest (2025-03-26).

> Compared to OpenAI o1 and OpenAI o3‑mini, GPT‑4.5 is a more general-purpose, innately smarter model. We believe reasoning will be a core capability of future models, and that the two approaches to scaling—pre-training and reasoning—will complement each other. As models like GPT‑4.5 become smarter and more knowledgeable through pre-training, they will serve as an even stronger foundation for reasoning and tool-using agents.

I’ve been following this case, and explained the history here: https://news.ycombinator.com/item?id=42889763.

The tile would be more accurate if it said “Apple requests a pause in US vs. Google while it attempts to intervene”.

I highly highly doubt that DOJ succeeds in convincing the Court to force Google to divest Chrome. That wasn’t really part of the trial, which focused on Google’s contracts to make its search engine the default on other platforms.

However, if DOJ does succeed then yes this case could make it easier in the future to force Microsoft to divest Bing if Bing ever gained a monopoly in search. As a comparison, the original Microsoft antitrust case precedent is very important to this search case. That’s why the government sued in this district (the same district they sued Microsoft in decades ago).

Prior to this lawsuit, Apple could bluff and say they would contract with Bing or some other search engine, and indeed evidence from the trial showed that Apple would invite Bing to submit bids but internally was clear that no amount of money would be enough to switch from Google. One key piece of evidence showed that Microsoft had offered to give Apple 100% of its Bing revenue on Apple devices and that still wasn’t enough to overcome Bing’s lack of quality.

If Google is removed as an available option, Apple can longer keep up the facade and will be forced to either suck it up and keep Google for free or give their users an inferior product. Mozilla tried switching from Google once and users hated it. Apple doesn’t want to piss off its users.

My gut says that Apple will probably get to keep what the Court gave it already (the ability to file two affidavits and participate in the post-hearing briefing) but not full intervention as it originally requested.

District courts have very broad discretion to order their affairs (going to the first point) and Apple’s request likely came too late to be permitted to intervene (going to the second point).

These types of motions for a stay are analyzed under a four-part test. Apple needs to show that 1) it is likely to succeed in its underlying request, 2) that without a stay it will suffer irreparable harm, 3) that no other party is unduly harmed by a stay, and 4) that the public interest is not disserved by a stay.

The four parts are weighed together, and a strong showing on one part can make up for a weaker showing on another part.

Assuming that Apple can make that showing, yes the standard procedure is to stay the case while Apple appeals. That preserves the status quo, which would be lost to Apple forever even if it ultimately succeeds on appeal if there was no stay (this is the “irreparable harm” part of the test).

I think people have the wrong idea when they see “defendant” in the title. Here’s what happened:

Apple has been watching this case closely from the beginning because it implicates the enormous amounts of money that Google pays them to be the default search engine on Apple devices. They’ve been involved as a nonparty, providing witnesses and documents requested by Google and the United States and have had lawyers present to observe from the gallery throughout the proceedings.

Until recently, Apple and Google were on the same side. Google was incentivized to win the case, which is also what Apple wanted because that would preserve the revenue share. But then Google lost the liability phase of the trial, and the government asked for truly extraordinary remedies including a divestiture of Chrome, significant changes to Android contracts, and a 10 year ban on any contract of any value whatsoever between Apple and Google.

Apple feels that its interests are no longer perfectly aligned with Google’s. First, Google has bigger fish to fry than its Apple revenue share because it needs to defend its Chrome and Android businesses. Second, it would actually benefit Google if they weren’t able to pay Apple revenue share if Apple kept Google as the default search engine because it’s the best on the market and Apple doesn’t want to provide an inferior option to its users.

So Apple filed a motion to intervene for the limited purpose of defending its right to contract during the remedies phase of the trial. That doesn’t make Apple a defendant. The Court can’t order Apple to do anything in its final remodels order. But it would allow Apple to present evidence during the remedy phase, call and cross-examine witnesses, and participate in the post-hearing briefing.

Apple says its purpose in participating in the remedies phase is to present evidence that neither Google nor the government is equipped to provide: information on Apple’s future plans. Apple says, regardless of what the Court does, it has no plans to invest the billions of dollars that would be required to enter the general search market. The government’s ban on Apple working with Google is premised at least in part on that competitive entry, which Apple aims to demonstrate is a fantasy.

The Court partially denied Apple’s request, fearing that other parties would also want to intervene. The Court will not allow Apple to call or cross-examine witnesses, but Apple can submit up to two affidavits from fact witnesses and participate in writing post-hearing briefs.

The government has represented to Apple that it will challenge the Court’s decision to give Apple that limited role during the remedies phase, and so the linked document is a request for a pause of the remedies phase while Apple appeals the partial denial of its motion to intervene to the D.C. Circuit.

Happy to answer any questions about the legal standards involved or the procedural history of the case.

This report was published by Common Sense Media, an advocacy organization with a clear interest in pushing this message as part of their lobbying efforts. Maybe their intentions are good (they seem to back some good bills!), but it’s not a neutral source of information.

If you look at polling data published as part of generic political surveys, you find that companies like Amazon and Meta are among the most trusted of U.S. institutions[0] (Amazon loses out only to the military).

[0] https://drive.google.com/file/d/1OgPzcB75uxXiFmTjUUb-ITIr7BF...

I would’ve asked to be taken off of this project if someone had asked me to build this. How embarrassing to need to stoop to this level.

>We now turn to the merits, which the parties have argued here in exemplary fashion. But the key flaw in the FCC’s arguments throughout is that the FCC elides the phrase “offering of a capability” as used in § 153(24). That phrase makes plain that a provider need not itself generate, process, retrieve, or otherwise manipulate information in order to provide an “information service” as defined in § 153(24). Instead, a provider need only offer the “capability” of manipulating information (in the ways recited in that subsection) to offer an “information service” under § 153(24). Even under the FCC’s narrower interpretation of “capability,” Broadband Internet Access Providers allow users, at minimum, to “retrieve information stored elsewhere. And we think it equally plain, for the reasons recited below, that Broadband Internet Service Providers offer at least that capability.

You're coming at this from the perspective of "it's weird to change the rules after the fact" which is reasonable but not how stare decisis works. The principle is that once something is decided once, it should be decided the same for all other parties litigating the issue unless and until the precedent is overturned. If something hasn't yet been litigated, it is still an open question.

It would be weird if, for example, the Supreme Court said that gay marriage is constitutionally protected but all the existing laws banning it could stand because they were enacted earlier.

The Federal government can regulate broadband providers, just not as telecommunications providers (subject to common carrier requirements) under the Telecommunications Act of 1996. Congress could update the law if it wanted to.

If the court had held otherwise, the Federal common carrier requirements would have preempted any common carrier duties imposed by California and New York. Since today the 6th Circuit says the FCC does not have that power (because broadband providers offer information services and not telecommunications services), California and New York are welcome to impose those duties within their state boundaries (as long as they don't step on any other areas of Federal preemption).

The Supreme Court said in Loper Bright that previous cases decided under Chevron are still good law. But this lawsuit challenged a brand new FCC order that had not been previously litigated.

Thanks to Loper Bright, we finally have certainty as to what the Telecommunications Act of 1996 actually means.

By way of background, the Obama administration first classified broadband providers as telecommunications services subject to common-carrier regulation under Title II in 2015. The FCC reversed itself under Trump, and then during the Biden administration changed its mind yet again. The law never changed.

Each time, the courts deferred to the agency's interpretation as "reasonable", despite the fact that the interpretations were completely inconsistent. There is a single right answer: either the law classifies broadband carriers as offering telecommunications services or information services.

Today, for the first time, a court has actually read the law[0] and concluded that broadband carriers offer information services, not telecommunications services. If that's not the policy result Congress wanted, then Congress can change the law instead of letting the President change his mind every 4 years.

[0] https://www.opn.ca6.uscourts.gov/opinions.pdf/25a0002p-06.pd...

Freedom of speech means that the government can't punish you for your speech. It has absolutely nothing to do with your speech being widely shared, listened to, or even acknowledged. No one has the right to an audience.

The only thing the Court says on the matter is:

>There are reasonable bases to conclude that intermediate scrutiny is appropriate even under these circumstances. We need not, however, definitively decide that question because we conclude the Act “passes muster even under the more demanding standard.”

It's possible that either Judge Ginsburg or Judge Rao think the strict scrutiny should apply, but if forced to put money on it I would say they both land on intermediate scrutiny.

While the Court based this decision solely on the non-classified portion of the record, I found this quote to be very interesting:

>Notably, TikTok never squarely denies that it has ever manipulated content on the TikTok platform at the direction of the PRC.

The Court held that the law could satisfy strict scrutiny (regardless of whether or not it applies), which requires that the Government prove that the restriction furthers a compelling interest and less restrictive alternatives would not accomplish the Government's goals. That's a high, high bar, and most laws subject to it are found wanting.

I doubt that the Supreme Court is going to want to hear this case. The most interesting legal question for them to decide was whether the law is subject to strict or intermediate scrutiny, but that is off the table now that the D.C. Circuit says it doesn't matter because the law could satisfy either standard.

Direct link to the opinion: https://media.cadc.uscourts.gov/opinions/docs/2024/12/24-111...

The Court did not say that the law is subject to strict scrutiny. The majority holds that they need not decide whether strict or intermediate scrutiny applies because the law satisfies the more stringent test.

The concurrence by Judge Srinivasan would instead hold that the law is subject to only intermediate scrutiny and uphold the law on that basis.

Why do it this way? It takes away a potential avenue of appeal for TikTok. They can’t ask the Supreme Court to hold that the law is subject to strict scrutiny, because that issue wouldn’t help them overturn this ruling.

If you had forced the rest of the panel to choose, I think they would have landed on intermediate scrutiny.

>judges are often paid multiples less than leading lawyers and they are not desireable jobs

The first part of this is true, but the second part is laughable. Federal judgeships are highly prized and nearly impossible to get. There are only on the order of 900 Article III judges in the country and they serve for life.

>The public often don't have access to all documents, transcripts and mostly what is published is the Judge's version of what the parties submissions are

Completely false. While there can be redactions for sensitive information like trade secrets, in general everything is public record. And in particular, there is a strong presumption rooted in the first amendment and the right to a public trial to unredact anything that the court relied upon in reaching its decision.

>When a judge dismisses a case and classifies it as not refileable, always raise an eyebrow.

You get a chance to refile when the judge thinks that you could plausibly plead more facts that, when taken as true, establish your claim. When the claims fail as a matter of law, leave to amend would be futile since you can’t plead around that.

The question before the Court in this case was whether the FTC has the power to issue substantive rules related to unfair methods of competition. The Court concludes[0] that the FTC does not have that power, and that Section 6(g) of the FTC Act permits the FTC only to issue "housekeeping" rules related to agency organization, procedure, or practice.

That follows from three main statutory and historical clues:

1) Section 6(g) contains no penalty provision—which indicates a lack of substantive force. Other grants of substantive rulemaking specify what happens if regulated entities don't follow the rule.

2) The location of Section 6(g) is suspect. It's the seventh in a list of twelve almost entirely investigative powers, and starts with a grant of organizational power to "from time to time classify corporations". This is hardly where you would expect to find sweeping substantive rulemaking power if Congress had in fact chosen to grant it.

3) The FTC has not historically seen itself as possessing the power to issue substantive rules on unfair methods of competition. Courts are skeptical of agencies discovering latent powers decades later.

Overall, this is exactly what commentators expected to happen. Banning noncompetes may well be good policy, but it's up to the legislature (whether Congress or in individual states) to enact that policy. In addition, nothing in this decision prevents the FTC from using its adjudicatory powers to go after individual examples of noncompetes which it believes are unfair methods of competition.

[0] https://storage.courtlistener.com/recap/gov.uscourts.txnd.38...

I liked this section of Justice Gorsuch's concurrence:

> How bad is the problem? Take just one example. Brand X concerned a law regulating broadband internet services. There, the Court upheld an agency rule adopted by the administration of President George W. Bush because it was premised on a “reasonable” interpretation of the statute. Later, President Barack Obama’s administration rescinded the rule and replaced it with another. Later still, during President Donald J. Trump’s administration, officials replaced that rule with a different one, all before President Joseph R. Biden, Jr.’s administration declared its intention to reverse course for yet a fourth time. Each time, the government claimed its new rule was just as “reasonable” as the last. Rather than promoting reliance by fixing the meaning of the law, Chevron deference engenders constant uncertainty and convulsive change even when the statute at issue itself remains unchanged.

This is a ruling from an Administrative Law Judge, which basically functions as a strong recommendation for how the National Labor Relations Board should decide a particular case. ALJs are Article II “in house” judges that specialize in the law of one particular administrative agency. They don’t have Constitutional life tenure protections and mostly do claims processing work for the Social Security Administration.

The next step in this case is for a 3-member panel of the NLRB (a 5 member board appointed by the President, currently with 4 members) to decide whether to accept the proposed ruling of the ALJ or to substitute their own opinion of how the case should come out.

Orders of the NLRB are not self-executing, so parties can appeal an adverse judgement to the Circuit of Appeals of their choice and the NLRB can cross-petition for enforcement of their decision. The Federal courts have final say over questions of law like “are non competes a violation of labor law”.

The entire process takes several years to play out.