Feed aggregator

Thirty years after Apple abandoned Copland, its doomed attempt to create a modern Mac operating system, a new emulator lets you try the final developer build in your browser. Copland was intended to replace Apple's System 7 with a PowerPC-native operating system built around a microkernel, protected memory, and preemptive multitasking for system services. Three unfinished developer builds – D7E1, D9, and D11E4 – are publicly known. Relatively few people outside Apple's developer community in the mid-1990s ever had the opportunity to use them. Now you can boot Apple Copland D11E4 in your browser and explore the unstable prerelease OS without installing anything. Developer Michael Steil, also known as Mist, made it possible with an experimental fork of the DingusPPC Power Mac emulator. Steil says the 11 patches that enable DingusPPC to boot Copland were written with AI assistance. The upstream project does not accept AI-assisted patches, so the fixes would have to be rewritten before they could be incorporated. Don't expect miracles. The emulator runs D11E4, the final build Apple distributed to developers, dating from June 1996. It remains very much unfinished: if the emulated system hits an assertion, it drops into a debugger and asks the user to continue manually. Apple CEO Gil Amelio had demonstrated Copland at the company's 1995 Worldwide Developers Conference, but the project never approached the stability required for release. Much of Copland existed as plans and documentation rather than finished code. In 2009, Steil assembled an extensive archive of that material in Apple Copland Reference Documentation. There was also a planned successor to Copland, codenamed Gershwin, but it remained nothing more than a plan. The modern Gershwin desktop we covered a year ago is a nod to that nonexistent OS. Anyone familiar with the classic Mac OS of the late 1990s will recognize parts of Copland, and for good reason. Copland overran so badly that Amelio recruited the late Ellen Hancock from IBM as Apple's chief technology officer and tasked her with getting the project back on track. After reviewing its progress, she instead recommended cancelling it and acquiring an operating system elsewhere. Apple considered several alternatives, including acquiring Be for BeOS and working with Sun on a system based on Solaris. It ultimately bought NeXT, bringing Steve Jobs back to the company along with NeXTSTEP. Hancock was later sidelined under Jobs and resigned in 1997. Apple acquired NeXT at the end of 1996, just months after cancelling Copland. The first developer release of its NeXTSTEP-based successor, Rhapsody, followed in 1997 – continuing the musical codenames, if not the code. Rhapsody evolved into Mac OS X Server 1.0 in 1999 and the consumer release of Mac OS X 10.0 in 2001. Replacing the Mac OS architecture took years. While that work continued, Apple folded some of Copland's more mature components into updates to its existing operating system. Mac OS 8, released in 1997, gained Copland's Platinum interface and a PowerPC-native, multithreaded Finder. Its Appearance control panel supported themes, although Apple shipped only Platinum. Renaming what had been planned as Mac OS 7.7 also helped Apple end the clone program, whose licensing agreements covered System 7. Mac OS 8.1 followed in 1998 with HFS+, Apple's new format for larger disks. Later that year, Mac OS 8.5 became the first PowerPC-only release. Further pieces of Copland's work appeared across Mac OS 8 and 9, allowing Apple to improve the old platform while Mac OS X was being prepared. When Apple bought NeXT, NeXTSTEP already ran on Motorola 680x0, Intel x86, Sun SPARC and HP PA-RISC processors. That portable foundation later helped Apple move the Mac to Intel, while OS X became the basis for the operating systems used by the iPhone and iPad. Apple's future might have looked very different had Copland succeeded. Now it's easier than ever to see why it didn't. ®

One of my many cataloged memories of the Trump era is from the run-up to the 2016 election, when one of my colleagues related to me what one of his sources in Brazil, fretful about Donald Trump’s rise, told him. “It’s OK for the world for Brazil to occasionally be a basket case nation,” he said. “But the thought of the U.S. becoming one is really disturbing.” This week, Trump addressed the U.N. General Assembly, cast in the role we once reserved for people like Mahmoud Ahmadinejad and Muammar Qaddafi: the batshit egomaniac dictator (only this time, with nukes). It seems to me that we’ve well and truly failed the basket case test. But a new poll has given me some fresh worries on this front—and a concern that maybe too many of us have developed a little taste for misrule.

The survey was published this week by Liberal Currents and it sought to catch the mood of the Democratic electorate as Election Day looms. Much of what the poll found was downright anodyne: A plurality of respondents favor Kamala Harris as the presidential nominee in 2028; “economy, inflation & cost of living” wins as the respondents’ biggest issue they want a future Democratic president to tackle. But one result made me do a double take. Under the section, “Feelings about Trump leaving office,” respondents were asked, “If you woke up tomorrow morning and saw the following news headlines, how would you feel?”

One of the choices, “To ‘safeguard the Constitution,’ a group of U.S. military leaders have removed President Donald Trump from power and assumed control of the United States” earned a net approval of 31 percent, with 59 percent approving and 28 percent saying no. That’s not good! Now the reason I’m not completely disassociating is because the ways respondents answered the other choices were more rational: Impeachment was by far the favored choice of removing the president, at 80 percent net sentiment, followed by removal via the Twenty-Fifth Amendment at 77 percent, and “dying peacefully in his sleep” at 40 percent. Assassination—by far the most immoral and dangerous way Trump could leave office, for all the reasons I lay out here—was the only choice with a negative sentiment, as it should have.

Still, I’m sorry, but the approval number for a military junta taking over the United States is way, way too high. Folks, you do not want a military coup in the United States under any circumstances.

Perhaps one reason this idea polls so well is that we’ve recently had a sample taste of what it’s like to be under unconstitutional leadership and it led to a desirable outcome. During the January 6, 2021, insurrection and for about a fortnight after, the United States was for all intents and purposes governed by a weird triumvirate: Nancy Pelosi, Mike Pence, and General Mark A. Milley. Trump was, during this time, meaningfully constrained from exercising some of his constitutional authority, while this troika took it up. It can be argued, of course, that this may have been for the best. This interregnum eventually yielded to the Biden administration and its comparative normalcy. But for many people, myself included, it was still fairly unsettling. In many places around the world, such takeovers end badly.

This is perhaps why the basket case anecdote came flooding back in my memory banks, as Brazil’s history is demonstrative of why military coups d’état are a bad idea. If the Latin American nation is any guide, once you start having military takeovers, you tend to keep having them. As the Berlin Journal’s James N. Green recalls, “Since 1889, when the Brazilian army ousted Emperor Pedro II and established a republic, the armed forces initiated nine attempted takeovers, five of which were successful.” Brazilian history also teaches us that military takeovers can be very difficult to dislodge. “The last military intervention into Brazilian politics took place in 1964, with the explicit support of the U.S. government,” writes Green. “It led to a 21-year-long military dictatorship.”

I feel it should be obvious to most that a Trump-ousting coup would be an incredibly combustible event that would shatter the political landscape and lead to widespread uproar that would be difficult to contain. As I’ve noted previously, the MAGA movement is steeped in conspiratorial thinking and bound in cultlike devotion to Trump. Martyring him is thus a very bad idea. A military-style coup d’état would likely reinflame the passions that led to the January 6 insurrection, and serve as rocket fuel for the growing number of people who TNR’s Matt Ford says claim “a purported Second Amendment right to murder civil servants and assassinate elected officials.”

As TNR’s Tim Noah wrote earlier this week, one Trump-removal scenario that’s on the table comes down to the GOP’s long-term political thinking: Should Republicans get battered in the midterm elections, there could be a strong incentive to convince Trump to retire or quit and pass the torch early to Vice President JD Vance, so as to allow him two years to essentially run as an incumbent. If that’s the way it goes, so be it. Trump would, of course, have strong incentives to stay, as his becoming a private citizen again would immediately embroil him anew in any number of legal battles.

Barring that, the constitutional methods of removal that most Democrats thankfully seem to favor are the better options—though it seems that impeachment is, for Democratic leadership anyway, off the table. Perhaps it would be impractical for a newly empowered Democratic Congress to try impeachment for a third time (though they say that’s “the charm”). Still, it’s worth remembering that Republican senators had two shots to end a derelict presidency—once immediately after he sicced a mob on them—and they failed in their duty to oust Trump and bar him from holding public office. In this way, impeachment should remain a salient issue in this autumn’s Senate races: All of the miseries we are experiencing right now would never have happened if enough Republican senators truly loved their country.

It needs to be said that if Trump’s reign is to end prematurely—that is, before his stepping down on the next Inauguration Day—things won’t be peaches and cream. Even if it comes about through constitutional means, the U.S. will have to navigate a period of tumult. If we’re contemplating the end of the Trump presidency, rather than indulge in fantasies about a military coup, we should all contemplate how each of us, in our own way, might help reinforce the civic foundation and serve as good stewards of a nation of laws. Perhaps, as Noah writes, we’re within sight of Trump’s fall. We should not take such matters lightly, because if the fall is all that’s left, it matters a great deal.

“Made in Santa Cruz: Why Art Matters” is a video series, published every other Sunday, exploring the artists who shape Santa Cruz County's creative landscape. Each episode offers a peek into an artist's practice, the places and experiences that inform their work, and the ways creativity helps us see our community differently. Today’s artist is I.B. Bayo.

Big AI has a problem. Its companies must ingest other people’s work. You know, books, news stories, photographs, code, websites, that one original meme you came up with, and practically everything else on the internet to train its large language models (LLMs). We all know that. We also know that AI companies hate paying for any of it, obeying the licenses attached to it, or, God forbid, sharing their revenue with the companies and people who created the work in the first place. Recently, Big AI's default way of doing business: "Take it now, argue about legality later," has become more in your face than ever. Look, for example, at the copyright fight between The New York Times and OpenAI and Microsoft. According to 404 Media’s reporting on recently unsealed court documents - arguments from the plaintiffs that the court has not yet ruled on - Microsoft allegedly knew what it was doing when it was importing the internet willy-nilly. Microsoft's Director of Applied Science, Dr. Brent Hecht, was quoted in the news plaintiffs' 92-page combined brief as saying: the case was about “an astonishing theft of unprecedented proportions." Indeed, it was possibly the “largest theft of labor in human history,” he was quoted as saying in the summary judgment brief from the journalists' lawyers [PDF].This, mind you, wasn't a comment by a member of the press; it was from a senior Microsoft staffer. Hecht wasn't the only one at Microsoft who commented. In an internal Microsoft policy document also quoted in the court papers, the authors admitted generative AI could “significantly disrupt the employment of the very people who generated the data on which the foundation model was trained [because] LLMs are a product that destroys its supply chain.” That, in turn, leads to model collapse. Ironically, AI is killing the content-creator goose that lays the gold eggs of worthwhile content. Judge Sidney H. Stein of the US District Court for the Southern District of New York has not yet issued a decision on the case. OpenAI and Microsoft are actively disputing the plaintiffs' allegations under a "fair use" defense. They contend that using public articles and books to train large language models helps push forward public knowledge, and isn't acting as a "unlawful economic market substitute". But the statements the documents cites are real. Big AI knows it needs the content. Some of the companies concerned don't give a damn; money now, worry later is their motto. Those who can look past the revenue numbers are well aware they're engaged in what Microsoft itself called a “doom loop” of AI content strategy in those unsealed court documents. Will that stop them? Nah. According to the filing [PDF], which refers to sworn deposition testimony of OpenAI’s own corporate representative, it also testified that its LLMs were happy to vacuum up other people's work even when a paywall nominally protected it. Its rep was quoted as saying they were unaware of “any effort to detect paywall content in its training datasets” or “to remove paywall content from its training datasets.” When OpenAI cofounder Greg Brockman was told OpenAI could hack its way through the firewall, he responded, “ah, nice.” Whether or not these statements are found to be reflective of a wider attitude, similar attitudes prevail amongst Big AI. Such policies could eventually come back and bite the LLM makers; it's already affecting the market for journalism, fiction writing, graphics creation, anything that demands humans actually get paid for producing original work. But Big AI just doesn't want to pay for it. That's not a side effect. It's the business model. Joe User couldn't care less. They just want a quick answer that sounds right. Some of them couldn't care less about getting the right answers. As for looking deeper to see if the information they're regurgitating is accurate, forget about it! As an OpenAI software engineer put it: “No matter how prominently we show the links, users won’t click.” Well, I click. That's a big reason why Perplexity is my AI of choice. It's not that it gives a more reliable answer. No, it's that, unlike most LLMs, Perplexity provides sources and links, and I check them before accepting what it tells me. But then I'm a journalist whose degrees are in history, where my professors drummed into me that you always - always - look at the primary sources. Big AI takes the same approach with its code generators. The US Ninth Circuit recently handed GitHub, Microsoft, and OpenAI a narrow win in the Doe v. GitHub lawsuit. The court ruled that the plaintiffs hadn't established a claim under one particular provision of the Digital Millennium Copyright Act (DMCA), which concerns the removal or alteration of copyright-management information (CMI). In short, the court said that generating new code without including CMI is not necessarily the same thing as removing or altering copyright information from an existing work. Judge Eric Miller wrote: “One who creates a new work and fails to include CMI cannot be said to have ‘removed’ or ‘altered’ anything.” This ruling didn't establish that GitHub Copilot or any other coding model may train on every open source project without restriction. It didn't determine that copying source code into a training set is always fair use. It did not decide that generated code cannot infringe copyright. And it certainly did not repeal the GPL, Apache, BSD, MIT, or any other open-source license. After all, open source isn't a synonym for “do whatever you want with it.” It just seems that way when AI gets its hands on open source code. That's the problem with AI-generated code. A programmer who receives a Copilot suggestion is very unlikely to check whether the GPL covers the code, or whether Apache covers the snippet from a library. The developer certainly can't tell which license terms may apply to the code that just popped up from Opus 5.5 or GPT 6. As the guy from OpenAI said, people don't click the links, and in AI-enabled programming pipelines, they don't even have the links anyway. An open source-savvy attorney friend of mine has a bigger worry than this narrow decision. He told me that what gives him "pause is a broader trend. There are multiple cases where parties are litigating open source licenses as contracts rather than as the IP / copyright licenses they were written to be. Contract theories let a plaintiff sidestep the questions copyright forces you to answer. 'Do you own the work? Is it protectable expression? Was it actually copied?'" He added: "Those questions are the foundation on which licenses are built. Open source licenses are grants of permission to use someone's intellectual property. If a case can't show that any intellectual property was owned or infringed, it's fair to ask whether enforcing the license as a bare contract supports open source licensing or quietly turns it into something else." With AI in the mix, this issue will eventually become one of those nasty IP matters developers hate, businesses want to avoid, but that courts must settle with expensive lawyers whose hourly rates worry even AI millionaires. After all, IP uncertainty isn't a minor paperwork issue. It is a supply-chain security and compliance problem. If you use AI-generated code, you may be importing unknown legal obligations into your software. If you are an open-source developer, your work may be used to improve a proprietary service that returns code without provenance, attribution, or meaningful reciprocity. And if you are a customer, you may be relying on software whose origins no one can fully explain. The Ninth Circuit ruling didn't solve any of that. We're in for a lot more open source litigation that will make SCO vs. the known Linux universe look like a kerfuffle over a parking ticket. Then there is the latest wrinkle: A federal antitrust lawsuit against Anthropic, OpenAI, SpaceXAI, and Google. The plaintiffs allege the companies coordinated to slow the development of advanced AI systems. The lawsuit points to public statements from AI leaders all calling for coordination to “pace” frontier AI development over a single weekend. Anthropic CEO Dario Amodei has argued for industry-wide coordination to slow frontier development. OpenAI CEO Sam Altman, Google DeepMind co-founder Demis Hassabis, and Elon Musk have also publicly expressed versions of the view that powerful AI systems need controls and careful deployment. On its face, that sounds sensible. AI safety is a real issue. Anyone who says otherwise has not been paying attention. But there's a huge difference between all big AI players agreeing among themselves on how quickly a market should develop. This looks like an awful lot like competitors deciding who gets to compete, what they get to build, and when they may build it. The Big AI companies all look like good guys while simultaneously strangling their smaller competitors. To mere mortals, this may sound like an open-and-shut case. It's not. Antitrust law requires evidence of an agreement and actual harm to competition; executives making similar public comments is not enough. The plaintiffs will need considerably more than quotes about AI safety or responsible development to win. Besides, if they do win, so what? What recently looked like huge antitrust wins has had its teeth pulled when it came to actually punishing the likes of Google and its ad business, or its earlier "lose the case, win the settlement" with Google Search. Do you see the pattern here? Big AI has effectively argued that copyright law, open source licenses, and competition rules shouldn't slow them down. When creators object, they are told that training is fair use. When developers object, they are told that generated code is new code. When competitors object, they are told security requires the largest companies to coordinate. Funny how the answer always seems to leave Big AI with the data, the market, and the money, isn't it? The central issue isn't whether AI should be allowed to develop. It will develop. The issue is whether the companies building it should be allowed to treat everyone else’s work as a free raw-material supply, then use the resulting products to take away the creators’ traffic, revenue, and bargaining power. Unless we stop Big AI now from this land grab with GPUs, only a handful of trillionaires will really benefit from AI, with the rest of us becoming serfs in a 21st-century post-technology feudal system. ®

PNOĒ, the Malden, Mass.-based startup whose breath-analyzing mask used to bear an unfortunate resemblance to Bane's, is launching a sleeker self-serve version on October 1 that lets gym-goers measure their VO₂ max and other metabolic markers in eight minutes without a trained operator.

The limited test covers select products and users, with a broader rollout planned for later in October.

Explaining the Right is a weekly series that looks at what the right wing is currently obsessing over, how it influences politics—and why you need to know. President Donald Trump’s decision to trash the First Amendment and bar news outlets from the White House would seem to be compelling evidence of antagonism between the right and the media. But despite Trump’s constant…

Source

A cartoon by Mike Luckovich. Related | GOP is already deep in the stages of grief as midterm losses loom…

Source

ph @bastet
Capocotta
Estate 2022

Blue Cross Blue Shield says hospital use of AI tools led to an additional $942M in healthcare spending over a two-year period.

By Gabby Birenbaum for The Texas Tribune Jill Tate has spent 30 years building and maintaining the Republican majority in Texas. In all that time, she’s never been more worried about an election. “This is truly the first time I’ve felt in my bones we could lose the Texas House, and we could lose a statewide seat,” Tate said. “And that’s very scary.” Now president of Texas…

Source

TikTok will pay Alabama at least $100 million in a settlement tied to allegations that the short-form video platform misled users about safety and was designed to addict children.

ALT