The DOJ’s press release,
The Justice Department, joined by 16 other state and district attorneys general, filed a civil antitrust lawsuit against Apple for monopolization or attempted monopolization of smartphone markets in violation of Section 2 of the Sherman Act.
The complaint, filed in the U.S. District Court for the District of New Jersey, alleges that Apple illegally maintains a monopoly over smartphones by selectively imposing contractual restrictions on, and withholding critical access points from, developers. Apple undermines apps, products, and services that would otherwise make users less reliant on the iPhone, promote interoperability, and lower costs for consumers and developers. Apple exercises its monopoly power to extract more money from consumers, developers, content creators, artists, publishers, small businesses, and merchants, among others. Through this monopolization lawsuit, the Justice Department and state Attorneys General are seeking relief to restore competition to these vital markets on behalf of the American public.
“Consumers should not have to pay higher prices because companies violate the antitrust laws,” said Attorney General Merrick B. Garland. “We allege that Apple has maintained monopoly power in the smartphone market, not simply by staying ahead of the competition on the merits, but by violating federal antitrust law. If left unchallenged, Apple will only continue to strengthen its smartphone monopoly. The Justice Department will vigorously enforce antitrust laws that protect consumers from higher prices and fewer choices. That is the Justice Department’s legal obligation and what the American people expect and deserve.”
“No matter how powerful, no matter how prominent, no matter how popular — no company is above the law,” said Deputy Attorney General Lisa Monaco. “Through today’s action, we reaffirm our unwavering commitment to that principle.”
“When corporations engage in anticompetitive conduct, the American people and our economy suffer,” said Acting Associate Attorney General Benjamin C. Mizer. “Today’s action against Apple sends a strong signal to those seeking to box out competitors and stifle innovation — that the Justice Department is committed to using every tool available to advance economic justice and root out anticompetitive practices, wherever they arise.”
“For years, Apple responded to competitive threats by imposing a series of “Whac-A-Mole” contractual rules and restrictions that have allowed Apple to extract higher prices from consumers, impose higher fees on developers and creators, and to throttle competitive alternatives from rival technologies,” said Assistant Attorney General Jonathan Kanter of the Justice Department’s Antitrust Division. “Today’s lawsuit seeks to hold Apple accountable and ensure it cannot deploy the same, unlawful playbook in other vital markets.”
As alleged in the complaint, Apple has monopoly power in the smartphone and performance smartphones markets, and it uses its control over the iPhone to engage in a broad, sustained, and illegal course of conduct. This anticompetitive behavior is designed to maintain Apple’s monopoly power while extracting as much revenue as possible. The complaint alleges that Apple’s anticompetitive course of conduct has taken several forms, many of which continue to evolve today, including:
- Blocking Innovative Super Apps. Apple has disrupted the growth of apps with broad functionality that would make it easier for consumers to switch between competing smartphone platforms.
- Suppressing Mobile Cloud Streaming Services. Apple has blocked the development of cloud-streaming apps and services that would allow consumers to enjoy high-quality video games and other cloud-based applications without having to pay for expensive smartphone hardware.
- Excluding Cross-Platform Messaging Apps. Apple has made the quality of cross-platform messaging worse, less innovative, and less secure for users so that its customers have to keep buying iPhones.
- Diminishing the Functionality of Non-Apple Smartwatches. Apple has limited the functionality of third-party smartwatches so that users who purchase the Apple Watch face substantial out-of-pocket costs if they do not keep buying iPhones.
- Limiting Third Party Digital Wallets. Apple has prevented third-party apps from offering tap-to-pay functionality, inhibiting the creation of cross-platform third-party digital wallets.
The complaint also alleges that Apple’s conduct extends beyond these examples, affecting web browsers, video communication, news subscriptions, entertainment, automotive services, advertising, location services, and more. Apple has every incentive to extend and expand its course of conduct to acquire and maintain power over next-frontier devices and technologies.
For over a century, the Department has enforced the antitrust laws against illegal monopolies, deploying the Sherman Act to unfetter markets and restore competition. As alleged in the complaint, the Department is seeking equitable relief on behalf of the American public to redress Apple’s long-running, pervasive anticompetitive conduct.
Apple Inc. is a publicly traded company incorporated in California and headquartered in Cupertino, California. In fiscal year 2023, Apple generated annual net revenues of $383 billion and net income of $97 billion. Apple’s net income exceeds any other company in the Fortune 500 and the gross domestic products of more than 100 countries.
As reported by BNN Bloomberg,
Meta Platforms Inc., Microsoft Corp., X Corp. and Match Group Inc. asked a federal judge to reject Apple Inc.’s plan for opening its App Store to outside payment options.
In a rare joint filing, the companies said Apple’s plan “comports with neither the letter nor the spirit” of a 2021 ruling which found iPhone maker in violation of California unfair competition laws and required it to allow app developers to direct users to their own payment systems.
Echoing arguments by Epic Games Inc. in its long-running antitrust fight with Apple over the App Store, the companies said Apple’s plan “imposes new restrictions on app developers that ensure the price competition that the injunction was designed to promote will never materialize.”
The filing marks the latest volley in the back-and-forth over Apple’s tight control over its app marketplace, which is one of the world’s two dominant stores alongside Alphabet Inc.’s Google Play. An appeals court last year upheld the 2021 decision by an Oakland trial judge who found Apple’s business model violated California law by limiting developers from communicating about their alternative payment systems, which could wind up costing users less. Apple takes a cut from every purchase made on its app store.
Epic, maker of the popular Fortnite game, and Apple had both asked the Supreme Court to review the lower-court rulings, but the high court declined in January.
Apple said at the time it would let all third-party apps sold in the US include an outside link to a developer website to process payments for in-app purchases. But Epic last week asked for Apple to be held in contempt of court, claiming the company made outside links “commercially unusable” by imposing new fees.
Apple’s plan — which is nine pages and 2,900 words — describes “dozens of requirements and limitations to which developers must adhere to be eligible to include an external purchase link within their apps,” according to the filing from Meta, Microsoft, X and Match.
Apple previously said it is complying with the court order while continuing to protect the privacy and security of its app store users.
Though Apple is allowing app makers to include an outside link to alternative payment systems, the company is still attempting to collect a 12% or 27% revenue share from developers that opt out of the system.
As reported by The Verge,
The European Commission is opening five non-compliance investigations into how Apple, Google, and Meta are complying with its new Digital Markets Act antitrust rules, the regulator announced today. “We suspect that the suggested solutions put forward by the three companies do not fully comply with the DMA,” the EU’s antitrust chief Margrethe Vestager said in a statement. “We will now investigate the companies’ compliance with the DMA, to ensure open and contestable digital markets in Europe.”
In particular, the Commission plans to investigate Google and Apple’s anti-steering rules in their app stores and whether Google is guilty of self-preferencing its own services within its search engine. Apple’s browser choice screen for iOS is also being investigated as well as Meta’s “pay or consent model” for ad targeting. In a press conference, the Commission said it plans to conclude the investigations within the next 12 months.
Additionally, the EU regulator is also looking into the fee structure Apple announced for distributing apps outside of the App Store, as well as whether Amazon is self-preferencing its own products on its store. The Commission has also announced that Meta has been given an additional six months to make Messenger interoperable with other messaging services.
“We are not convinced that the solutions by Alphabet, Apple and Meta respect their obligations for a fairer and more open digital space for European citizens and businesses,” EU Commissioner Thierry Breton said in a statement. “Should our investigation conclude that there is lack of full compliance with the DMA, gatekeepers could face heavy fines.”
Following the investigation, the Commission will tell each gatekeeper what must be done to address concerns, as well as what measures the regulator is planning to take. If found not to be in compliance, the Commission can fine each company up to 10 percent of their annual global revenue under the DMA, or even 20 percent in cases of “repeated infringement.”
Earlier this month, the six major tech companies designated as gatekeepers under the DMA had to start complying with its rules. These include having to give customers the option of changing default apps and uninstalling the gatekeeper’s pre-installed applications, a ban on ranking a gatekeeper’s first-party services higher than rivals, and allowing third-party app stores.
The EU’s antitrust chief Margrethe Vestager previously indicated to Reuters that the Commission would be taking a closer look at how Apple is complying with the regulation over concerns that it may “de facto not make it in any way attractive to use the benefits of the DMA.”
The announcement follows fierce criticism of how Apple in particular is complying with the Digital Markets Act. Although the company is allowing alternative app stores on iOS as required by the new rules, it’s doing so with a new fee structure that its critics claim will dissuade developers from distributing apps outside of Apple’s App Store. Spotify called Apple’s compliance “a complete and total farce” while Epic CEO Tim Sweeney called the changes “a new instance of Malicious Compliance.”
Meta’s “pay or consent model” has also been the subject of complaints from various EU watchdogs. Last year, it launched a new paid tier for Facebook and Instagram in the EU that allows users to pay €9.99 a month to use each service without ads. The subscription was designed to be a way to get user consent to collect their data if they decide not to pay, but the Commission is concerned with the “binary choice” that Meta is offering. Last week, Meta said it had offered to reduce the monthly price of ad-free access to €5.99 a month to appease regulators.
In a series of statements, Apple, Meta, Google, and Amazon defended their approaches to complying with the DMA, and said they’ll continue to work with the Commission to address its concerns.
“We’re confident our plan complies with the DMA, and we’ll continue to constructively engage with the European Commission as they conduct their investigations,” Apple spokesperson Julien Trosdorf said in a statement to The Verge.
“Subscriptions as an alternative to advertising are a well-established business model across many industries,” Meta spokesperson Matt Pollard said. “We will continue to engage constructively with the Commission.”
“To comply with the Digital Markets Act, we have made significant changes to the way our services operate in Europe,” Google’s competition director Oliver Bethell said, adding that the company has “engaged with the European Commission, stakeholders and third parties in dozens of events over the past year to receive and respond to feedback.” Bethell said Google will defend its approach.
“Amazon is compliant with the Digital Markets Act and has engaged constructively with the European Commission on our plans since the designation of two of our services,” Amazon spokesperson Sam Barratt said in a statement. “We continue to work hard every day to meet all of our customers’ high standards within Europe’s changing regulatory environment.”
As reported by The Guardian,
The Florida governor, Ron DeSantis, has signed a law that has given his state one of the US’s most restrictive social media bans for minors, though it must still withstand expected legal challenges.
Once it takes effect, the bill signed on Monday bans social media accounts for children under 14 and requires parental permission for 14- and 15-year-olds. It was slightly watered down from a proposal DeSantis vetoed earlier in March, a week before the annual legislative session ended.
The new law was the top legislative priority for Republican state house speaker, Paul Renner. It takes effect on 1 January.
“A child in their brain development doesn’t have the ability to know that they’re being sucked into these addictive technologies and to see the harm and step away from it, and because of that we have to step in for them,” Renner said at a ceremony for the bill signing held at a Jacksonville school.
The bill DeSantis vetoed would have banned minors under 16 from popular social media platforms regardless of parental consent. But before the veto, he worked out compromise language with Renner to alleviate the governor’s concerns, and the legislature sent DeSantis a second bill.
Several states have considered similar legislation. In Arkansas, a federal judge in August blocked enforcement of a law that required parental consent for minors to create new social media accounts.
Supporters in Florida hope the bill will withstand legal challenges because it would ban social media formats based on addictive features such as notification alerts and autoplay videos, rather than on their content.
Renner said he expects social media companies to “sue the second after this is signed”.
“But you know what? We’re going to beat them,” Renner remarked. “We’re going to beat them and we’re never, ever going to stop.”
DeSantis also acknowledged the law will be challenged on free speech rights enshrined in the US constitution’s first amendment. He also bemoaned the fact the Stop Woke Act he signed into law two years ago was recently struck down by an appeals court with a majority of Republican-appointed judges. They ruled it violated free speech rights by banning private business from including discussions about racial inequality in employee training.
“Any time I see a bill, if I don’t think it’s constitutional, I veto it,” said DeSantis, a lawyer, expressing confidence that the social media ban will be upheld. “We not only satisfied me, but we also satisfied, I think, a fair application of the law and constitution.”
Khara Boender, a state policy director for the Computer & Communications Industry Association, said in a news release that she understood the concern for online safety but expressed doubt the law would “meaningfully achieve those goals without infringing on the first amendment rights of younger users”.
As reported by Ars Technica,
On Tuesday, Anthropic’s Claude 3 Opus large language model (LLM) surpassed OpenAI’s GPT-4 (which powers ChatGPT) for the first time on Chatbot Arena, a popular crowdsourced leaderboard used by AI researchers to gauge the relative capabilities of AI language models. “The king is dead,” tweeted software developer Nick Dobos in a post comparing GPT-4 Turbo and Claude 3 Opus that has been making the rounds on social media. “RIP GPT-4.”
Since GPT-4 was included in Chatbot Arena around May 10, 2023 (the leaderboard launched May 3 of that year), variations of GPT-4 have consistently been on the top of the chart until now, so its defeat in the Arena is a notable moment in the relatively short history of AI language models. One of Anthropic’s smaller models, Haiku, has also been turning heads with its performance on the leaderboard.
“For the first time, the best available models—Opus for advanced tasks, Haiku for cost and efficiency—are from a vendor that isn’t OpenAI,” independent AI researcher Simon Willison told Ars Technica. “That’s reassuring—we all benefit from a diversity of top vendors in this space. But GPT-4 is over a year old at this point, and it took that year for anyone else to catch up.”
Chatbot Arena is run by Large Model Systems Organization (LMSYS ORG), a research organization dedicated to open models that operates as a collaboration between students and faculty at University of California, Berkeley, UC San Diego, and Carnegie Mellon University.
We profiled how the site works in December, but in brief, Chatbot Arena presents a user visiting the website with a chat input box and two windows showing output from two unlabeled LLMs. The user’s task it to rate which output is better based on any criteria the user deems most fit. Through thousands of these subjective comparisons, Chatbot Arena calculates the “best” models in aggregate and populates the leaderboard, updating it over time.
Chatbot Arena is important to researchers because they often find frustration in trying to measure the performance of AI chatbots, whose wildly varying outputs are difficult to quantify. In fact, we wrote about how notoriously difficult it is to objectively benchmark LLMs in our news piece about the launch of Claude 3. For that story, Willison emphasized the important role of “vibes,” or subjective feelings, in determining the quality of a LLM. “Yet another case of ‘vibes’ as a key concept in modern AI,” he said.
The “vibes” sentiment is common in the AI space, where numerical benchmarks that measure knowledge or test-taking ability are frequently cherry-picked by vendors to make their results look more favorable. “Just had a long coding session with Claude 3 opus and man does it absolutely crush gpt-4. I don’t think standard benchmarks do this model justice,” tweeted AI software developer Anton Bacaj on March 19.
Claude’s rise may give OpenAI pause, but as Willison mentioned, the GPT-4 family itself (although updated several times) is over a year old. Currently, the Arena lists four different versions of GPT-4, which represent incremental updates of the LLM that get frozen in time because each has a unique output style, and some developers using them with OpenAI’s API need consistency so their apps built on top of GPT-4’s outputs don’t break.
These include GPT-4-0314 (the “original” version of GPT-4 from March 2023), GPT-4-0613 (a snapshot of GPT-4 from June 13, 2023, with “improved function calling support,” according to OpenAI), GPT-4-1106-preview (the launch version of GPT-4 Turbo from November 2023), and GPT-4-0125-preview (the latest GPT-4 Turbo model, intended to reduce cases of “laziness” from January 2024).
Still, even with four GPT-4 models on the leaderboard, Anthropic’s Claude 3 models have been creeping up the charts consistently since their release earlier this month. Claude 3’s success among AI assistant users already has some LLM users replacing ChatGPT in their daily workflow, potentially eating away at ChatGPT’s market share. On X, software developer Pietro Schirano wrote, “Honestly, the wildest thing about this whole Claude 3 > GPT-4 is how easy it is to just… switch??”
Google’s similarly capable Gemini Advanced has been gaining traction as well in the AI assistant space. That may put OpenAI on guard for now, but in the long run, the company is prepping new models. It is expected to release a major new successor to GPT-4 Turbo (whether named GPT-4.5 or GPT-5) sometime this year, possibly in the summer. It’s clear that the LLM space will be full of competition for the time being, which may make for more interesting shakeups on the Chatbot Arena leaderboard in the months and years to come.
As reported by PCMag,
To meet the demands for AI computing, Nvidia is introducing the Blackwell architecture to produce its fastest GPU chips yet.
Depending on the task, Blackwell GPUs promise to deliver a seven to 30 times performance improvement over Nvidia’s H100 product, which tech companies across the globe have been buying to train AI programs such as ChatGPT.
“Blackwell GPUs are the engine to power this new industrial revolution,” said Nvidia CEO Jensen Huang while introducing the chip at the company’s GTC event in San Jose.
Despite the huge performance gains, the new chip uses up to 25 times less energy, the company says. Blackwell is built with an improved version of TSMC’s 4-nanometer process, which Nvidia previously used to produce the H100, first introduced in 2022. The big difference is that Blackwell contains a whopping 208 billion transistors, up from 80 billion in the H100.
The company pulled this off by combining two large chip dies, which communicate over a 10TB/s interlink, to create a unified GPU.
Despite the huge performance gains, the new chip uses up to 25 times less energy, the company says. Blackwell is built with an improved version of TSMC’s 4-nanometer process, which Nvidia previously used to produce the H100, first introduced in 2022. The big difference is that Blackwell contains a whopping 208 billion transistors, up from 80 billion in the H100.
The company pulled this off by combining two large chip dies, which communicate over a 10TB/s interlink, to create a unified GPU.
Nvidia plans on selling the technology via the GB200 Superchip, which will contain two Blackwell GPUs and an integrated Grace GPU processor. Blackwell will also be sold through a server product called the GB200 NVL72, which will contain 72 Blackwell GPUs all connected over the company’s upgraded NVLink technology to deliver a “1.8TB/s bidirectional throughput.”
Whether Blackwell will one day end up in gaming GPU was left unclear. But the technology will power the latest craze in tech: generative AI. Nvidia has been reporting record earnings on the tech world’s growing appetite for GPUs to fuel AI development. There’s no word on Blackwell’s cost, but Nvidia says Microsoft, Meta, OpenAI, Google, and Amazon all plan on adopting the technology. The first Blackwell GPUs are slated to launch later this year.
As reported by TechCrunch,
In 2016, Facebook launched a secret project designed to intercept and decrypt the network traffic between people using Snapchat’s app and its servers. The goal was to understand users’ behavior and help Facebook compete with Snapchat, according to newly unsealed court documents. Facebook called this “Project Ghostbusters,” in a clear reference to Snapchat’s ghost-like logo.
On Tuesday, a federal court in California released new documents discovered as part of the class action lawsuit between consumers and Meta, Facebook’s parent company.
The newly released documents reveal how Meta tried to gain a competitive advantage over its competitors, including Snapchat and later Amazon and YouTube, by analyzing the network traffic of how its users were interacting with Meta’s competitors. Given these apps’ use of encryption, Facebook needed to develop special technology to get around it.
One of the documents details Facebook’s Project Ghostbusters. The project was part of the company’s In-App Action Panel (IAPP) program, which used a technique for “intercepting and decrypting” encrypted app traffic from users of Snapchat, and later from users of YouTube and Amazon, the consumers’ lawyers wrote in the document.
The document includes internal Facebook emails discussing the project.
“Whenever someone asks a question about Snapchat, the answer is usually that because their traffic is encrypted we have no analytics about them,” Meta chief executive Mark Zuckerberg wrote in an email dated June 9, 2016, which was published as part of the lawsuit. “Given how quickly they’re growing, it seems important to figure out a new way to get reliable analytics about them. Perhaps we need to do panels or write custom software. You should figure out how to do this.”
Facebook’s engineers solution was to use Onavo, a VPN-like service that Facebook acquired in 2013. In 2019, Facebook shut down Onavo after a TechCrunch investigation revealed that Facebook had been secretly paying teenagers to use Onavo so the company could access all of their web activity.
After Zuckerberg’s email, the Onavo team took on the project and a month later proposed a solution: so-called kits that can be installed on iOS and Android that intercept traffic for specific subdomains, “allowing us to read what would otherwise be encrypted traffic so we can measure in-app usage,” read an email from July 2016. “This is a ‘man-in-the-middle’ approach.”
A man-in-the-middle attack is an attack where hackers intercept internet traffic flowing from one device to another over a network. When the network traffic is unencrypted, this type of attack allows the hackers to read the data inside, such as usernames, passwords, and other in-app activity.
Given that Snapchat encrypted the traffic between the app and its servers, this network analysis technique was not going to be effective. This is why Facebook engineers proposed using Onavo, which when activated had the advantage of reading all of the device’s network traffic before it got encrypted and sent over the internet.
“We now have the capability to measure detailed in-app activity” from “parsing snapchat [sic] analytics collected from incentivized participants in Onavo’s research program,” read another email.
Later, according to the court documents, Facebook expanded the program to Amazon and YouTube.
Inside Facebook, there wasn’t a consensus on whether Project Ghostbusters was a good idea. Some employees, including Jay Parikh, Facebook’s then-head of infrastructure engineering, and Pedro Canahuati, the then-head of security engineering, expressed their concern.
“I can’t think of a good argument for why this is okay. No security person is ever comfortable with this, no matter what consent we get from the general public. The general public just doesn’t know how this stuff works,” Canahuati wrote in an email, included in the court documents.In 2020, Sarah Grabert and Maximilian Klein filed a class action lawsuit against Facebook, claiming that the company lied about its data collection activities and exploited the data it “deceptively extracted” from users to identify competitors and then unfairly fight against these new companies.
As reported by TechCrunch,
In 2016, Facebook launched a secret project designed to intercept and decrypt the network traffic between people using Snapchat’s app and its servers. The goal was to understand users’ behavior and help Facebook compete with Snapchat, according to newly unsealed court documents. Facebook called this “Project Ghostbusters,” in a clear reference to Snapchat’s ghost-like logo.
On Tuesday, a federal court in California released new documents discovered as part of the class action lawsuit between consumers and Meta, Facebook’s parent company.
The newly released documents reveal how Meta tried to gain a competitive advantage over its competitors, including Snapchat and later Amazon and YouTube, by analyzing the network traffic of how its users were interacting with Meta’s competitors. Given these apps’ use of encryption, Facebook needed to develop special technology to get around it.
One of the documents details Facebook’s Project Ghostbusters. The project was part of the company’s In-App Action Panel (IAPP) program, which used a technique for “intercepting and decrypting” encrypted app traffic from users of Snapchat, and later from users of YouTube and Amazon, the consumers’ lawyers wrote in the document.
The document includes internal Facebook emails discussing the project.
“Whenever someone asks a question about Snapchat, the answer is usually that because their traffic is encrypted we have no analytics about them,” Meta chief executive Mark Zuckerberg wrote in an email dated June 9, 2016, which was published as part of the lawsuit. “Given how quickly they’re growing, it seems important to figure out a new way to get reliable analytics about them. Perhaps we need to do panels or write custom software. You should figure out how to do this.”
Facebook’s engineers solution was to use Onavo, a VPN-like service that Facebook acquired in 2013. In 2019, Facebook shut down Onavo after a TechCrunch investigation revealed that Facebook had been secretly paying teenagers to use Onavo so the company could access all of their web activity.
After Zuckerberg’s email, the Onavo team took on the project and a month later proposed a solution: so-called kits that can be installed on iOS and Android that intercept traffic for specific subdomains, “allowing us to read what would otherwise be encrypted traffic so we can measure in-app usage,” read an email from July 2016. “This is a ‘man-in-the-middle’ approach.”
A man-in-the-middle attack is an attack where hackers intercept internet traffic flowing from one device to another over a network. When the network traffic is unencrypted, this type of attack allows the hackers to read the data inside, such as usernames, passwords, and other in-app activity.
Given that Snapchat encrypted the traffic between the app and its servers, this network analysis technique was not going to be effective. This is why Facebook engineers proposed using Onavo, which when activated had the advantage of reading all of the device’s network traffic before it got encrypted and sent over the internet.
“We now have the capability to measure detailed in-app activity” from “parsing snapchat [sic] analytics collected from incentivized participants in Onavo’s research program,” read another email.
Later, according to the court documents, Facebook expanded the program to Amazon and YouTube.
Inside Facebook, there wasn’t a consensus on whether Project Ghostbusters was a good idea. Some employees, including Jay Parikh, Facebook’s then-head of infrastructure engineering, and Pedro Canahuati, the then-head of security engineering, expressed their concern.
“I can’t think of a good argument for why this is okay. No security person is ever comfortable with this, no matter what consent we get from the general public. The general public just doesn’t know how this stuff works,” Canahuati wrote in an email, included in the court documents.In 2020, Sarah Grabert and Maximilian Klein filed a class action lawsuit against Facebook, claiming that the company lied about its data collection activities and exploited the data it “deceptively extracted” from users to identify competitors and then unfairly fight against these new companies.
As reported by Fortune,
Generative artificial intelligence might be driving the markets and forcing businesses to prepare for a new potential future, but the CEO of OpenAI isn’t real impressed about where the technology stands at the moment.
In an interview with computer scientist and podcaster Lex Fridman, Sam Altman had some surprising words about GPT-4, the latest AI model from his company, when asked about its most impressive capabilities.
“I think it kind of sucks, relative to where we need to get to and where I believe we will get to,” Altman said. “I expect that the delta between [GPT-]5 and 4 will be the same that was between 4 and 3 and I think it is our job to live a few years in the future and remember that the tools we have now are going to kind of suck, looking backward at them, and that’s how we make sure the future is better.”
GPT-4, Altman noted, was a huge advancement over GPT-3. But people who have used GPT-4 now scoff at its predecessor, which at the time seemed revolutionary. That’s where he stands now, as GPT-5 gets closer to release. (OpenAI has said its next-generation model will release in the next year, but has not been more specific than that.)
“Look, I don’t want to downplay the accomplishment of GPT-4, but I don’t want to overstate it either,” said Altman. “And I think this point that we are on an exponential curve, we’ll look back relatively soon at GPT-4 like we look back at GPT-3 now.”
Altman said he uses GPT-4 as “sort of like a brainstorming partner” and notes it can sometimes help on longer-horizon tasks, breaking them into smaller steps and assisting with those.
“When that works, which is not very often, it’s very magical,” he said.
Chinese government blocks use of Intel, AMD chips in hardware
Neuralink’s first human patient talks on livestream, plays chess
Researchers demonstrate exploit that could let hackers open any hotel room door with their phone
Apple’s in-the-box iPhone update machine revealed
ETH Zurich researchers’ ANYmal quadruped robot can, terrifyingly, do parkour
um what the sigma
Do Sonic characters have feet?
J.C. Thornton on X: “Excuse me for the Sonic feet pictures; I …
In every incarnation of the series, the feet design changes in many ways; sometimes, it’s three toes, four toes, and five toes, or in some cases two toes and blob feet. It’s happened in cartoons, comics, and even video games before and after the “Dreamcast” era, where they’re known to have the blob design.