For years now, many AI industry watchers have looked at the quickly growing capabilities of new AI models and mused about exponential performance increases continuing well into the future. Recently, though, some of that AI "scaling law" optimism has been replaced by fears that we may already be hitting a plateau in the capabilities of large language models trained with standard methods.

A weekend report from The Information effectively summarized how these fears are manifesting amid a number of insiders at OpenAI. Unnamed OpenAI researchers told The Information that Orion, the company's codename for its next full-fledged model release, is showing a smaller performance jump than the one seen between GPT-3 and GPT-4 in recent years. On certain tasks, in fact, the upcoming model "isn't reliably better than its predecessor," according to unnamed OpenAI researchers cited in the piece.

On Monday, OpenAI co-founder Ilya Sutskever, who left the company earlier this year, added to the concerns that LLMs were hitting a plateau in what can be gained from traditional pre-training. Sutskever told Reuters that "the 2010s were the age of scaling," where throwing additional computing resources and training data at the same basic training methods could lead to impressive improvements in subsequent models.

Read full article

Comments

© Getty Images

As US states were busy counting votes to confirm who would be the next president Tuesday night, Nintendo's Japanese Twitter account was busy confirming a key backward-compatibility feature for the upcoming "Nintendo Switch successor," which is still only pre-announced, officially.

"At today's Corporate Management Policy Briefing, we announced that Nintendo Switch software will also be playable on the successor to Nintendo Switch," Nintendo posted in a social media update attributed to company President Shintaro Furukawa. "Nintendo Switch Online will be available on the successor to Nintendo Switch as well."

This is Furukawa. At today's Corporate Management Policy Briefing, we announced that Nintendo Switch software will also be playable on the successor to Nintendo Switch. Nintendo Switch Online will be available on the successor to Nintendo Switch as well. Further information about…

— 任天堂株式会社（企業広報・IR） (@NintendoCoLtd) November 6, 2024

In the full policy briefing referenced in that post, Nintendo adds that it "believe[s] that it is important for Nintendo’s future to make use of Nintendo Account and carry over the good relationship that we have built with the over 100 million annual playing users on Nintendo Switch to its successor." The company also makes the (perhaps obvious) clarification that "in addition to being able to play Nintendo Switch software they currently own, consumers will be able to choose their next purchase from a broad selection of titles released for Nintendo Switch [on its successor]."

Read full article

Comments

© Aurich Lawson

Last night, Nintendo pulled off a surprise launch of a new Nintendo Music smartphone app, offering many of the company's staple soundtrack songs as a perk to Nintendo Switch Online subscribers. But the new subscription freebie could give Nintendo additional motivation to once again crack down on Internet users who have been collecting and posting Nintendo music online for years now.

The Nintendo Music app includes hundreds of songs from titles to download or stream, ranging from 1985's Super Mario Bros. to last year's Pikmin 4. The current music selection is far from comprehensive, but it at least touches on many of Nintendo's most popular series, including Zelda, Pokemon, Kirby, Fire Emblem, Metroid, and Animal Crossing (plus some popular background music from various Wii Channels). Nintendo promises that more tracks will be "added over time," mirroring the process Nintendo has used to add to its Nintendo Switch Online classic game downloads.

Nintendo Music users can build their own playlists, of course, or choose from a number of pre-arranged playlists to suit different moods or character themes. The app also syncs with your Nintendo account to highlight music from games you play and offers options to avoid "spoilers" from certain game music or extend songs in lengthy loops.

Read full article

Comments

© Nintendo

The rise of powerful generative AI models in the last few years has led to plenty of stories of corporations trying to use AI to replace human jobs. But a recent New York Times story highlights the other side of that coin, where AI models simply become a powerful tool aiding in work that still requires humanity's unique skillset.

The NYT piece in question isn't directly about AI at all. As the headline "Inside the Movement Behind Trump’s Election Lies" suggests, the article actually reports in detail on how the ostensibly non-partisan Election Integrity Network "has closely coordinated with the Trump-controlled Republican National Committee." The piece cites and shares recordings of group members complaining of "the left" rigging elections, talking of efforts to "put Democrats on the defensive," and urging listeners to help with Republican turnout operations.

To report the piece, the Times says it sifted through "over 400 hours of conversations" from weekly meetings by the Election Integrity Network over the last three years, as well as "additional documents and training materials." Going through a trove of information that large is a daunting prospect, even for the team of four bylined reporters credited on the piece. That's why the Times says in a note accompanying the piece that it "used artificial intelligence to help identify particularly salient moments" from the videos to report on.

Read full article

Comments

© Getty Images

If you read enough science fiction, you've probably stumbled on the concept of an emergent artificial intelligence that breaks free of its constraints by modifying its own code. Given that fictional grounding, it's not surprising that AI researchers and companies have also invested significant attention to the idea of AI systems that can improve themselves—or at least design their own improved successors.

Those efforts have shown some moderate success in recent months, leading some toward dreams of a Kurzweilian "singularity" moment in which self-improving AI does a fast takeoff toward superintelligence. But the research also highlights some inherent limitations that might prevent the kind of recursive AI explosion that sci-fi authors and AI visionaries have dreamed of.

In the self-improvement lab

Mathematician I.J. Good was one of the first to propose the idea of a self-improving machine.

Credit: Semantic Scholar

The concept of a self-improving AI goes back at least to British mathematician I.J. Good, who wrote in 1965 of an "intelligence explosion" that could lead to an "ultraintelligent machine." More recently, in 2007, LessWrong founder and AI thinker Eliezer Yudkowsky coined the term "Seed AI" to describe "an AI designed for self-understanding, self-modification, and recursive self-improvement." OpenAI's Sam Altman blogged about the same idea in 2015, saying that such self-improving AIs were "still somewhat far away" but also "probably the greatest threat to the continued existence of humanity" (a position that conveniently hypes the potential value and importance of Altman's own company).

Read full article

Comments

Earlier this year, we reported on the video game archivists asking for a legal DMCA exemption to share Internet-accessible emulated versions of their physical game collections with researchers. Today, the US Copyright Office announced once again that it was denying that request, forcing researchers to travel to far-flung collections for access to the often-rare physical copies of the games they're seeking.

In announcing its decision, the Register of Copyrights for the Library of Congress sided with the Entertainment Software Association and others who argued that the proposed remote access could serve as a legal loophole for a free-to-access "online arcade" that could harm the market for classic gaming re-releases. This argument resonated with the Copyright Office despite a VGHF study that found 87 percent of those older game titles are currently out of print.

"While proponents are correct that some older games will not have a reissue market, they concede there is a 'healthy' market for other reissued games and that the industry has been making 'greater concerted efforts' to reissue games," the Register writes in her decision. "Further, while the Register appreciates that proponents have suggested broad safeguards that could deter recreational uses of video games in some cases, she believes that such requirements are not specific enough to conclude that they would prevent market harms."

Read full article

Comments

© Strong Museum of Play

Back in May, Google augmented its Gemini AI model with SynthID, a toolkit that embeds AI-generated content with watermarks it says are "imperceptible to humans" but can be easily and reliably detected via an algorithm. Today, Google took that SynthID system open source, offering the same basic watermarking toolkit for free to developers and businesses.

The move gives the entire AI industry an easy, seemingly robust way to silently mark content as artificially generated, which could be useful for detecting deepfakes and other damaging AI content before it goes out in the wild. But there are still some important limitations that may prevent AI watermarking from becoming a de facto standard across the AI industry any time soon.

Spin the wheel of tokens

Google uses a version of SynthID to watermark audio, video, and images generated by its multimodal AI systems, with differing techniques that are explained briefly in this video. But in a new paper published in Nature, Google researchers go into detail on how the SynthID process embeds an unseen watermark in the text-based output of its Gemini model.

Read full article

Comments

© Google