philip lelyveld The world of entertainment technology

Artificial Intelligence in 2030

...

Some industry leaders, including Alexander Karp, the chief executive of Palantir Technologies, have argued that the U.S. needs a program to accelerate development of A.I. technology, similar to how it established the Manhattan Project to develop nuclear weapons, to keep it from falling behind the rest of the world. At the DealBook Summit, Marc Raibert, the founder of Boston Dynamics, the robotics company, disagreed. “It seems to me we have about three or four or five of them already if you look at the big companies who are investing 10s or 20s or 30s of billions of dollars in it,” he said, referring to the handful of companies building generative A.I. models, which includes Meta, Google and OpenAI, which is spending more than $5.4 billion a year to develop A.I.

Eugenia Kuyda, the founder of Replika, an A.I. companion company, said that if the U.S. government wanted to accelerate A.I. research, it should start by making it easier for A.I. scientists to immigrate. ...

In another live poll, six of the 10 panelists indicated they believed A.I. will create more jobs than it destroys. ... But the vision of widespread economic prosperity that some think A.I. puts within reach isn’t a given.  ...

President Trump, in his previous term, tried to push the limits in a bunch of different ways, tried to tell people underneath him to do things that were norm violating or illegal, and they pushed back. If all those people under him were instead 10 times as brilliant, but perfectly loyal — programmed to be perfectly loyal — that could be a destabilizing situation. ...

One immediate fear cited by Hinton, the Nobel Prize-winning researcher, is that A.I. will flood the internet with so much false content that most people will “not be able to know what is true anymore.” ... “A.I. slop” could increase the value of things that are created by humans.  ...

See the full story here: https://www.nytimes.com/2024/12/19/business/dealbook/artificial-intelligence-in-2030.html

Need a research hypothesis? Ask AI.

... MIT researchers have created a way to autonomously generate and evaluate promising research hypotheses across fields, through human-AI collaboration. In a new paper, they describe how they used this framework to create evidence-driven hypotheses that align with unmet research needs in the field of biologically inspired materials. ...

“There’s a lot of stuff you can do without having to go to the lab,” Buehler says. “You want to basically go to the lab at the very end of the process. The lab is expensive and takes a long time, so you want a system that can drill very deep into the best ideas, formulating the best hypotheses and accurately predicting emergent behaviors. Our vision is to make this easy to use, so you can use an app to bring in other ideas or drag in datasets to really challenge the model to make new discoveries.”

See the full story here: https://news.mit.edu/2024/need-research-hypothesis-ask-ai-1219

Gemini 2.0 Flash Thinking Experimental

...

Reasoning models, like Gemini 2.0 Flash Thinking Experimental, approach problem-solving differently than traditional large language models (LLMs). Given a prompt, these models pause to consider related questions and explain their reasoning before summarizing what they believe to be the most accurate response. Jeff Dean, Chief Scientist for Google DeepMind, noted that this model was “trained to use thoughts to strengthen its reasoning,” and said the company sees promising results when increasing inference time computation—essentially giving the model more time and resources to process queries.

The trade-off? Time. Unlike conventional LLMs, reasoning models often take significantly longer to respond—sometimes seconds or even minutes. ...

From Shelly Palmer daily newsletter

This is where the data to build AI comes from

They audited nearly 4,000 public data sets spanning over 600 languages, 67 countries, and three decades. The data came from 800 unique sources and nearly 700 organizations. ...

Today, most AI data sets are built by indiscriminately hoovering material from the internet. Since 2018, the web has been the dominant source for data sets used in all media, such as audio, images, and video, and a gap between scraped data and more curated data sets has emerged and widened. ...

See the full story here: https://www.technologyreview.com/2024/12/18/1108796/this-is-where-the-data-to-build-ai-comes-from/

Orion glasses: Meta’s bold leap into augmented reality’s future

...

Meta, the owner of Facebook, has introduced its new invention – Orion Glasses. This augmented reality prototype aims to revolutionize the way we use AR technology. ...

Meta emphasizes gesture control, voice commands, and neural wrist interfaces to ensure maximum user comfort. ...

See the full story here: https://www.msn.com/en-ca/lifestyle/shopping/orion-glasses-metas-bold-leap-into-augmented-realitys-future/ar-AA1rgxSE?apiversion=v2&noservercache=1&domshim=1&renderwebcomponents=1&wcseo=1&batchservertelemetry=1&noservertelemetry=1

AI ENTERTAINMENT STUDIOS: HOW GEN AI TOOLSETS ARE TRANSFORMING PRODUCTION WORKFLOWS

• AI studios use off-the-shelf tools but are also developing in-house workflow solutions to streamline production
• Studios are pursuing “hybrid” production paths that still purposefully incorporate human artists and their creative work 
• Still, they expect using generative AI in a workflow to diminish traditional previsualization and post-production stages

See the full story here; https://variety.com/vip/ai-entertainment-studios-how-gen-ai-toolsets-transforming-production-workflows-1236252190/

Report on deepfakes: what the Copyright Office found and what comes next in AI regulation

December 18, 2024 - On July 31, 2024, the Copyright Office released part one of its Report on Copyright and Artificial Intelligence, specifically addressing the topic of digital replicas, or "deepfakes" — i.e., AI-generated video, image, or audio recordings that realistically but falsely depict an individual. The report is the result of a broad initiative to explore the intersection of copyright and AI, informed by a series of listening sessions and meetings with stakeholders, as well as more than 10,000 public comments from authors, artists, publishers, lawyers, academics, industry groups, and more.

The report's conclusions are stark: It finds that existing laws, in copyright and other intellectual property areas, are vastly insufficient to redress the harm posed by unauthorized digital replicas, which have the potential to threaten not only those in entertainment and politics, but private individuals, too. ...

 Most alarming, the report warns that digital replicas pose a danger to our political system and news reporting "by making disinformation impossible to discern." ...

The report found existing federal laws largely inapplicable. For example, it is black-letter law that copyright does not "protect an individual's identity in itself, even when incorporated into a work of authorship." Thus, while it might be a copyright violation to reproduce a copyrighted image or song that contains the copyright owner's likeness or voice, merely replicating someone's image or voice in a deepfake would not implicate copyright protections.

The Lanham Act "require[s] proof of commercial use and a likelihood of consumer confusion," so it is not useful in cases of harmful but personal and non-commercial deepfakes. ...

Meanwhile state laws were deemed uneven and often too narrow. ...

Based on this analysis, the Copyright Office did not mince words as to its conclusion. It stated in blunt terms: "new federal legislation is urgently needed." Looking ahead, the call for action is likely to accelerate congressional focus on an issue that has already sparked numerous pieces of draft legislation. ...

See the full story here: https://www.reuters.com/legal/legalindustry/report-deepfakes-what-copyright-office-found-what-comes-next-ai-regulation-2024-12-18/

The 51 most disruptive startups of 2024

See the full story with brief summaries of each startup here: https://techcrunch.com/2024/12/13/the-51-most-disruptive-startups-of-2024/#P

Memo to AI: Does it hurt when we pull your plug?

... Fish’s report ties into the idea of artificial general intelligence (AGI) a goal of AI reaching human-level abilities—thinking, learning and even potentially feeling. OpenAI’s search for an “AI Welfare Specialist” signals a growing concern: if AI ever becomes conscious, how should we treat it? ...

The discussion centers on a question that has long challenged philosophers and computer scientists alike: what defines consciousness? Smolinski points to philosophical perspectives that define consciousness as the ability to independently affect oneself and the surrounding world. Current AI systems, he notes, operate purely by reacting to external inputs, rather than through independent agency. ...

Researchers exploring AI consciousness propose borrowing tools from animal cognition studies to hunt for signs of machine awareness, while conceding that definitively proving artificial consciousness remains a formidable challenge. ...

See the full story here: https://www.ibm.com/think/news/ai-welfare-debate

Where Humans Still Have the Edge on AI [HBR]

...

Areas Where Humans Still Have an Edge Over AI

In speaking to hundreds of experts, consumers, and skeptics of AI over the past few years, four strongholds for humans keep coming up:

• Emotion: Understanding, connecting with, and responding sensitively to human feelings.
• Complexity: Navigating ambiguous, broad-context challenges with holistic problem-solving.
• Physicality: Tasks requiring dexterity and interaction with the physical world, particularly where human presence and responsiveness matter.
• Creativity: The ability to generate original, novel ideas and solutions.

...

Despite the fanfare surrounding the launch of new AI media-generating technologies, we don’t care much for it as consumers. Spotify now includes AI music in its catalog, but all the big hits are by flesh-and-blood musicians. Very few people are watching AI-generated films, nor do they want to. And although an AI-generated painting recently sold for $1 million, it’s a long way off the $450 million paid for da Vinci’s Salvator Mundi. We value history and scarcity, and these two qualities are beyond AI. ...

See the full story here: https://hbr.org/2024/12/where-humans-still-have-the-edge-on-ai