C_H_R_I_S - Network

Interesting ...https://garymarcus.substack.com/p/irony-too-funny-for-words#ai

Interesting ...

https://garymarcus.substack.com/p/irony-too-funny-for-words

#ai

Gary Marcus’s Take on DeepSeek: https://garymarcus.substack.com/p/five-things-most-people-dont-seem#ai

Gary Marcus’s Take on DeepSeek: https://garymarcus.substack.com/p/five-things-most-people-dont-seem

#ai

Testing AI models seems to be becoming an increasingly complicated...

Testing AI models seems to be becoming an increasingly complicated topic.

https://time.com/7203729/ai-evaluations-safety/

https://www.computing.co.uk/news/2025/ai/ai-models-cheating-benchmark-tests

#ai

“We’re at this tipping point for AI in biology" (Eran Segal,computational biologist at the Weizmann Institute of...

“We’re at this tipping point for AI in biology" (Eran Segal,computational biologist at the Weizmann Institute of Science).

https://www.theatlantic.com/technology/archive/2025/01/generative-ai-virtual-cell/681246/

#ai

Will AI boost productivity? If so, when? A podcast from MIT Technology Review sheds light on the pace of diffusion and what it depends...

Will AI boost productivity? If so, when? A podcast from MIT Technology Review sheds light on the pace of diffusion and what it depends on.

https://open.spotify.com/episode/7JT07ngGUVrw0XegHH9VpI

#ai

Nvidia and others are set to revolutionize non-playable characters in games. Beyond entertainment, this innovation could unlock fascinating...

Nvidia and others are set to revolutionize non-playable characters in games. Beyond entertainment, this innovation could unlock fascinating possibilities for museums and other spaces that mediate content. I wonder how AI-powered agents, featuring diverse characters, might contribute to more engaging and thought-provoking experiences in exhibitions or virtual spaces at institutions focused on art, science, or education.

https://www.nvidia.com/en-us/geforce/news/nvidia-ace-autonomous-ai-companions-pubg-naraka-bladepoint/

#ai

„The AI apocalypse is here and it’s far stupider and more depressing than we were promised. Instead of being hunted down by a gleaming...

„The AI apocalypse is here and it’s far stupider and more depressing than we were promised. Instead of being hunted down by a gleaming metal skeleton in a post-apocalyptic wasteland, we are surrounded by zombies endlessly repeating our own posts back to us.“

https://gizmodo.com/author/mjgault

#ai

Gary Marcus has 25 AI predictions for 2025: https://garymarcus.substack.com/p/25-ai-predictions-for-2025-from-marcus#ai

Gary Marcus has 25 AI predictions for 2025: https://garymarcus.substack.com/p/25-ai-predictions-for-2025-from-marcus

#ai

"If trust is our north star as we navigate this new landscape, agents will empower us to make a meaningful impact on an unprecedented...

"If trust is our north star as we navigate this new landscape, agents will empower us to make a meaningful impact on an unprecedented scale."

When I read about the potentials of the 'agentic era,' it comes across as PR. If there’s a lesson to take from history, it’s that new technologies are rarely deployed for the benefit of society as a whole. Big Tech in 2024 shows this trend has only intensified.

Nevertheless, here’s Marc Benioff’s article for TIME: https://time.com/7178872/agents-unlimited-age/

#ai

“The true power of AI will be unlocked when we achieve truly autonomous agents capable of collaborating at scale."Robert Yang,...

“The true power of AI will be unlocked when we achieve truly autonomous agents capable of collaborating at scale."

Robert Yang, founder of the AI startup Altera, envisions large-scale AI civilizations. He experiments with AI-controlled characters in Minecraft that exhibit emergent behaviors such as becoming sociable or introverted, specializing in specific tasks, and evolving over time.

Here's Niall Firth's article for MIT Technology Review: https://www.technologyreview.com/2024/11/27/1107377/a-minecraft-town-of-ai-characters-made-friends-invented-jobs-and-spread-religion/

#ai

Red-teaming was inspired by cybersecurity, where experts try to identify vulnerabilities in computer systems. Since 2022, red-teaming has...

Red-teaming was inspired by cybersecurity, where experts try to identify vulnerabilities in computer systems. Since 2022, red-teaming has evolved into a collaborative effort between humans and LLMs. Critics argue that these efforts lag behind the pace of developing and deploying new models—and that red-teaming general-purpose systems is ultimately infeasible because their full range of capabilities is impossible to predict.

Here's an article by Will Douglas Heaven: https://www.technologyreview.com/2024/11/21/1107158/how-openai-stress-tests-its-large-language-models/

#ai

Can a two-hour—or even just a 30-minute—interview really provide enough information about a person to create their digital twin? It...

Can a two-hour—or even just a 30-minute—interview really provide enough information about a person to create their digital twin? It seems some people may underestimate the complexity and fluidity of what it means to be human, encompassing all our experiences, values, beliefs, and creativity. Humans are inherently unpredictable and tend to change significantly over the course of their lives, don’t they?

Read James O'Donnells article for MIT Technology Review: https://www.technologyreview.com/2024/11/20/1107100/ai-can-now-create-a-replica-of-your-personality/

#ai

With tools like "Gaze" and "Nightshade," scientists and artists are fighting back against tech companies who scrape...

With tools like "Gaze" and "Nightshade," scientists and artists are fighting back against tech companies who scrape their work without permission or compensation.

Read Melissa Heikkilä's article for the MIT Technology Review: https://www.technologyreview.com/2024/11/13/1106837/ai-data-posioning-nightshade-glaze-art-university-of-chicago-exploitation/

#ai

"Just adding data and compute and training longer worked miracles for a while, but those days may well be over."Gary Marcus...

"Just adding data and compute and training longer worked miracles for a while, but those days may well be over."

Gary Marcus reflects on rumors about the end of "scaling laws" in his Substack.: https://garymarcus.substack.com/p/an-ai-rumor-you-wont-want-to-miss

#ai

"Controlling search means controlling how most people access every other digital thing—it’s an incredible platform to gain trust...

"Controlling search means controlling how most people access every other digital thing—it’s an incredible platform to gain trust and visibility, advertise, or influence public opinion."

AI-driven search "could completely reorient our relationship to knowledge, prioritizing rapid, detailed, and abridged answers over a deep understanding and the consideration of diverse sources and viewpoints," says Matteo Wong in his article for The Atlantic: https://www.theatlantic.com/technology/archive/2024/11/ai-search-engines-curiosity/680594/

#ai

"[...] the next Beatles will play with KLAY"Universal Music Group (UMG) partners with Klay Vision to create an "ethical"...

"[...] the next Beatles will play with KLAY"

Universal Music Group (UMG) partners with Klay Vision to create an "ethical" foundation model for AI music generation: https://www.theverge.com/2024/10/28/24282030/universal-music-group-partners-with-klay-ai-company-ethical-music-generator-foundational-model

#ai

"In a few thousand days, superintelligent virtual experts will create almost anything we can imagine," says Sam Altman, who is...

"In a few thousand days, superintelligent virtual experts will create almost anything we can imagine," says Sam Altman, who is seeking VC.

"Powerful AI will likely resemble today’s LLMs and will defeat most diseases, expand our biological and cognitive freedom, lift billions out of poverty, and spark a renaissance of liberal democracy and human rights," says Dario Amodei, who is also seeking VC.

"Today's LLMs are not smarter than a house cat," says Yann LeCun, who doesn’t need VC.

#ai

Finance for European AI startups in the scale-up and growth phases consists of a mixture of local billionaires, public funding through...

Finance for European AI startups in the scale-up and growth phases consists of a mixture of local billionaires, public funding through national development banks, existing sectoral champions, corporate venture investments, and international (primarily US) venture capital.

(This picture was taken from the report: AI Now - Redirecting Europe’s AI Industrial Policy: From Competitiveness to Public Interest: https://ainowinstitute.org/wp-content/uploads/2024/10/AI-Now_EU-Industrial-Policy_October-2024-1.pdf).

#ai

Already a key player in the global semiconductor industry, Tawain is aiming to become a hub for innovation in advanced AI. There's a...

Already a key player in the global semiconductor industry, Tawain is aiming to become a hub for innovation in advanced AI. There's a strategic plan to strengthen five key sectors: semiconductors, AI, defense, security, and next-generation communications.

Here's Yu-Tzu Chiu's article for IEEE Spectrum: https://spectrum.ieee.org/ai-in-taiwan

#ai

“Everything I’m saying could very easily be wrong […] All of these are somewhat vague ideas […] They may be unrealistically utopian...

“Everything I’m saying could very easily be wrong […] All of these are somewhat vague ideas […] They may be unrealistically utopian […].”

Although Dario Amodei is obviously unsure if he is right, he predicts that AI will contribute to "the defeat of most diseases, the growth in biological and cognitive freedom, the lifting of billions of people out of poverty, and a renaissance of liberal democracy and human rights."

Here’s his 14,000-word essay: https://darioamodei.com/machines-of-loving-grace#taking-stock

#ai