News

eu. 36kr. com
eu. 36kr. com > en > p > 3779195342083337

Elorian AI Raises $55 Million: Exploring the Physical World AGI via "Visual Reasoning"

4+ hour, 24+ min ago  (1113+ words) The capabilities of large AI models have, in some aspects, exceeded those of the average person, such as in programming and mathematics. It is reported that Anthropic has almost achieved 100% AI programming internally. Google's Gemini Deep Think solved 5 out of…...

eu. 36kr. com
eu. 36kr. com > en > p > 3776116648215049

AI and New Materials: The "GPT Moment" and Paradigm Revolution in Materials Science

2+ day, 10+ hour ago  (1160+ words) Yunxiu Capital systematically sorts out the integrated development path and technological trends of AI and material science, analyzes the competitive advantages in this industrial ecosystem for you, and explores the investment opportunities therein. The integration of AI and material science…...

eu. 36kr. com
eu. 36kr. com > en > p > 3767583273829125

Peking University and Llama-Factory Collaborate to Launch Data Flex: Industrial-Grade Dynamic Data Training System

1+ week, 1+ day ago  (1147+ words) As large model training enters the deep - water zone, the key to competition is no longer just "how to adjust model parameters," but is gradually shifting to a more core and difficult - to - systematically - solve problem: What data does the…...

eu. 36kr. com
eu. 36kr. com > en > p > 3767812746937095

AI-Driven De Novo Design of Diverse Small Molecule-Binding Proteins: Korean Team Finds Proteins to Selectively Recognize Stress Hormones

1+ week, 1+ day ago  (520+ words) The relevant research results, titled "Small-molecule binding and sensing with a designed protein family", have been published in Nature Communications. The NTF2 fold is composed of three "-helices and a curved six-stranded "-sheet. These structures together form the large internal binding…...

eu. 36kr. com
eu. 36kr. com > en > p > 3767822090777088

Can Humans Control AI? Anthropic's Experiment with Qianwen

1+ week, 1+ day ago  (1747+ words) If one day, AI becomes smarter than humans, what on earth should we organisms do? If they turn around and try to wipe us out, how can we resist? Similar questions have been discussed in various science - fiction movies, but…...

36 Kr
eu. 36kr. com > en > p > 3766418380816899

Transforming Scattered Data into an AI Platform: My Journey

1+ week, 2+ day ago  (956+ words) A hierarchical architecture that integrates dozens of heterogeneous data sources, implements artificial intelligence, and feeds the results back into the entire operation, with zero risk and no need to rewrite any line of code in the original system. When a…...

36 Kr
eu. 36kr. com > en > p > 3760808720777986

Beyond Agents and World Models: Meta Team Proposes "Neural Computer"

1+ week, 6+ day ago  (317+ words) Recently, a research team from Meta AI and KAUST proposed a new machine form: Neural Computer. This concept aims to overcome the fundamental limitations of current AI systems at the execution level: the separation of the model from the computing…...

36氪
eu. 36kr. com > en > p > 3755298778792705

Harnessing New Intelligence: How to Respond When AI Attempts Reverse Fine - Tuning of Humans

2+ week, 2+ day ago  (1768+ words) "Everyone is asking me if I should quickly study lobsters. Actually, I think most people are using it wrongly." "Product managers in the AI field shouldn't study lobsters. They should study the entire system of Anthropic. I think this is…...

36 Kr
eu. 36kr. com > en > p > 3750643236782855

The First Year of Physical AI: A Trillion - Dollar Gamble on "How the World Works"

2+ week, 6+ day ago  (956+ words) In March 2026, AMI Labs, founded by Yann Le Cun, a Turing Award winner and former chief AI scientist at Meta, announced the completion of a $1. 03 billion seed round of financing. Almost at the same time: These events did not occur…...

36氪
eu. 36kr. com > en > p > 3749147549401605

Meituan Long Cat-Next: Convert Images, Sounds, and Texts into Tokens - What's Next?

3+ week, 13+ hour ago  (1216+ words) This is a discrete native autoregressive multimodal large model built on the Long Cat-Flash-Lite Mo E architecture. It has a total of 68. 5 B parameters and only 3 B active parameters, capable of simultaneously processing three modalities: text, images, and audio within…...