Tests on 14 datasets show that TabPFN and TabICL can match or beat tuned boosting models using in-context learning, potentially eliminating the need for hyperparameter tuning.
Eric Gullichsen, who has served on NVIDIA's Technical Advisory Board since 1993, is seeking a legal resolution from the company, claiming he lost massive stock options due to contractual errors.
Hitachi, Ltd. is introducing Forward Deployed Engineers (FDEs), who work directly at customer sites to handle everything from requirement extraction to implementation and value creation. As AI streamlines system development, there is an increasing demand for proactive roles that solve customer challenges together.
A new implementation uses a genetic algorithm to evolve a car's genome, enabling self-parking capabilities through iterative optimization of 180-bit DNA.
Accounting automation platform Basis observed a 50% reduction in task completion time and a 20% increase in internal evaluation scores when using GPT-6 Astra compared to previous models.
Apple researchers have introduced the Local Inexact Proximal Point Algorithm with Extra Step (LIPPAX) to address convergence gaps and client drift in stochastic variational inequalities.
IDC research suggests that AI agents are driving a structural evolution within enterprise applications by running inside existing systems of record, rather than causing a "SaaS apocalypse."
Tsukuba Memorial Hospital is working to prevent turnover by using the organizational analysis AI "MediTune" to visualize workplace issues within each department. As a result, the nursing staff turnover rate decreased from 16.2% in FY2021 to 6.7% in FY2025.
Lister, a data provider for the manufacturing sector, has released a report analyzing the roles of companies involved in the implementation of Physical AI and their relationship with various industries.
At an event in Japan, a US HP executive discussed the company's strategic direction, highlighting the growth of AI PCs and the necessity of integrated security to prevent data leaks.
Alibaba's Damo Academy has released "RADAR," an open-source medical vision-language model capable of identifying approximately 150 types of abdominal diseases from CT scan images. By employing a learning method that does not require manual annotation, it demonstrates high diagnostic performance for a variety of diseases, including cancer.
According to a U.S. federal government factsheet, President Donald Trump and Chinese President Xi Jinping have agreed to refer to the emerging technology as "Super Intelligence (SI)" instead of "Artificial Intelligence (AI)." The two nations have also agreed to establish a "U.S.-China Super Intelligence (SI) Dialogue" regarding the risks and benefits of SI, and to set up a communication channel for incident response.
Yaskawa Electric introduced an AI robot designed to replicate the sensory skills of experienced workers for arc welding, while FANUC demonstrated an AI welding agent.
Engram uses local "tiny AI" models to mangle audio and generate experimental sounds, moving beyond simple text-to-song generation.
Anthropic CEO Dario Amodei is scheduled for a one-on-one dinner with President Donald Trump, marking a potential dialogue amidst conflicting stances on AI safety and regulation.
As LLMs enable developers to build software in hours, the time required for legal setup becomes a significant bottleneck for launching micro-businesses.
Meta's Muse AI agent focuses on consumer applications through social integration, contrasting with the enterprise focus of competitors like OpenAI and Anthropic.
A new library, Imp, ports the DSPy framework to the BEAM, enabling developers to build self-improving, typed language-model programs in Elixir using OTP's concurrency model.
Fireworks Research has introduced Ember-1, a specialized model built on Kimi K3 that delivers equivalent quality with approximately 40% fewer tokens by optimizing reasoning efficiency.
Cast member Jane Wickline performed a comedic impression of Anthropic CEO Dario Amodei on Saturday Night Live, satirizing recent discussions regarding AI safety and existential risks.