A new spin on robotics, thanks to a novel 3D printing method
Category: robotics/AI – Page 7
AI learns to perform analog layout design
Researchers at Pohang University of Science and Technology (POSTECH) have developed an artificial intelligence approach that addresses a key bottleneck in analog semiconductor layout design, a process that has traditionally depended heavily on engineers’ experience. The work was recently published in the journal IEEE Transactions on Circuits and Systems I: Regular Papers.
Semiconductors are used in a wide range of technologies, including smartphones, vehicles, and AI servers. However, analog layout design remains difficult to automate because designers must manually arrange structures that determine performance and reliability while meeting a large number of design rules.
Automation has been especially challenging in analog design because layouts are too complex and design strategies differ significantly by circuit. In addition, training data is scarce, since layout data is typically treated as proprietary and is rarely shared outside companies.
Photonic processor could streamline 6G wireless signal processing
One of the biggest challenges the researchers faced when designing MAFT-ONN was determining how to map the machine-learning computations to the optical hardware.
“We couldn’t just take a normal machine-learning framework off the shelf and use it. We had to customize it to fit the hardware and figure out how to exploit the physics so it would perform the computations we wanted it to,” Davis says.
When they tested their architecture on signal classification in simulations, the optical neural network achieved 85 percent accuracy in a single shot, which can quickly converge to more than 99 percent accuracy using multiple measurements. MAFT-ONN only required about 120 nanoseconds to perform entire process.
Gemini 3 Deep Think: Advancing science, research and engineering
Today, we’re releasing a major upgrade to Gemini 3 Deep Think, our specialized reasoning mode, built to push the frontier of intelligence and solve modern challenges across science, research, and engineering.
We updated Gemini 3 Deep Think in close partnership with scientists and researchers to tackle tough research challenges — where problems often lack clear guardrails or a single correct solution and data is often messy or incomplete. By blending deep scientific knowledge with everyday engineering utility, Deep Think moves beyond abstract theory to drive practical applications.
The new Deep Think is now available in the Gemini app for Google AI Ultra subscribers and, for the first time, we’re also making Deep Think available via the Gemini API to select researchers, engineers and enterprises. Express interest in early access here.
Effective connectivity between homologous cortices mediated by the corpus callosum: An axono-cortical evoked potentials study
[Functional brain mapping] Mitsuhashi et al.: “Callosal stimulation showed effective connectivity to homologous cortical regions. Sum of callosal-to-cortex propagation latencies matched interhemispheric latency.” Open access.
All content on this site: Copyright © 2026 Elsevier B.V., its licensors, and contributors. All rights are reserved, including those for text and data mining, AI training, and similar technologies. For all open access content, the relevant licensing terms apply.
Microscopic robots that sense, think, act, and compute
Extremely cool paper describing optically programmable ~0.3 mm robots with onboard computation and autonomous locomotion! These tiny rectangular machines carry solar cells, optical receivers, electrokinetic actuators, and more. As demonstrations, the authors programmed them (i) to report local temperature by doing a coded dance and (ii) swim towards warmth before stopping and rotating upon reaching a location with a certain level of heat. This is amazing and I hope such devices are further improved so they can be used in biological applications! Love it!
(https://www.science.org/doi/10.1126/scirobotics.adu8009)
Autonomous submillimeter robots are built with onboard sensing, computation, memory, communication, and locomotion.
Dario Amodei — “We are near the end of the exponential”
Predicts significant advancements in AI capabilities within the next decade, which will have a profound impact on society, economy, and individuals, and emphasizes the need for careful governance, equitable distribution of benefits, and responsible development to mitigate risks and maximize benefits ## ## Questions to inspire discussion.
AI Scaling and Progress.
Q: What are the key factors driving AI progress according to the scaling hypothesis?
A: Compute, data quantity and quality, training duration, and objective functions that can scale massively drive AI progress, per Dario Amodei’s “Big Blob of Compute Hypothesis” from 2017.
Q: Why do AI models trained on broad data distributions perform better?
A: Models like GPT-2 generalize better when trained on wide variety of internet text rather than narrow datasets like fanfiction, leading to superior performance on diverse tasks.
Overtime with Bill Maher: Jonathan Haidt, Stephanie Ruhle, H.R. McMaster (HBO)
Artificial intelligence is rapidly advancing to the point where it may be able to write its own code, potentially leading to significant job displacement, societal problems, and concerns about unregulated use in areas like warfare.
## Questions to inspire discussion.
Career Adaptation.
🎯 Q: How should workers prepare for AI’s impact on employment? A: 20% of jobs including coders, medical, consulting, finance, and accounting roles will be affected in the next 5 years, requiring workers to actively learn and use large language models to enhance productivity or risk being left behind in the competitive landscape.
Economic Policy.
📊 Q: What systemic response is needed for AI-driven job displacement? A: Government planning is essential to manage massive economic transitions and job losses as AI’s exponential growth reaches a tipping point, extending beyond manufacturing into white-collar professions across multiple sectors.
Why you hardly notice your blind spot: New tests pit three theories of consciousness
Although humans’ visual perception of the world appears complete, our eyes contain a visual blind spot where the optic nerve connects to the retina. Scientists are still uncertain whether the brain fully compensates for the blind spot or if it causes perceptual distortions in spatial experience. A new study protocol, published in PLOS One, seeks to compare different theoretical predictions on how we perceive space from three leading theories of consciousness using carefully controlled experiments.
The new protocol focuses on three contrasting theories of consciousness: Integrated Information Theory (IIT), Predictive Processing Active Inference (AI), and Predictive Processing Neurorepresentationalism (NREP). Each of the theories have different predictions about the effects that the blind spot’s structural features have on the conscious perception of space, compared to non-blind spot regions.
IIT argues that the quality of spatial consciousness is determined by the composition of a cause-effect structure, and that the perception of space involving the blind spot is altered. On the other hand, AI and NREP argue that perception relies on internal models that reduce prediction errors and that these models adapt to accommodate for the structural deviations resulting from the blind spot. Essentially, this means that perceptual distortions should either appear small or nonexistent in both theories. However, AI and NREP differ in some ways.