OpenAI's new GPT-5.4 clobbers humans on pro-level work in tests - by 83%

GPT-5.4 is also more reliable, producing 18% fewer errors and 33% fewer false claims than GPT-5.2, according to OpenAI.

I used Nano Banana 2 to make perfect sketchnotes: 5 lessons learned (and how I broke Gemini)

The results were fascinating, impressive, and sometimes surprisingly bad. Here are five tips that can help you get better results faster.
Nature

AI can write genomes — how long until it creates synthetic life?

The Evo2 genomic language model can generate short genome sequences, but scientists say further advances are needed to write ...
GitHub

circuit-synth

What is Code-Based Circuit Design? Circuit-synth brings software engineering practices to hardware design by letting you define circuits in Python code instead of ...
IEEE

Topo-Geometrically Distinct Path Computation Using Neighborhood-Augmented Graph, and Its Application to Path Planning for a Tethered Robot in 3-D

Abstract: Many robotics applications benefit from being able to compute multiple geodesic paths in a given configuration space. Existing paradigm is to use topological path planning, which can compute ...
IEEE

Vehicle Trajectory Prediction Using Hierarchical LSTM and Graph Attention Network

Abstract: Vehicle trajectory prediction (VTP) is an important task that can enhance the safety and efficiency of autonomous driving. However, existing VTP methods often struggle to fully extract ...