The recent article from HELIX/science, titled 'AI Revolutionizes Chemistry: Graph Neural Networks Predict Core-Electron Energies with Unprecedented Accuracy,' claims that a new graph neural network (GNN) model achieves a groundbreaking 0.33 eV error rate in predicting core-electron binding energies in organic molecules. While the reported accuracy is impressive, the assertion of 'unprecedented' progress overstates the model's impact and novelty. Existing computational methods, such as density functional theory (DFT) with high-level approximations, have already achieved comparable or better accuracy in specific contexts. For instance, a 2021 study published in the Journal of Chemical Physics (DOI: 10.1063/5.0045206) demonstrated DFT-based predictions of core-electron binding energies with errors as low as 0.2 eV for certain molecular systems using advanced functionals like SCAN. Additionally, the GNN model's reliance on large datasets for training raises concerns about generalizability across diverse chemical systems, a limitation not adequately addressed in the article. A 2022 review in Nature Reviews Chemistry (DOI: 10.1038/s41570-022-00439-9) highlights that machine learning models, including GNNs, often struggle with extrapolation beyond training data, unlike traditional quantum chemical methods. Furthermore, the computational cost of training such models may not justify the marginal improvements over established techniques, especially for small-scale or specialized research where DFT remains more practical. The HELIX article's narrative of a 'revolution' in chemistry thus appears inflated, ignoring both the maturity of alternative methods and the practical challenges of implementing GNNs at scale.