JAMA Dermatology continued to thrive in 2019, maintaining an impact factor of 7.99 and our high rank among all dermatology journals. Manuscript submissions rose; the journal received 2942 submissions in the past year, an increase from 2584 in 2018 and 2153 in the prior year (Table). In 2019, 1259 of 2942 submissions (42.8%) were reports of research, including Original Investigations, Brief Reports, and Research Letters. A total of 9% of submissions overall were accepted, reduced from a 14% acceptance rate during the previous year. Reducing the time to publication continues to be an ongoing priority and area for improvement for the journal, with a median reviewer turnaround time of 9 days, median receipt to rejection time of 4 days, and median time from acceptance to publication of 75 days. The editorial leadership team continues to strive to find ways to improve the authors' experience to ensure authors the fastest and best publication of their work. The journal continued to have broad digital reach, with 4.4 million full-text views and PDF downloads in the past year, and connecting with a growing number of followers (more than 51 000) on social media channels Twitter and Facebook. JAMA Dermatology was pleased to be involved in the JAMA Network's arrival on Instagram (@jamanetwork), as we predict that our visual-based content will thrive in that space. The journal received 3700 media mentions this year, another important indicator of the journal's success. Three JAMA Dermatology articles were among the Altmetric top-scoring dermatology articles in 2019, including 2 Original Investigations titled "Association Between Surgical Skin Markings in Dermoscopic Images and Diagnostic Performance of a Deep Learning Convolutional Neural Network for Melanoma Recognition" 1 and "Association of Vitamin A Intake With Cutaneous Squamous Cell Carcinoma Risk in the United States" 2 as well as an Editorial titled "Natural Does Not Mean Safe-The Dirt on Clean Beauty Products." 3 The popularity of the first study is striking, as it is a highly technical computational article. What is the significance of this well-deserved attention? Machine learning is undoubtedly here to stay and likely in time will be a valuable commonplace tool to augment our abilities to provide expert care for cutaneous disease. This work by Winkler et al 1 highlights that it is essential to rigorously evaluate the algorithms used to train machines to detect skin lesions in images, as seemingly innocuous elements-in this case surgical pen markings intended to identify a skin lesion of concern-may significantly depreciate the computer's accuracy. Improved understanding of when machine learning can improve or impede clinical practice, as well as the road to achieving this improved understanding, is thoughtfully discussed in the accompanying Editorial. 4 Research manuscripts received b 1259 Acceptance rate, % Overall 9 Research 9 Peer reviewer turnaround, median, d 9 Receipt to first decision without peer review, median, d 4 Receipt to first decision with peer revi...