I agree. Any testing has value. For software development, with limited time and budget, we have to get the priority right. For example, performance testing is good, unless extreme cases (e.g. Steve Jobs' demand for Safari), we don't run continuous performance testing a few times a day.

My advice to avoid "Visual Regression Testing" is based on nearly all these activities I witnessed, plus Cypress Component Testing alike, were conducted where there was no Automated E2E Functional (via UI) Regression Testing. Then, what's the point? If one tester spend 2 hours on 'Visual Regression Testing' on a build, shortly after, a business analyst or manual tester reports a few functional regression defects, which means that 2-hour "Visual Regression Testing" was wasted.

"QA Engineers, Get your End-to-End (UI) Automated Testing Done Properly Before Talking/Working on Other Blackbox Testing" https://zhiminzhan.medium.com/why-visual-regression-testing-is-a-wrong-approach-698018c80db7

"Verify a Dynamic Chart in Selenium WebDriver"

https://zhiminzhan.medium.com/verify-a-dynamic-chart-in-selenium-webdriver-f7f16ee83d78