I definitely need to check out your recent chat with Henri
I definitely need to check out your recent chat with Henri
Yes, Iβm familiar with that. Your talks are gold! π€©
Learning 6 languages at once? Wow! Thatβs parallel processing at its finest π Iβve never heard of most apps in that list, so thanks! At least when it comes to app-based learning, I was pretty satisfied with Babel and donβt have an urgent need to try something else.
Ever give Babbel a try? It was the best language learning app I've come across after trying a handful.
It was legendary! π
I've been working on RUM-based alerting and got to share some lessons learned at Performance.sync() in Amsterdam this week. Here are my slides: screenspan.net/talks/noise-...
Last Wednesday I had the pleasure of speaking at the performance.sync() organized by @mozilla.org in Amsterdam.
My talk was a 15-minute session where I shared practical techniques for debugging and optimizing slow interactions on websites.
Post and slides: www.andreaverlicchi.eu/blog/improvi...
- CSS Containment can be used to reduce unnecessary rendering work. (Check out my blog post www.speedkit.com/blog/field-t... on a test we did last year, and that we should re-test)
I canβt wait for next yearβs Performance.now()!
* The Speculation Rules API is already proving very effective in improving load times in Chrome. But devs need to be aware of pitfalls. (In this vid youtu.be/rz6mpodOPy4?..., my colleague @erikwitt.bsky.social explains how we leverage RUM data for better precision with Speed Kit)
- The Long Animation Frames API in Chrome can be super helpful for identifying bottlenecks in rendering, execution and interaction. That data can also be used to see how user engagement or business metrics are affected. (Iβve found LoAF to be useful for identifying processing hogs)
- Some browser vendors are quicker than others in developing features and tooling aimed at perf optimization. We as a community need to work together to push for standardization across all browsers by actively requesting and testing new features, submitting bug tickets and pull requests.
Here are some of my takeaways:
- As AI continues to influence how we build the web and users use it, we need to be all the more diligent in avoiding slop and bulk, always aiming to create delightful, accessible experiences.
Tim Kadlec speaking at Performance Now in Amsterdam, 2025
This marks my fifth year attending my favorite conference, #perfnow in Amsterdam. Again, I leave super-charged from the inspiring talks, nerdy chats and unforgettably fun times. The perf community is just so amazing!
Itβs an honor to see my company #speedkit mentioned on stage at #perfnow by a web performance VIP and ambassador such as Barry @tunetheweb.com Pollard. π Congrats to the Speed Kit team who implemented speculation rules in such an effective way in our product!
Fellow #perfnow fellows, a reminder that PerfPlanet calendar is coming this December for another edition. And it needs YOU. Please ask if you have any contribution ideas or concerns calendar.perfplanet.com/contribute/
Yes! I played around with it a while ago to see if CSS Containment budged CWV: www.speedkit.com/blog/field-t...
Would love to run some more tests, however!
Videos from our meetup are now posted! Here's the playlist:
www.youtube.com/playlist?lis...
#webperf
π£ Reminder!!
In 2 days, we have @toddhgardner.com and @inesakrap.bsky.social giving talks in our first meetup! For full details - check out our website (and add to calendar to get the join links)!
performanceobserver.dev
1/ π Prediction 4: AI Takes Over Performance Optimization
@screenspan.net predicts "AI will play an even bigger role in how we surf, build, debug, and optimize the web"
Letβs dive into the key insights for 2025. π§΅π
Very cool that you worked on the INP polyfill. I definititely wanna give that a look π€©
π Real User Monitoring (RUM) is crucial for identifying issues on websites and tracking regressions. But the challenge multiplies when monitoring hundreds or thousands of sites. I got to share insights into our collaborative efforts in Amsterdam in Nov.: www.youtube.com/watch?v=XBsv...
Solid, Barry! πͺπ½
Nice sleuthing, Harry! π€© Good points to keep in mind when debugging LCP. Looks like Lighthouse would have called out the <article> (and a massive render-blocking CSS file) as well.
Needless to say as well: the gobs of JS also bog down the main thread and make interactions laggy.
Iβm not on the olβ personal computer right now, but judging from these WebPageTest repeat view runs, it looks like the fonts are not being loaded from cache. www.webpagetest.org/result/24120...
I reached out to support about that
Hi support@bsky.app, I've been enjoying BSKY :-)
Just wanted to point out something that @kojordan.com spotted: the app is loading TTF fonts, which are very large. You should be using WOFF2 β and if you're preloading them, you need the "crossorigin" attribute :-)
Thanks! π Geesh, you're right β massive font and JS files. I wonder why they're serving TTF...