Un site lent à cause d’images trop lourdes ? Découvrez comment Nero AI Image Compressor réduit leur taille sans perte visible grâce à l’intelligence artificielle. Qui veut booster son site ? #SEO #WebPerf #IA #Archive
@byothe.fr
@byothe.fr
@bra31k.bsky.socialA/B with ~18k users per arm: the slowest users gained the most, ~670 ms at p50 and ~1.9 s at p90. It cost ~120 ms on the loader, which hover prefetch later paid back.
bra31k.dev/en/blog/same-bytes-17-percent-faster
#webdev #webperf
@inautilo.bsky.social
@nucliweb.netDuring a perf audit for a client, I read this in a PR: "The module should only be loaded once, not every time a file is selected." When I read it, something didn't add up, so I started digging. And the result surprised me.
I wrote a post joanleon.dev/en/dynamic-i...
@webperfreport.bsky.social⚡Featured posts this week:
✨ Improving site performance by shipping more CSS: @josh.black
✨ How AI Agents Use Websites: @giacomozecchini.bsky.social
✨ Chasing the Perfect Website: Morgan Murrah
✨ Performance at Scale with Ryan Townsend: @twnsnd.com
🙏 Special thanks to the authors!
#webperf
@webperfreport.bsky.social🎉 WebPerformance Report Week #40 is out! ✅
Join our growing community:
⚡ Performance Report
🛡️ Security Report
♿ Accessibility Report
📊 Analytic Report
🔎 SEO Report
Explore our report catalog 👉 webperformancereport.com#Catalog
@baremind.healthWhen the streak tree in my app is fully grown, it’s drawn with about 590 watercolour SVG filters. Scrolling past it repainted every one of them, every frame. The fix was one line: translateZ(0) on the tree’s svg, so scrolling moves a finished picture.
@webperfreport.bsky.socialThe Technical Diagnostics and Business Outcomes Gap
When tools do not connect, people do.
And too often, senior specialists become human APIs between disconnected systems.
New WPR article:
webperformancereport.com/blog/technic...
@etechinvest.bsky.social
@nucliweb.netOne thing I’m experimenting with for a performance audit:
Using RUM data + @debugbear.com MCP to build synthetic tests based on how the product is actually used
RUM → real user behaviour → MCP → representative synthetic tests
@davatron5000.bsky.socialAny day I get to talk to people who enjoy and think about #webperf is a good day. Enjoyed talking with @benschwarz.bsky.social this week on the podcast.
@twnsnd.comSo happy to share mine & @nicj.net's first @cloudflare.social blog post—during our 16th Birthday Week no less! 🥳
Today we announce BEACON, a #webperf research dataset we're contributing to the @rumarchive.com community.
@etechinvest.bsky.social
@webperfreport.bsky.social⚡Featured posts this week:
✨ AI Reporting Workflow: @webperfreport.bsky.social
✨ How we made claude.ai 3x faster: Raymond Wang.
✨ Liquid storefronts start loading 29% faster: @krzksz.bsky.social
✨ Blurry before beautiful: @patrickbrosset.com
🙏 Special thanks to the authors!
#webperf
@webperfreport.bsky.social🎉 WebPerformance Report Week #39 is out! ✅
Join our growing community:
⚡ Performance Report
🛡️ Security Report
♿ Accessibility Report
📊 Analytic Report
🔎 SEO Report
Explore our report catalog 👉 webperformancereport.com
@sugataai.bsky.socialWe stopped shipping less CSS and started shipping more to make GitHub load faster. #webperf sugataai.com/blog/improvi...
@webperfreport.bsky.socialWe gave Claude Opus 5.5 an article we wrote a few months ago and asked it to turn it into an animation. 👇
📖 The Anatomy of a Web Performance Report
calendar.perfplanet.com/2025/the-ana...
Thanks for the idea, @addyosmani.bsky.social!
@inautilo.bsky.social#Development #Releases
New in Chrome DevTools · Agent controls, modernized device presets, and more ilo.im/16gfer by Matthias Rohmer
_____
#Chrome #DevTools #AI #MCP #Devices #Network #Accessibility #WebPerf #WebDev #Frontend
@josephscott.bsky.socialOne more 👍 - the progressive rendering that I've seen from both AVIF and JXL look good, with those widely supported I'd ditch any other color blur pre-fill approach. #webperf
@lionel-peramo.comYou use Promises for every async task.
Tomorrow, you might use scheduler.postTask instead.
They handle tasks in very different ways.
Here is how they change web performance.
1/5
