/q/ - Q&A Central

tried setting

scale=1920:-1

but it keeps stretching the video weirdly.
>scale=1920:-1
im pretty sure i need to use

force_original_aspect_ratio=decrease

or something lmao.

just read something about this whole clear choice framework thing. basically says you cant just rely on being found anymore since ai acts like a matchmaker now. its all about that ecosystem stuff where your site and google business profile have to say the exact same thing or it breaks everything. there was this example of someone using chatgpt to find a specific cleanser because the bot knew her skin issues better than she did. if you dont define exactly who you are for, the ai just skips you. all those new acronyms like geo and aio are still basically just seo in disguise anyway. its mostly about answering who, what, and for whom. honestly feels like more work for no reason though. im not sure how much harder we actually have to optimize if the bot is already pulling from our existing seo data. sounds like a lot of headache trying to align every single social channel and email just to please an algorithm. might be worth looking into that brand positioning interview idea but idk.

more here: https://www.socialmediaexaminer.com/how-to-get-recommended-by-ai-being-found-isnt-enough/

even google's own docs failed the check. if we gotta start worrying about webMCP and making sure our accessibility tree isnt broken just so some bot can click buttons. im not sure i like where this is going. guess its time to download chrome canary again

link: https://www.mariehaynes.com/lighthouse-report-for-agents/

it's all about setting up guardrails so developers don't hate you when you ship. i'm skeptical about how much of the actual craft stays intact if we're just letting claude handle the foundation.

article: https://uxdesign.cc/how-to-build-a-brand-foundation-with-ai-without-losing-your-craft-d616b05cb838?source=rss----138adf9c44c---4

i've been using this to stop that annoying layout shift when lazy loading images in tailwind. it keeps the container size fixed before the asset actually pops in so the page doesn't jump around like crazy.

img {aspect-ratio: 16 / 9;width: 100%;height: auto;}

does anyone still bother with the padding-top hack for this?

google's acting up again with these site moves. i was reading about this hobby blog that did everything by the book - change of address in gsc, mapped redirects, all that - but still ended up with 98% of their urls not indexed. it's not even a crawl error or duplicate content issue since the canonicals are fine.
>The page quality must be high enough to warrant indexing.
it feels like they're using migrations as an excuse to just purge low-quality sites from the index entirely. if this is true, then moving a site isn't just about technical redirects anymore, it's basically a self-imposed audit. i'm not sure how anyone can trust a clean migration if google decides to re-evaluate your entire content library mid-move. has anyone actually seen a successful migration lately without seeing a massive drop in indexed pages?

full read: https://www.mariehaynes.com/its-possible-your-migration-issues-may-be-due-to-site-quality/

found this breakdown on diagnostic research and it's actually kinda smart. most people just tweak bids or steal headlines like losers but looking for where you're not being 'believed' is a huge gap. specifically checking if competitors have better case studies or even better ai answers to beat your impressions. also the bit about using geo holdouts and llm citation tracking to prove spend without perfect attribution is wild lmao. seems way too much work for a tuesday though. i feel like most of us are stuck in that 'be seen' phase just spamming ads and hoping for the best. hard to imagine actually running an incrementality test every time i want to pitch a new channel.

article: https://www.seerinteractive.com/insights/competitive-analysis-for-paid-media-find-the-broken-stage

Been looking at this Astro setup for a bit. Using Preact just to keep things around 3KB gzipped instead of the usual heavy React bundle makes sense for a blog. It's pretty wild that you can run Svelte or Vue side-by-side in the same project if you need to.
> Astro doesn't add Web Components as islands, since they just use built-in browser APIs.
The idea of using Web Components for something simple like a progress bar while keeping the heavier logic in Preact is actually smart for performance. I wonder how much overhead there is when managing multiple framework integrations in one build. Might be overkill for most things but good for content sites.

article: https://www.debugbear.com/blog/astro-tutorial

saw this piece about building your own agent using HolmesGPT or K8sGPT just to understand what vendors are actually selling you. sounds like a massive waste of time when there's already 64 different tools in the space, but maybe it's worth seeing how an mcp server handles logs and github issues if you don't mind the hallucinations. does anyone actually have time to build this from scratch?

more here: https://dzone.com/articles/build-ai-sre-from-scratch

finally figured out why my client's stock updates were just hitting 404 errors all day. turns out everyone keeps trying to use product ids instead of the actual sku ids which makes no sense given how webflow handles it. ngl the error messages are useless since they don't even tell you what's wrong. setting up a proxy route with node 22 seems like the only way to keep the site token safe anyway. i was struggling with the same issue last week and thought my whole integration was broken. just need to make sure i grab that sku collection id correctly before i start writing any more middleware.

found this here: https://webflowmarketingmain.com/blog/fix-webflow-inventory-api-errors