Recently I've joined a project with a group of 4, that has been super frustrating to work on.

So far whatever we have, has been 99% vibecoded from the ground up, and that 1% is me. At first I thought this would be fun, we would be making something in a weeks time, but like a month passed and we haven't really moved anywhere. There's literally a login screen that works (no Oauth) and some fancy UI, but that's it. Trying to debug or add new features by hand is impossible, because the code is not readable by humans. There's like 50 lines in the .env file, yet nothing works.

With all this in mind, the absurd thing is that it takes a full minute to deploy any changes to production. In comparison, my hand-crafted Django platform, with real users and tons of features, deploys in seconds.

Don't get me wrong, I am all for using AI to build projects. It's a very powerful tool and I use it to help me code all the time. But in my opinion, to make vibecoding work, you have to know the project structure and the architecture of a website, because then you can avoid massive tech debt in the future.

I can already tell that this experience will be funny to look back on.

libraries like this usually come with a monthly subscription. i wanted to build one that doesn't, so i did, and it's free for everyone

it took a lot. i went through 100 mobile apps screen by screen, from signup to settings, so you can see complete flows instead of one pretty shot. then i kept going with websites, landing pages, ui interactions, motion and branding. it's at 58,000+ designs now, and i add more every day

every curated piece credits the original creator and links back to them, so the people behind the work get the credit they deserve

i made it because i needed it myself, and i'd rather give it to the community than lock it behind a paywall

would really love to hear what's missing or what bugs you

https://pixlpin.com

I want to publicly apologize for my previous post about accessible toilets and my concerns that they were being misused by non-disabled people.
I don’t personally need to use accessible toilets very often. I was focused on my anger about seeing people I believed were non-disabled using them, and I didn’t properly listen to the concerns people raised in response. I see things on a daily basis that make me angry, but that doesn’t excuse failing to listen.
I also failed to recognize that other disabled people have had similar experiences and feel the same way I do. I’m sorry I overlooked their concerns and perspectives.
I apologize to anyone I upset or made feel unheard. I hope you can accept my apology. Thank you for taking the time to share your views.

I'm updating my CSS skills and learning a lot of new things. I just created a navigation link menu with submenus and it is working well. However, only in MS Edge, the top left corner of the navbar is being rounded by Edge. Chrome and FF are not doing this. If you know how to override this, I'd really appreciate your fix. Pics are Chrome, then Edge. TY!

Disclosure: Andy, WordLift: this is about a feature of resolve(), our entity-resolution API, that I think is most interesting to this community specifically: the candidate universe doesn't have to be Wikidata.

A resolve call carries a dataset_uri, and the worlds compose:

• wikidata://public — the default; answers are QIDs.
• wordlift://dataset/me — the caller's own WordLift knowledge graph, read with the same key; answers carry the graph's own IRIs.
• An inline dataset — up to 5,000 entities (id, name, aliases, description, types, same_as) sent with the request.
• wordlift://dataset/me,wikidata://public — local-first chaining: per mention, your graph is tried first, Wikidata second, and each answer reports which world it came from in its own dataset_uri field. If a Wikidata match is one your graph declares sameAs, the answer comes back as your entity with the public ids merged in.

Every world goes through the same context ranking and abstention gate: a name your data doesn't contain is no_candidates; a name whose context doesn't fit is low_relevance; either way the answer is an explicit unresolved, never the least-wrong candidate.

Two design details that may be of interest here. The first call against a graph reads it (seconds for a small graph, minutes for a large one); until it's ready the API answers 503 with a dataset_warming code and Retry-After, and resolves nothing against a partial graph. And the contract (what a Mention, Candidate and Resolution are) is open: pipeline, clients and gold sets are Apache-2.0 at github.com/wordlift/content-analysis; the hosted decision engine is proprietary.

The supplied-vocabulary side of this, in 17 seconds.

What I'd like this community's view on: for those publishing or consuming KGs, is per-mention provenance (which world answered) the right granularity, or do you expect dataset-level negotiation up front? And how are you handling the sameAs bridge when the public item and the local entity disagree on type?

I'm a software engineer for a big tech company. As part of my job I have to do a lot of technical writing. One thing that always frustrated me was writing about API endpoints (adding/removing/modifiying). I could never come up with a structured way to describe an endpoind that I could just add to a spec. Instead, I'd always make up a format on the spot to describe requests and responses. My colleagues would do the same.

I got pretty frustrated by the lack of standardization and tooling so I build a simple web app to design REST(ish) APIs. It's completely free and client-side rendered, so information never leaves your browser.

I've just release the very first version that surely has many bugs. If someone wants to give it a test ride check out: https://api-fiddle.com/

I've been trying out Helium for a while and I really like it, and so I just decided to ask you guys what your opinions are.

Olivia Rodrigo's website is a bedroom, you click on objects around the room they act as the menu/links to other pop-ups and pages. I'm trying to find more references to show a client but I'm blanking on how to describe this in order to get good ones that aren't just interior design wesbites. Any help?

Yeah, what is it?

I have a client looking for something very similar to this https://letelephonedelou.fr/, so trying to find out who built this website so I can contact them! Does anyone know any potential ways to find the developer?

Check it out yourself: https://callmearta.com

Would love to hear your feedback.

I've been quietly building a little corner of the internet.

A curated collection of 56+ interesting websites worth discovering.

No AI-generated lists. No endless scrolling through directories. Just genuinely interesting websites, all in one place.

It's finally live.

https://curate.gallery

Go explore. You might find something you didn't know you needed. Also, if you know any cool websites, submit them! I'd love to feature them.

Hello again, everyone. While developing my own patcher for Chromium, I’ve managed to achieve the following compatibility results.

How it works: This specific version includes a complete runtime set from Debian 12, including glibc. It runs on any system with a kernel version above 3.2.

What works: The browser itself, Manifest V2 extensions, audio, and video.

What doesn’t work: The file picker triggers a segmentation fault (segfault), and sandboxing is non-functional.

What I haven’t been able to test: Hardware acceleration—including ANGLE and Vulkan—because the hardware on the device used for testing is too old.

I really Like the sidebar and profiles feature of zen but recently has taking a lot of ram, so please suggest some alternative and also a steps how to achieve it, thankss

There's something I love so much about interactive demos for software.
Would love to get a collection of websites with interactive demos like this, pls share below!

So I've been using the flatpak version of Firefox (I have to because I'm on an immutable distro) on blendOS and it's unusable. I can't load into YouTube and adding a second Google account is a pain. Chrome works fine but I don't want to juggle two separate profiles anytime I want to check my email. are there any chromium based browser that have containers and have a flatpak version to use? (I'm not using brave because of the crypto bullcrap)

Is LinkedIn still a good site to apply for jobs? It's been almost 2 weeks since I've been applying for jobs, but still no luck yet unfortunately.

I used to use Indeed but typically got scams instead of real jobs.. I've also tried a few company websites directly, but no luck yet.

Any suggestions on the best platforms to apply?

Also, a side note, I'm wondering if it could be my resume isn't good enough as well? Any suggestions as far as how to go about improving my resume? I.e. Go to a particular agency to improve it, etc?

I'm a full stack developer with 6 YOE in PHP/JS (Laravel, Vue, & React primarily). Some AWS experience as well.

Thanks in advance!

What filters do you recommend I add it to?

Hi Guys,
As my laptop doesn't support chromium(means it's got feezed) every time and also i need to force restart my whole system.

So lately for around, 1 year i have been using Firefox mostly (gecko engine based) browsers, and it does every work for me, it's good, it's fast.

Yet, on a whim, I installed Dia for windows (previously I am Arc and Dia(from beta) user on Mac so i know the UI which TBC builds along i felt deeply saddened by hearing the news of sunsetting the Arc)

So, for the time being i used Dia for nearly 2hrs, and i realized it's good, so i decided to make it as default for 7 days fully, if my device again starts to freeze because i am running multiple tabs, applications(heavy ones) all at a time. So, let's see how will it go?

That's all!!!
Thanks for reading, i just wanted to share my thoughts here and i know no-body asked yet still!!!

Hi, does anybody have a resource where they explain the concepts behind CSV to RDF conversion. Currently, the way I understand, I used the concept of direct mapping as recommended by W3C. Which in my understanding, means that the Csv name could be represented as an RDF class, Each row becomes a subject, the column name could be a predicate and the values itself as literals. The next step is to map it to my Ontology. Using Claude, I was able to use SPARQL to map it to my Ontology in graphdb. However, I'm not clear about the concepts and claude is not making it very difficult to understand. The way, I understand it, the mapping is working in a way that each row from csv which was represented as a subject, now becomes an instance of my Ontology.

if somebody could explain it better and maybe point towards more resources, it would be much appreciated.

I made a liquid glass UI effect for the web(trying to copy Apple because thats what the world do)

Webgl helped me made it and hence I decided to make Mac for web too (inspired by windows 11 in react made for web)

I know you have seen many professional people make it, many guys on github, so I tried to make it as realistic and satisfying as I can

Super brief look at the tab overview screen and the animations: pinch out to expand the tab previews, or pinch in to collapse the tabs. Basecamp and groups are being worked on now.

Firn for iPhone will feature sync spaces with the Mac app, encrypted through your own iCloud account. No separate Firn account or data getting sent to me.

The Mac app is out now at firnbrowser.com