HN user

llaxsll

77 karma

https://x.com/ludo_antonov

Posts7
Comments29
View on HN

SuperMe | Software Engineer | SF (In-Person) | Full-time | https://www.superme.ai

We’re building the professional network for the AI age - think Perplexity meets LinkedIn, powered by expert-modeled AI profiles. Team is small (ex-Lyft, Whatnot, Pinterest, Eventbrite, ex-founders) and everyone codes (even our ops person). Our designer is cracked and ships production code regularly.

We're looking for a product-minded engineer who like to work across the stack — spanning product, infrastructure, and ML. If you're excited about the intersection of consumer and AI, we should chat.

Location: In-person in SF.

Check out our the product and tell us a bit more about yourself at: https://superme.ai/ludo , https://superme.ai/casey

Hi mynameishere,

I'm the developer on the Growth Team @ Pinterest that built the experience in question.

First, thanks for the candid feedback. We always strive to understand users and better their experience (in this instance from the very first second they come in touch with Pinterest).

The overhaul of the unauth home page, which is internally dubbed "inspired wall", is a concept that we introduced and refined over a period of a few months. As the name suggests the idea behind it is to instantly connect with you on the level of a story(or a use case) of Pinterest that can inspire you and stick with you as you go through the signup and orientation process. It is a hint of what Pinterest could be for you not written in words but rather expressed through a visual composition. The further you get into the product the more we tailor the experience around things that you love and the story that they tell about you.

We actually did quite extensive testing around this experience because it is such an essential and sensitive part of the product. This included concept research (where we tried to identify what got people to use Pinterest), A/B testing (where we quantitatively showed that results were better than the previous home page), and user testing (where we qualitatively showed that people get better understanding and satisfaction with the product). One of the key findings was that people were able to instantly grasp how Pinterest is different from other popular services such as Instagram or Flickr. To put in music channel terms - it did a good job at explaining how MTV is different than VH1.

Obviously the work is not done and things can be improved. The inspired wall is still a relatively fresh concept and we continue iterating over it.

We plan on releasing an engineering blog post soon that will detail the process of rebuilding our unauth home page and the several iterations that it went through to be successful. Stay tuned and again thanks for the feedback.

Ludo

Hi monkeygus! Thanks for the comment!

Yes, that has floated around as an idea. Totally not out of the realm of possibilities! We think that eating out is such a big problem for people with dietary needs, that it often affects their social life very negatively in many ways. This is the big challenge we are after.

But we are definitely on the lookout of our users telling us what they want to see in Eatfindr and following their needs. We've been live for a couple of weeks now, so our selection database is still growing, but we have over 15,000 gluten-free restaurants from across the US, as well as over 1,000 humane & organic restaurants.

We got rejected :). We didn't build our startup for YC, though. In fact the YC application was just a super great opportunity for us to have a 'checkpoint' at our startup. So, I think this was great. Good luck to everybody else! Check out http://eatfindr.com/ if you'd like to find out more about what we applied with. ShowHN coming soon. :)

I have a low-key server at home. Ubuntu + dyndns. I even host some of my low-key sites on it. I've found it is a pretty good combo. It has come to be a very useful tool over the months.

GWT alternatives?? 14 years ago

I concur. I've lately switched away from GWT/GAE to something more like rails/jquery hybrid.

GWT does have some nice things to it, but I think in the grand scheme of things, the best part of it is the separation between server/client, and the forcing function of having to do rendering on the client, where it should be done in the first place.

With jquery, tho, you can achieve a very similar effect .. just have a div with an id and convert it to widget in one line. There are far more , and much more well written widgets on jquery, than on gwt. It is a little more work, but I think it is well worth it.

Thanks! That's a great feedback. That's one of the things I left out and will address with the next push.

I didn't want to make the event creation very complex, so I was thinking of pushing the location/map specification as a step after the base event is created.

If you actually read further along in the thread, it looks like this guy is trying to implement coroutines. However, he cut some essential part of a sample code that he was initially looking at and then assumed that the problem is with cygwin. And then came some silly flame war.

I personally haven't done anything with coroutines but looks like from a low level perspective, ultimately what he is trying to do is not insane… aka do some form of task switching. Arguably, he shouldn't be doing this by himself and he should've resorted to something that is already available. Just a quick look at (http://en.wikipedia.org/wiki/Coroutine#Coroutine_alternative...) should've been sufficient. Yet.., maybe he had some reason to try and do it himself. Looks like at the end he managed to figure out what he wanted to do.

Dammit, MySQL 16 years ago

It's an ok environment. There is a lot of tooling around it which is to its advantage.

Once you start doing something more advanced, though, you start seeing its flaws. I've had pretty miserable time with its query optimizer and sprocs.

Bing's UI is really hands down much better than Google's new image search UI. In terms of results I don't really find a big difference anymore.

I sometimes resorted to google's image search because it was faster. Dynamic image loading really does take a big toll on performance(at least for me). I wonder why neither Bing nor Google have switched to SilverLight/Flash respectively yet. Maybe they are just waiting for html5?

facebook had thefacebook.com initially and I don't think it was very expensive. I actually found thefacebook.com to be equally satisfying for me.

It was sort of funny to me how this video and Steve Jobs had essentially the same message and points - everybody makes mistakes, if you dont want an iphone 4 bring it back. Hmm. For the record i think it was a pretty good conference on the issue

Interesting. Even if you actually scope it down to the name, some of these principles still hold true. For example, take flickr - its simple, and unexpected as syntax. Kind of when I think of it now, it might be part of the reason why I remembered it in the first place. Althought now with everybody in Web 2.0 using some sort of an interesting syntax.. it is not as unexpected, and thus not as memorable anymore.

For simple apps, yes. But actually I was surprised to find out that they have been on and off plans in attempt to kill linq2sql and actually if you inspect the produced sql code(out of the box) .. well it can use a lot of optimizations.

I do think o/r mappers are the way to go, although having better tools does not mean that you should start handwaving when it comes to db internals.