Jemaclus

Like everyone else, I love Divvy and am sad it's been orphaned. I'll check this out!

My understanding is that the underlying tech is pretty old and well-established. So if you understand those older models, this is old news. A lot of the negativity is basically "yeah, this is old stuff."

But what Jev actually has done that's new and interesting is that it allows you to define a new semantic decision at runtime in natural language, without training a model for that task, and get a fast, cheap, bounded, probability-bearing result designed to be consumed directly by software.

In other words, you don't have to fine tune a model. You should if super high quality is really important to you -- but you should also do that with other things that people reach for LLMs for as well. But for the rest of us for whom good enough is, well, good enough, and who aren't interested in training or fine-tuning models? Jev unlocks a lot off-the-shelf.

You should be impressed. It's doing something more novel than the haters suggest.

Thanks! I have an alpha version of Holler on Android native with two of my friends that have Android phones. We're still in testing phases, but I'm hoping to have it published in the next couple of months. Stay tuned. :)

Totally. But now we're getting into location tracking, and for various reasons (privacy being one of them), I'm mildly averse to tracking location.

And as you approach the "coworkers vs family" problem, you start to drift toward individual invitations, which again adds friction and decision making.

What I think I've settled on is the least friction for the most value. It's not perfect, and I'm not chasing an endless roadmap of features for the sake of features. I've hit upon what really works for me and my social group, so I'm trying to be really, really thoughtful about any additional features.

I do have a draft of personas on my beta app, but I'm not totally happy with it, so I'm going to continue tweaking it and testing it with my social group before releasing it.

I don't have a ton of features planned.

Randomisation is one I've thought of before, but ultimately rejected because it's a thing I've played with in the past and never liked. I dunno if you remember an app called Urbanspoon from back in the day. It was basically a randomized Yelp. More often than not, I didn't want to do the randomised activity, so I'd just spin it again and again and then get frustrated and go to my normal places. But maybe that's just me.

I do have a couple of other ideas, though.

One idea is to re-holler. If I say I'm going to this brewpub and that hollers out to my friends, then one of my friends could say "Yeah, I'm going here, and I'll invite my friends, too." But then you start getting into some weird network effects that might be problematic.

Another is the idea of personas or profiles. There are events where the acceptability of your friend group changes. If I'm going to a farmer's market, it's totally appropriate to have a coworker join you, but at a BBQ at my house, I'd rather not have my coworkers join. Similarly, my family can join a BBQ but I'd rather not meet them at a professional conference.

Similarly, there are some events where I think it's OK to bring a plus-one and in other circumstances absolutely not.

Another concept is that sometimes I might go to, say, ComicCon with some friends in a faraway city. Now my definition of "local" has changed, so if I holler while I'm in San Diego, my friends back home in Boulder won't be able to attend anything -- but my friends that are with me at the convention might totally want to know what's up. So having some sort of mechanism for saying "I'm not at home base, I'm somewhere else and the rules are different" might become useful.

So far I've rejected all of those for one simple reason: they increase friction. I want this to be as simple as it can possibly be to hang out with my friends. The more friction you add, the less likely people are to keep using it. At least, that's my philosophy. So I've rejected just about every other feature I can think of.

As noted elsewhere, the website was made with the assistance of AI. I designed it myself in Canva, used AI to generate the HTML/CSS, but I designed it myself. The app itself is 100% hand-coded by me, though if you consider some of the AI-suggested autocomplete to be AI slop, maybe that's fair.

This app is designed to make the world a marginally better place, so maybe instead of shitting on someone's hard work, providing constructive feedback would be more in the spirit of collaboration and positivity. Have a nice day. :)

Ah, yes, that makes sense! I think that's a great solution.

I think perhaps one sort of semantic/slight difference is that Holler is hyper-local. If you live in LA and your friend lives in London, they obviously aren't going to the local farmer's market with you. I'd say the vast majority of my WhatsApp friends aren't local (I'm based in the States.)

BUT I can totally see that being an easy way to do things if your local friend group is on WhatsApp!

One of my questions would be whether anyone (strangers?) can see that or not. And again this is a US-centric concern, but one of the things I think a lot about is ICE and the government tracking non-white people, and so if there's any semblance of a public "I'll be at X place at Y time," then I'd rather not make that public. So Holler solves that.

But yeah, that's a creative use case for WhatsApp I hadn't considered! Very neat. Thank you for explaining.

Not sure I understand. Can you elaborate?

Ah, Foursquare was so much fun. Definitely inspired this, except this is a little more private -- only your explicitly-added friends get to see your hollers.

I spent a lot of time thinking about what a "public" feed might look like and ultimately decided that my goal wasn't to make a social media app, but to specifically hang out with my friends in person. There's also a locality angle to this, where if you live in LA, you won't add a friend in Montana to your group, because they could never attend your events. So it becomes hyper-local by default.

Group chats, I think, are a fantastic medium for directed messaging (you have to explicitly choose the friends that you're group texting), and if that's something you like, then there's no reason to use Holler. This app, on the other hand, is about undirected messaging, where you simply broadcast your activity. The friction of choosing the group text and possibly excluding someone or accidentally including someone you didn't mean to goes away.

It does introduce other problems, to be fair, but I think it's a trade-off that I'm happy to make.

Thank you for trying! Sadly for ease of use but excellently for privacy reasons, Apple isn't very nice about giving names. It's just something we have to work around.

Yes! You get it! That's awesome. This is basically "comment what you want to do on a FB group and see who shows up" as an app. You got it exactly.

Sort of but not really?

As someone else pointed out, there's pressure in group chats to reply. This explicitly avoids that.

Also, let's say you want to go to a park for a picnic, and you have 100 friends. Which 100 friends are you going to add to a group chat? That's thinking you have to do, friction to getting the engagement going. You have to start thinking about "if I invite Sue, then I can't invite Dan because he's her ex," and blah blah blah.

Group texts are excellent directed communication media. But this is undirected event management. Rather than me inviting 13 specific people, I'm simply broadcasting the destination and time vector.

It's a different way of doing things. I'm not surprised at your question -- it's often the first one I get. But I'm explicitly attempting to solve a different problem than group texts solve.

Ha, yeah. It's a fox with a speech bubble, because foxes live in hollers. Plus my daughter loves foxes :)