Ekşi Sözlük (eksisozluk.com) has always been like that, and it still is. People wait over 4–5 years just to become a user, while non-users can only read. It remains one of the biggest websites in Türkiye, yet the design is still very simple, with only one or two new features added over the years. It reached more and more users, but it never really scaled in true meaning. It still like a weekend project
Do things that don't scale, and then don't scale
21–30 of 216 posts
Re: Do things that don't scale, and then don't scale
#22> It would grab the photo, grab the caption, and send it through a direct mail API to my mom. Anyone know what API they are talking about? If I could quickly email a photo somewhere and have it appear as a 4x6 in my mailbox in a few days i would have a much cooler fridge.
There are a few but I use directmailers.com!
Re: Do things that don't scale, and then don't scale
#23> There’s an old startup mantra — probably from the early Airbnb or Y Combinator days — that goes: “Do things that don’t scale.” Yes, it was posted again just the other day: https://news.ycombinator.com/item?id=44913359
Haha funny timing! I wrote this post last weekend and haven’t checked HN much this week
Re: Do things that don't scale, and then don't scale
#24This reads like a LinkedIn post
Re: Do things that don't scale, and then don't scale
#25> It would grab the photo, grab the caption, and send it through a direct mail API to my mom. Anyone know what API they are talking about? If I could quickly email a photo somewhere and have it appear as a 4x6 in my mailbox in a few days i would have a much cooler fridge.
If you search “physical mail API”, I see several options. The first one that I saw charged $1 per letter.
Re: Do things that don't scale, and then don't scale
#26> The cost to build is so low now.
The cost seems deceivingly low right now because those AI companies are fighting for monopoly, but in reality the cost is huge – not only capital, but also trust, privacy, and environmental.
Re: Do things that don't scale, and then don't scale
#27> Could it be bigger? Sure. But at some point — maybe even before 1,000 people — the vibe breaks. The intimacy evaporates. You stop recognizing names. People talk less because it’s harder to know who’s listening. Growth would make it worse, not better. > > Some things work precisely because they’re small. I'd argue this is true for social networks like Facebook actually. There was a magical period in Facebook between…
Yeah, Facebook's best days were when it still required a .edu email to sign up. Makes me wonder if there would still be a market for a smaller, niche social media like that, but on the open web and not locked behind something like Discord servers.
https://www.thegradcafe.com/survey/new/index.php
comes to mind... those were the days, circa 2014 for me, chilling with folk, waiting for thier grad admissions letters
Re: Do things that don't scale, and then don't scale
#28> The cost to build is so low now. The cost seems deceivingly low right now because those AI companies are fighting for monopoly, but in reality the cost is huge – not only capital, but also trust, privacy, and environmental.
If the concern is about the inference cost -- we do have open-weight models that are getting more powerful, and hardware to run small-ish models cheap. I run agents using small local models in my MacBook.
Re: Do things that don't scale, and then don't scale
#29> The cost to build is so low now. The cost seems deceivingly low right now because those AI companies are fighting for monopoly, but in reality the cost is huge – not only capital, but also trust, privacy, and environmental.
If the concern is about the inference cost -- we do have open-weight models that are getting more powerful, and hardware to run small-ish models cheap. I run agents using small local models in my MacBook.
Yes – but when people say "AI", most often they are talking about the latest OpenAI/Claude development.
Re: Do things that don't scale, and then don't scale
#30> The cost to build is so low now. The cost seems deceivingly low right now because those AI companies are fighting for monopoly, but in reality the cost is huge – not only capital, but also trust, privacy, and environmental.
If the concern is about the inference cost -- we do have open-weight models that are getting more powerful, and hardware to run small-ish models cheap. I run agents using small local models in my MacBook.
Can you use something like Cursor with your local models? Is the quality comparable? Is the speed acceptable?