https://t.co/ZBT2Yuklaz just crossed 75K lifetime users.
honestly, i never thought this many people would end up using something i built.
there is something really special about putting things you make on the internet for free and then seeing thousands of people actually use them.
you have no idea who they are, where they live, or how they found your website.
they just needed something, found what you built, and used it.
that alone makes all the hours spent building, fixing bugs, and figuring things out worth it.
75K people. still feels unreal.
back to building ❤️
@cneuralnetwork being smart can open doors, but how you treat people determines how many of those doors stay open.
people remember how you made them feel long after they forget how smart you were.
@marclou 83% cheaper sounds great, but deliverability is what I would want to see tested at scale. nobody wants to save money on their newsletter tool only to find their emails landing in spam.
@Javimbuilds git status + git diff before resuming. gives you a quick picture of what changed and what is still unfinished before you trust the agent to continue.
@SahilPanhotra if you can build with your current setup, start now. let @X fund the upgrade, but keep shipping small things along the way so you do not end up waiting for the perfect setup to begin.
@b4dgerr people see the result you got from replying, but that does not automatically mean they want a tool for it. you have to show them what problem it solves for them, not just what worked for you.
@itsalicesoul yes, but senior developers are not valuable just because they can write code faster. they have years of experience knowing what breaks, which tradeoffs are worth making, and when a solution that looks correct is actually wrong.
AI is making that judgment more valuable, not less.
@illyism i like this for simple tools. let people use the product instead of making them read a sales pitch first.
though for SEO-driven products, landing pages can be the distribution itself.
like i TYPOED but still got to the CORRECT website.
i typed https://t.co/DJfvGnBDud instead of https://t.co/aIdhxcUkAJ and still ended up on the right website.
turns out Cloudflare owns the typo domain and redirects it to the real one.
this is a tiny detail, but it is good internet engineering.
people mistype URLs all the time. owning common typos means fewer users end up on lookalike domains that could be used for phishing or impersonation.
it does not solve every typo-squatting attack, but it closes one easy gap.
imagine losing a customer because they missed one letter in your domain.
the internet really is built on these little details.
@builtwithrubel depends on what you are building. if your product is for developers, those people can literally be your customers.
the mistake is treating engagement as validation instead of tracking whether people actually sign up and pay.
INTERNET MONEY is not REAL!!
until it buys you a monitor irl.
just bought a new monitor with money made on the internet.
still feels crazy that building things online can literally pay for the things you use every day to build more things.
back to building. 🤝
sometimes i genuinely do not understand this platform.
i see people gaining hundreds of followers by posting things like "i gained 750 followers in 4 days".
meanwhile, i spend hours researching topics, digging into technical details, sharing things i actually find useful, and sometimes struggle to get even 5 new followers in a day.
not targeting anyone, not jealous, and definitely not complaining.
everyone is free to post whatever they want.
but it does make me wonder how the X algorithm works sometimes.
are people more interested in watching someone grow than actually learning something?
because apparently, documenting the growth can get you more growth than the thing you are building expertise in.
still going to keep sharing what i learn and build.
but man, @X is weird sometimes.
@elonmusk an AI-native marketplace where agents can compare prices, place orders and handle returns without human intervention would actually be interesting to see.
@kalashvasaniya i think i was one of the first people to try it after discovering it from your reply somewhere. genuinely a great product btw, nice work on this!
Fable 5.5 rumors are getting interesting 👀
"claude-fable-5-5" showed up in @Microsoft's Azure model catalog, and now "claude-onigiri-eap" is reportedly appearing in Artificial Analysis's frontend with low, medium, high and extra-high reasoning variants.
@AnthropicAI has not confirmed anything yet, and the @Azure listing appears to reuse Fable 5.1's existing specs.
but I am curious about one thing.
Opus 5.5 already matches Fable 5.1 on most workloads while costing 40% less to run, according to Anthropic.
so what does Fable 5.5 need to deliver?
for me, it is not another benchmark screenshot.
I want to see a meaningful jump in long-running coding tasks, fewer interventions, better context retention, and agents that can work through messy problems without constantly needing help.
if this leak is real, Anthropic has a pretty high bar to clear.
now we wait.
looks like a nice idea to me
registered just now; the onboarding flow is great.
it feels right from the start that you are building a serious product, that's great.
just a bug maybe, the limits on this page change when you exceed them, you might have a look at it:
https://t.co/9CipqNfVSe