
Maks
@fadeone_x
Followers
98
Following
176
Media
47
Statuses
493
SWE. Day 49 of 100 - Building: https://t.co/QmwmbXgM90 - Chatbot widget assistant that transforms questions into conversions. Pre-launch stage.
In the middle of Nowhere
Joined June 2025
Simply add this widget to your website:
whilio.com
Transform customer conversations into insights that drive sales with Whilio AI Chatbot widget.
0
0
4
Exactly 6 months ago I've started my current project. When you work solo and going live in a week with something really big. It feels insane. This is my hobby. This is what I'm obsessed with. Building small things and ship fast - not my way.
1
0
7
2020: everyone is obsessed by online privacy and security. 2025: everyone pays for ai browsers sending every user step to the cloud.
0
0
2
A 32 days of challenge behind. Doing at least a small thing every day works like magic. Launch day is coming. Day 66 is soon. Then 34 days of sales and marketing.
0
0
1
The only thing you need to know is how to properly define a task. Black box will do the rest.
0
0
0
๐ this simple example shows much more than you think. Developing things, executing tasks became much easier than ever before.
1
0
0
Imagine you have a complicated task:. Build an API that reviews a given webpage for missing SEO tags. Before LLM: you have to use an HTML / XML parser library, build your own solution that checks every possible meta tag, structured representation like json+ld. .
1
0
0
Fun and unpopular fact!. Attracting 100 followers in X is 10x harder than reaching out 10 customers directly and selling your product.
2
0
3
Day 30 of 100. 6 DMs.4 leads .4 demos.2 trials. Having successful use cases and feedback, it's time to connect 3 more enterprise clients. How did I found them? That's just my 3 local banking services. Wrote 3 emails. Potential revenue in case of 100% conversion ~$2.5k-6k / MRR.
0
0
3
My app has identified issue that costs ~$3-5k / mo to my trial customer. That's not quite huge victory to the customer who earns $100k+ / mo, but it's a serious proof of value to my project. And it's just two days of live tests. Details in (thread)
1
0
2
I hate integrating payment gateways. So much buerocracy, so much pain. .
1
0
3
Cursor, claude code. can cost you 50$/month. and even less. The problem, is that they are too generic and require too much AI analysis on each turn. If we build a coding agent that's:.- adjusted for a single framework .- uses deterministic code chunking .- uses embeddings.
0
0
0
22nd day of being and building in public. Great day today. Going live with two enterprise leads. That's only a test flight of whole system, but is a huge step on the way to production.
1
0
3
It's finally alive. God dmn. I had so much side tasks to do. But there is a lot things left. Three more weeks and we are going launch on ProductHunt.
0
0
1
LLMs have a lot of built-in knowledge, that can imply hidden sense. All you need to know as a prompt engineer: . The smaller model you use, the more details you have to provide.
0
0
0
All we need is to provide proper description, where LLM understand that 5.9 != 5th of September and 5.11 != 5th of November. Not the best prompt, just a showcase below:
0
0
0