montenegrohugo

10 ms·

91.
Vanced: YouTube adblocker for Android(vancedapp.com)
290 points by montenegrohugo 5y ago|346 comments
92.
by montenegrohugo 5y ago
Wait until we make first contact (I'm sure this is scheduled later in 2022)
93.
by montenegrohugo 5y ago
Not crass, I like sharing! We're all here to learn from each other. The monetization model is just referral links to Namecheap, where I get a 10% commission. I want to make that a bit more elegant (especially for people with uBlock Ori
94.
by montenegrohugo 5y ago
You guys are overloading the domain checking API :( Good problem to have I suppose ^^
95.
by montenegrohugo 5y ago
Thanks! Was a fun project to do. I have a bunch of ideas to make it better, but I decided to let it rest for a bit and focus on other stuff. Might put in a bit more effort if it keeps getting the interest it's getting now!
96.
by montenegrohugo 5y ago
Thanks! Yeah, I did NOT expect HN's traffic. It's overloading the domain checking API and making it fail
97.
by montenegrohugo 5y ago
I'm sorry! I didn't expect the HN hug of death. About 200 concurrent people right now, so the domain check API is failing :(
98.
by montenegrohugo 5y ago
I start tons of projects, and it's always a bother naming them. I didn't find existing domain generators at all useful, and since my background is in AI, I made my own. → https://www.namy.ai It currently has a modest b
99.
by montenegrohugo 5y ago
Heyo, cool product. I know setting up and deploying ML on your own is a big PITA from own experience, so ML as a service is a no brainer. Questions: What do you guys see as the long term vision here? Where would you like to go, if you could
100.
by montenegrohugo 5y ago
It's not just in principle, I personally know some startups that do this. Crypto certainly helps them make this easy. It is borderline scammy though.
101.
by montenegrohugo 5y ago
As an intern at my first job a few years ago I basically did something like that. We had a data hunting that would go out into supermarkets and collect data, but oftentimes with errors. Supervisors would sift through all of them and spend 4
102.
by montenegrohugo 5y ago
Reference for what exactly? That input dimensionality is important when determining net size? That seems quite self-explanatory; try training a image classifier with only 100 parameters. Maybe I understood that question wrong, but regardles
103.
by montenegrohugo 5y ago
GPT-3 has been used in a ton of commercially important applications. To name a few: - GitHub CoPilot (transformative imo) - Markcopy.ai, jenni.ai, etc.... Tons of content generation and SEO tools startups - AI Dungeon and such - Plenty of c
104.
by montenegrohugo 5y ago
The point is your simple mapping with zero error on the training dataset also has zero prediction power in both the test dataset and in real life. It's learned nothing; it's at the extreme scale of overfitted. Input dimensionality
105.
by montenegrohugo 5y ago
Super small feedback: there's a typo on the landing page. synchronized instead of syncrhonized
106.
by montenegrohugo 5y ago
The wordplay in that article is just fantastic. German has this specific kind of humor that is hard to describe, but definitely stands out. Great read, thanks for linking it.
107.
by montenegrohugo 5y ago
Thanks sireat! You're totally right on both points. 1) I'll be implementing sometime soon 2) should be fixed now. I had a bit of an error in the backend (growing pains...) and it would show domains available that weren't. Als
108.
by montenegrohugo 5y ago
Such an enlightening comment. This is why I come back to HN, I had no idea why it was so hard to discover classical music (and I've experienced that pain point myself)
109.
by montenegrohugo 5y ago
oh please do! I've tried using GodotAIGym myself, but it was a bit too complicated to make it work, and I was getting non-trivial errors too (in fact last commit on the repo is a small pull request from me). I'd love a way to make
110.
by montenegrohugo 5y ago
For Spain (my country), New Zealand is straight at the other end. Neat.
111.
by montenegrohugo 5y ago
Thanks!! Here's a few answers: - AI Model is a bit primitive still, I've been focusing on building the frontend recently. Now that I'm somewhat happy with that (viewport errors notwithstanding ^^), I'm focusing on the da
112.
by montenegrohugo 5y ago
I'm currently building namy.ai (AI domain name generator) solo, and here's my tech stack: BACKEND: - Python for the backend. Just a great, batteries included language that lets me do what I want fast and doesn't get in my way
113.
by montenegrohugo 5y ago
Stumbled upon this whilst scraping some data for a ML project I'm doing. I tweeted about it and someone noticed that that's exactly the 16th power of two, so some bitshifting shenanigans may be at fault. Thought HN would find this
114.
A website with 65536 empty lines of code(twitter.com)
2 points by montenegrohugo 5y ago|1 comments
115.
by montenegrohugo 5y ago
If you're interested in human-centric urbanism like this, I wholeheartedly recommend the twitter account @wrathofgnon [1]. One of the most knowledgeable and diverse sources of information I've seen on the topic. Really valuable co
116.
by montenegrohugo 5y ago
This is absolutely hilarious and illuminating at the same time. Don't think I'll be able to look at butterfly pictures without remembering this article. It's also a great example of what I'd consider good HN content. Not
117.
What I Wish I Knew When I Was Younger (2015)(youtube.com)
1 points by montenegrohugo 5y ago|1 comments
118.
by montenegrohugo 5y ago
Appreciated!
119.
by montenegrohugo 5y ago
Codepen for this: https://codepen.io/uwwgo/pen/GRmEKJz (nothing revolutionary, does the exact same thing.)
120.
by montenegrohugo 5y ago
I think to define what is "fun" I'd take the reverse approach: What's NOT a fun stack to work with? For me personal hell is a giant stack that I barely understand, a stack that I'm very slow to iterate on, a stack t

More ›