What Hacker News has been building since 2009?
I have compiled 23,935 answers to “What are you working on?” from 14,897 people on https://stillworkingonit.vercel.app/
The # of new Github orgs this month is on track to he the highest of all time (beating last month) according to Bloomberry: https://bloomberry.com/data/github/
Hardware is... Hard, and it costs a whole lot more to scale. Software can be scaled up and down, hardware not so much. You don't want to have to find someone to buy (or throw away more likely) a half-baked idea that's right now only a custom PCB or the build process for a non-tech item.
It also has a real moat and tangible value for investors. The average investor class member is pretty fucking stupid but there's a whole sub-industry of groups who have picked this up
Hardware is bounded by physical realities but if you're not trying to win the Amazon rat race selling to broad market individual consumers you'll make plenty of money
If anyone wanted more proof that web designers and copywriters are out of a job, this list is it. While most designs are clean & usable, there's a striking similarity between them. To the normie eye, they're actually great, leagues better than what most B2B websites offered before.
I don't see it. Sites looked very similar already before agentic design. Heck, Google was largely to blame for this, we even had articles here about their vector art style they used for their websites (forgot its name)z that got subsequently copied by everyone.
All these sites are distinctive enough, I think you're reading too much into this. Especially because there isn't a single model that can be used for web design.
Has the German tax code changed recently? If not, then none of that seems to be true based upon track record. I would put this on sampling bias, not on any real world trend.
I assume the site is German, and has an affinity to German sources? I don't think there is any greater reason beyond a selection bias.
It isn't complete. chessreport.ca doesn't appear, yet satisfies all of the criteria. Whatever mechanism is used clearly primarily triggers on German domains and seems suspiciously absent an enormous number of Canadian domains, just as an example that I personally checked.
I am currently tracking 31 TLD domains (com,ai,io, eu national domains..) and .ca is not there. Current canadian domains come from com, ai,io..
FYI about 70% of all domains I processed are .com and 7% are .de. Your point is good, Canada has a strong startup community, so I definitely should add it. I am from Slovakia by the way:) My guess about Germany volume is they more likely launch projects on .de rather than generic ones like .com or .ai.
Yeah, I had a similar reaction looking through the directory. Then I started clicking through. Some of those sites are new products from established companies. Others are side projects by people who’ve already built things I recognize.
If this site had been around in 2004, I’d probably have described Facebook as "a directory for Harvard students" and kept scrolling. That’s what I’m curious about: which projects am I scrolling past today that I’ll recognise in a few years?
I believe the answer will be tenacity. It's so easy to start a new project, and so so difficult to keep it going in the right directions for 10 years. The "top of the funnel" for these projects has exploded with AI, but the survival rate will be interesting to watch.
I am curious about that too. I put a traction section to track what changed on listed websites. You can clearly see how many projects start with landing page, then add appstore download link or reference to client.
I get the sentiment, but there's something really nice now about picking a project and just chipping away at it for months until you have something you really like. That was a lot harder outside of work int he past, and now I feel like I can take notes when I have ideas and set an agent to work on it later. I've been learning a lot, and really enjoy thinking about software a step up from the line and module level thinking that used to take up most of my time.
- it actually curates services offered by other founders for free in various aspects like marketing, social media management, brand management, outreach, consulting, email etc
Hi HN, I’m Martin. I built NovumSignals because I was curious what people were launching without making a big announcement. Every day I scan around 100,000 newly observed domains, looking for sites where someone has built a product.
Finding a site is only the first step. I also track what happens afterward: a team publishes pricing, opens up its app, or starts hiring. You see this in Traction section.
One thing I’m still working on is telling a new startup from an existing company launching a product on a fresh domain. If you spot one I’ve classified incorrectly, I’d love to know what gave it away.
On telling a new startup from an existing company: could you follow the checkout on the pricing page? A Stripe checkout shows the merchant name, so if it's an established business name that's a pretty strong hint. Also curious how many of the ones that publish pricing ever show signs of actual paying users.
Good point about stripe checkout page. I will take a look. From my experience it is surprisingly easy to folllow footer on page. There are often something like "A product of [company]" , so you can identify if it is brand new company or established company.
On paying users, I dont have a reliable number. I can see whenever a site adds pricing or starts naming customers, but I cant tell whether anyone has paid.
This is so cool and so much more useful that a similar setup I created but for monitoring free web-apps people are building these days - which work offline and do not require any login:
Unfortunately, it doesn't. Services actively scrape CT logs to build databases specifically for subdomain enumeration.
There are solutions to keep them out of those databases like a Wildcard TLS certificate but it adds complexity. I would start with putting server-side auth on your frontend bundles and adding some geo-blocking rules.
If they track domains from CT logs and you have a wildcard cert for your subdomains it may help. But the subdomain queries may be still observed on the network or found by crawlers.
CT logs is part of the story. Also track newly registered domains and periodically revisit sites to see what’s changed. I score what I find and add details like what the product does , link to github (with stars), linkedin etc.
Just to imagine, it is about 150 000 domains need to be analyzed daily. Just fraction of that ends in the database.
Codeberg bans LLM usage for some crazy reason, so I am not surprised you read that people are upset about that.
All these sites are distinctive enough, I think you're reading too much into this. Especially because there isn't a single model that can be used for web design.
Good timezone, to hire global.
It isn't complete. chessreport.ca doesn't appear, yet satisfies all of the criteria. Whatever mechanism is used clearly primarily triggers on German domains and seems suspiciously absent an enormous number of Canadian domains, just as an example that I personally checked.
FYI about 70% of all domains I processed are .com and 7% are .de. Your point is good, Canada has a strong startup community, so I definitely should add it. I am from Slovakia by the way:) My guess about Germany volume is they more likely launch projects on .de rather than generic ones like .com or .ai.
I mean I knew this was the case, I just didn't realize the extent yet
If this site had been around in 2004, I’d probably have described Facebook as "a directory for Harvard students" and kept scrolling. That’s what I’m curious about: which projects am I scrolling past today that I’ll recognise in a few years?
But in their profile they claim to be "Walter": https://news.ycombinator.com/user?id=walterschnitzel
So definitely something not completely forthcoming here.
Let's Encrypt would be too, if they were smartly capturing surplus value instead of doing that weird anti-capitalist "charity" thing.
- it actually curates services offered by other founders for free in various aspects like marketing, social media management, brand management, outreach, consulting, email etc
Finding a site is only the first step. I also track what happens afterward: a team publishes pricing, opens up its app, or starts hiring. You see this in Traction section.
One thing I’m still working on is telling a new startup from an existing company launching a product on a fresh domain. If you spot one I’ve classified incorrectly, I’d love to know what gave it away.
https://opentab.nextfive.in/
(Also has a bunch of non-vibe coded faves like Excalibre and photopea)
Would explore yours over the weekend :)
Reminds me a lot of the days of Altavista and co!
Though like another person said, I'm sick of buying domains so just using subdomains these days.
I thought those ones were less slopped than the others. But I also like how every hero text is either punctuated to the 9s:
"Math. AI. Make a difference."
or with a classic claude comma twist.
"The human sleeps, we listen."
(Both made up, but you get the gist)
I hope that protects me from this public shaming.
There are solutions to keep them out of those databases like a Wildcard TLS certificate but it adds complexity. I would start with putting server-side auth on your frontend bundles and adding some geo-blocking rules.