A two-decade retrospective of launches Hacker News dismissed. And what happened next.
“For a Linux user, you can already build such a system yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem. From Windows or Mac, this FTP account could be accessed through built-in software.”
“It does not seem very ‘viral’ or income-generating.”
“It’s pretty nice, and I was thinking to myself — hey cool, I could make an online backup of my code. Then it occurred to me — who the hell is this guy, and why should I trust my code to be on his server!?”
“It’s a pretty crowded space. And XDrive gets you 5 GB for free, 50 GB for $9.95 a month. I think competitors can duplicate Dropbox’s nice front end.”
The most famous bad take in HN history — the ‘I could build this myself with existing unix tools’ archetype.
What happened
Dropbox IPO’d in 2018 at a $12B valuation. Drew Houston later thanked BrandonM by name when Dropbox went public.
“I still don’t exactly understand what they are offering? Is there an advantage to using GitHub versus dumping some (yet to be created) virtual machine image on a cheap virtual server?”
“Don’t you think that git’s advantage over SVN evaporates when there is only one user on a team? I run my private Subversion repository which I use for everything (not just code).”
“Doesn’t the pricing seem a bit too granular, though? I suspect the pricing categories will collapse into 3, maybe 4, levels eventually.”
The opening comment literally couldn’t see the point. GitHub was perceived as ‘just a git host’ — the social layer, the network effects, the open source ecosystem it would enable were all invisible.
What happened
Microsoft acquired GitHub in 2018 for $7.5 billion. GitHub now hosts 100M+ developers and 420M+ repositories. It became the connective tissue of the entire software industry.
“Well this is an exceptionally cute idea, but there is absolutely no way that anyone is going to have any faith in this currency.”
“I’m having trouble wrapping my head around the logistics of this...”
The entire thread had just 3 comments and 5 upvotes. Three comments. For what would become a $2 trillion asset class.
What happened
A single bitcoin went from fractions of a cent in 2009 to over $100,000 by 2024. Total crypto market cap exceeded $3 trillion.
“I can’t ever see anyone saying ‘just duckduckgo it.’ The name just sounds silly. It makes me think it’s a search engine for toddlers.”
“DuckDuckGo is childish. I think that name will hold them back.”
“How many people would go to Google and search for ‘new search engine’? DuckDuckGo is not even in the top 10 pages.”
“I don’t find their actual search engine very useful at all.”
The name — the name! — was the biggest objection. Nobody could get past it. Meanwhile, Google itself was once mocked for being a misspelling of a number.
What happened
DuckDuckGo grew to over 100 million daily search queries and became the default search engine in many privacy-focused browsers. Valued at over $600 million.
“Unfortunately taxis are a regulated industry in most major cities. The entrenched interests of the taxi companies are simply too big — and they have the political clout — to let this one slide under the radar.”
“If this service became at all popular, it is very likely that cities would immediately include ‘mobile hailing’ as also requiring a license.”
“Driving a gypsy cab (which is what UberCab is) is a dangerous business. A bad guy could simply place an order for an out-of-the-way alley or warehouse and know that the cabbie is going to be driving a really nice car.”
“The first time an UberCab driver gets into a wreck without insurance or licensing should be interesting.”
“This drastically idealizes UberCab profiles. It gets a lot shadier when UberCab is one of 10 companies doing this, and when it starts to become worth it to game profiles.”
Two months after this thread, Uber received an actual cease-and-desist from San Francisco — seemingly validating every skeptic. Travis Kalanick’s response was to ignore it and expand to five more cities.
What happened
Uber IPO’d in 2019 and is now worth over $160 billion. NYC taxi medallions, which sold for $1.3M in 2014, collapsed to under $80K. The regulation that was supposed to stop Uber became its origin story.
“All my experiences with it as a user have been too unreliable to expect that it can scale to truly massive usability. I just don’t see it swallowing up the whole hotel industry.”
“This exchange cements my concerns about Airbnb only being huge if they can end-run the hotel regulatory system.”
“Airbnb is almost more like a dating service than a marketplace… a buyer and seller who prove compatible will never need to use the service again.”
“Airbnb is great unless you’re the kind of person that doesn’t trust strangers. Sadly, in the United States, the tendency to not trust strangers has been on the upswing for the last few decades.”
The top comment sided with the skeptics. Commenters argued Airbnb couldn’t scale, couldn’t survive regulation, and couldn’t solve the fundamental trust problem of sleeping in a stranger’s home.
What happened
Airbnb IPO’d in 2020 at a $100B+ valuation and is now worth over $80 billion. One of YC’s most successful companies ever.
“I really don’t get or see how Stripe is different? Why would I use it instead of PayPal, 2CheckOut, e-junkie, etc?”
“I have no need of a fancy API either — PayPal lets me specify the basics and fire off a simple Post from my PHP code.”
“Stripe gets added to the bookmark collection for ‘services to use should I ever have a problem with PayPal.’”
“Pretty much every company in payment processing that does not use segregated merchant accounts sooner or later goes bust.”
The launch thread was full of commenters doing unfavorable price comparisons to PayPal. Posted by Patrick Collison himself.
What happened
Stripe reached a $106B+ valuation and processed $1.4 trillion in payments in 2024. The ‘fancy API’ became the default payments infrastructure for the internet.
“A lot of really smart people have tried and failed to accomplish this sort of thing before. Amazon invested $60 million in Kozmo.com back in the late 90’s, and they couldn’t make it work.”
“I just do not see how this scales, as your marginal labor costs have got to be a very high portion of your revenues.”
“Having a delivery fee is a non-starter. ‘I can get it in 2 days free with Amazon, or $4 today...’ People will spend huge amounts of time and effort to not pay delivery charges.”
“I’ve built a few real-time delivery businesses, and I’m pessimistic. Real-time operations are costly to manage. Not being Amazon and not being able to control inventory hurts.”
The top comment pointed to the graveyard of companies that tried before. The entire thread read like a post-mortem for a company that hadn’t even launched yet.
What happened
Instacart IPO’d in 2023 and is worth over $12 billion. COVID-19 turned grocery delivery from a novelty into a necessity.
“It includes code to load up various analytics tools even if you never use them. For example, if I only use GA and Mixpanel, do I really want to serve the bytes for all the other plugins?”
“It’s going to be really hard to make a generic, non-lossy mapping in a static, stateless JS script.”
“I was hoping that this would be an open source version of Google Analytics.”
“Google Analytics has a new API currently in beta that is also called analytics.js. This will be confusing.”
Commenters argued the abstraction layer couldn’t work across fundamentally different analytics providers. The founders later wrote ‘From Show HN to Series D.’
What happened
Segment was acquired by Twilio for $3.2 billion in 2020 — one of the largest acquisitions in developer tooling history.
“I think you’d be a damned fool to invest in this technology for any serious project. Right now this is a toy.”
“I have more than a sneaking suspicion that this project is essentially a proof-of-concept, and that it is not heavily used at Microsoft.”
“Where’s all that great refactoring support if everything is made dynamic and stringly typed?”
Microsoft + new language + compile-to-JS triggered every distrust reflex at once. The phrase ‘damned fool’ was deployed with full sincerity.
What happened
TypeScript is now used by 80%+ of JavaScript developers and is the default language for virtually every major web framework. It’s arguably the most impactful programming language of the 2010s.
“This is terrible. Did we really not learn anything from PHP days? Are we seriously going to mix markup and logic again?”
“OMG, JSX… why? Just why?? Stop ruining JS people!”
“The current fad of quasi-declarative web components looks like early Ext to me, and I think everyone knows how that turned out.”
“Mixing JS and XML syntax, variables in docblocks, DOM components that are not really DOM components… Yikes. Thank you, but no, thank you.”
“Meh.”
The developer community overwhelmingly felt React violated fundamental software engineering principles. ‘Separation of concerns’ was the rallying cry against it.
What happened
React became the most popular UI library in the world, used by over 20 million developers. In 2025, Meta donated React to the Linux Foundation.
“I mean this in the most helpful way possible: the interface is really, really bad at serving one of its basic, fundamental functions.”
“I can get everything I need on HN. Ultimately the best products will make the front page here, no need to look around.”
“I looked at the page for like 30 seconds, thinking to myself, ‘What is this?’… literally incomprehensible to a typical reader.”
“Here’s a few things I couldn’t easily figure out on your site: What is a best new product? How is this different than a linked list on a blog? Is this site for me or for someone else?”
Commenters couldn’t figure out what Product Hunt was after 30 seconds. It went on to become the definitive launchpad for tech products.
What happened
Product Hunt became the go-to platform for launching tech products, was accepted into Y Combinator, raised from Andreessen Horowitz, and was acquired by AngelList.
“No way this a spreadsheet. This is just a CRUD app with data displayed in rows. Zero chance of catching with spreadsheet users.”
“The demand for an Access-like or ‘better spreadsheet’ product is all of the ‘Oh yeah, it sounds cool’ variety that never results in sales.”
“Very difficult to get non technical peeps just suddenly ditch spreadsheets.”
“Your app seems sluggish to scroll compared to Google Docs at that size, and the record density seems low.”
Commenters predicted zero market demand. The ‘better spreadsheet’ category was seen as a graveyard of failed attempts.
What happened
Airtable reached an $11 billion valuation and is used by over 300,000 organizations including Netflix, Shopify, and Nike. It created an entirely new product category.
“This service allows me to solve this communication problem by asking designers to learn this tool — which is new to them, requires time, and also isn’t as powerful as Photoshop.”
“The main differentiator is ‘we’re making this run in the browser.’ But nowhere does it explain why that’s better for designers.”
“I just want a solid desktop app that isn’t a web wrapper or lives in the browser. I can’t stand web apps to be honest.”
“$18MM to spend only to see if you got it wrong is a rather interesting approach.”
“MEH”
An entire comment. Just “MEH.” For a company that would be valued at $20 billion.
What happened
Adobe tried to acquire Figma for $20 billion in 2022. Figma IPO’d in 2025 at a $60B+ market cap.
“I don’t understand Tailwind. The entire point of CSS is to separate style from structure. How does applying composable utility classes differ from the old days of using HTML attributes for styling?”
“This is essentially the same as inlining all of your styles in a style attribute on every element. I don’t see how you would ever reasonably want to use this in a project.”
“Wasn’t the whole point of CSS to separate presentation from data, and move away from things like <font color=...>? This is still considered bad practice, right?”
“I don’t get it either. Start putting CSS in the style attribute while you’re at it.”
“The emperor has no clothes.”
The exact same ‘separation of concerns’ argument was levelled against React in 2013. HN missed it twice.
What happened
Tailwind CSS became the most-downloaded CSS framework in the world with 100M+ npm downloads per month. It’s now the default styling choice for millions of developers, used by Shopify, OpenAI, Netflix, and GitHub itself.
“No one should use a for-profit terminal emulator, especially one created by a VC-backed startup, full stop.”
“Downloaded the image, installed it and was greeted by a mandatory login. Next step was uninstall and delete the dmg image. What a waste of time.”
“You like people to contribute for free but refuse to give them an actual FOSS client. This is bound to fail.”
“Warp’s VC decide they want an exit and Warp becomes 50usd/month SaaS. Your workflow, scripts, etc. are basically dead.”
A VC-backed terminal? Requiring login? With telemetry? HN reached for the pitchforks. The thread read like a restraining order.
What happened
Warp raised $78 million, grew to hundreds of thousands of users, and became the most popular modern terminal. Login removed, telemetry made optional — every criticism addressed.
“I love these super-ambitious projects (see Parcel, Rome.js) because after several years they will still fail in many areas at once!”
“Moving to a reimplementation of core Node APIs is a terrifying prospect.”
“Something has done a bit wrong if you’re running any of those tools in production.”
One commenter preemptively grouped Bun with Parcel and Rome.js — ambitious projects that burned out. 1,431 upvotes said otherwise.
What happened
Bun 1.0 shipped 14 months later. It’s now a legitimate Node.js alternative with 80,000+ GitHub stars, backed by $7M+ in funding, and used in production worldwide.
“I asked the chatbot broad questions and got answers that were straight-up false. These chatbots instill a delusion of consciousness in people. Every new technology has initially had cases where people could be deluded into thinking it was magic.”
“The complete lack of humour or wit is what breaks the illusion for me. Its responses are extremely human-like, but mind-numbingly dull.”
“What if the real result of ‘AI safety’ is making models like this boring as hell?”
“It’s interesting how ChatGPT feels much more filtered, reserved, and somehow judgmental than GPT-3.”
Posted the same day ChatGPT launched. Commenters called it dull, filtered, and delusional — the same day the fastest-growing consumer app in history was born.
What happened
ChatGPT became the fastest-growing consumer app in history — 100 million users in 2 months. OpenAI reached a $157B valuation.
“There’s nothing on the official website or GitHub that indicates what this software is, other than a cropped screenshot that looks like VSCode with a prompt pop up over it.”
“So looking through the dependencies, it’s CodeMirror with a VSCode theme on top of it, that includes Copilot. Why wouldn’t I just use an existing editor with Copilot support?”
“AI is still in an info-phase Bitcoin was in before 2017. Expected to see an avalanche of fake/fraud/phony products based on it.”
The first Show HN got just 14 upvotes. Fourteen. The thread had 5 comments. One commenter couldn’t tell if it was ‘some sarcastic joke software.’
What happened
By 2025, 1 billion lines of code were being written on Cursor every day. Valued at $10 billion.
“While and after watching the video, I wasn’t sure if the whole thing isn’t just a parody of AI companies.”
“It’s a cool idea but I really don’t see how this is any different from Cursor IDE.”
“Of course it’s not going to be sustainable.”
“Totally useless and I’m sure I will not be subscribing to it at any cost. It gets easily confused and cannot troubleshoot or understand a bit of the environment.”
One commenter thought the entire launch video was a parody. Another gave it exactly three minutes before declaring judgment.
What happened
Windsurf was acquired in a deal worth $2.4 billion, with its CEO and key employees joining Google.
“It’s clear that progress is incremental at this point. Anthropic and OpenAI are bleeding money. It’s unclear to me how they’ll shift to making money while providing almost no enhanced value.”
“I paid for it for a while, but I kept running out of usage limits right in the middle of work every day. I don’t recommend using it in a professional setting.”
“The biggest complaint I have is that we continuously hit the limit via the UI after even just a few intensive queries.”
“The API could cost $100/hr. All of these coding tools seem to work only with the API and are therefore either too expensive or too limited.”
Critics focused on rate limits and cost. The thread got 2,127 points and 963 comments — a sign of just how much people actually cared.
What happened
Claude Code hit $1B in annualized revenue within 6 months of GA — faster than ChatGPT. By early 2026 it surpassed $2.5B ARR.
“This thing chews through tokens. I’ve spent $300+ in the last 2 days doing fairly basic tasks. Also, it’s terrifying — no directory sandboxing. It can modify anything on my machine.”
“There are 300 open GitHub issues. One of them is a security report claiming hundreds of high-risk issues, including hard-coded, unencrypted OAuth credentials. I am disinclined to install this software.”
“I just don’t trust an AI enough to run unprompted with root access to a machine 24/7. Most of the cool stuff here you can also just vibecode in an afternoon using regular Claude Code.”
“Layers and layers of security practices over the past decade are just going out the window. It’s quite wild to give root access to a process that has access to the internet without any guardrails.”
“This is all starting to feel like the productivity theater rabbit hole people went down with Notion/Obsidian. It is clearly capable of doing a lot of stuff, but where is the real impact?”
The project hit 60,000 GitHub stars overnight. Critics called it hype — then Anthropic asked for a name change, and OpenAI acquired the creator.
What happened
Creator Peter Steinberger joined OpenAI to work on AI agents. The project surpassed 145,000+ GitHub stars and spawned an entire ecosystem of derivative projects.