HN

Comments (148)

• Menu

  tananaev
  I have an open source project and started receiving a lot of security vulnerability reports in the last few months. A lot of them are extremely corner cases, but there were some legit ones. They're all fixed now. Closed source software won't receive any reports, but it will be exploited with AI. So I definitely agree with the message of this article.
• Menu

  CodesInChaos
  > The reasoning provided by their CEO, Bailey Pumfleet, is that AI has automated vulnerability discovery at scale,That sounds like an excuse. The real reason is probably that it's hard to make a viable business out of developing open source.
• Menu

  pradn
  Brilliant piece of content marketing:1) Pulls you in with a catchy title, that at first glance seems like a dunk on Cal.com (whatever that is).2) Takes the "we understand your pain" approach to empathize w/ Cal.com, so you feel like you're on the good vibes side.3) Provides a genuine response to the actual problem Cal.com is dealing with. Something you can't dismiss out of hand.4) But in the end of the day, the response aligns perfectly with the product they're promoting (a click away to the homepage!)This mix of genuine ideas and marketing is quite potent. Not saying this is all bad or anything, just found it a bit funny. The mixed-up-ness is the point!
• bobkb
  Security by obscurity is flawed.
• Menu

  janalsncm
  Reading between the lines, it seems like they were working with cal.com and used red team bots to find vulnerabilities in cal.com’s code. And they probably found bugs a lot faster than cal.com could fix them. So the CEO balked at the estimated cost of fixing and took his ball home.This article is effectively an announcement that cal.com is riddled with vulnerabilities, which should be easy to find in an archive of their code.
• JoshTriplett
  I wonder whether cal actually has concerns about security (in which case, they're wrong, this argument was false when people made it decades ago), or whether they just took a convenient excuse to do something they wanted to do anyway because Open Source SaaS businesses are hard.
• keeda
  >Security through obscurity is a losing bet against automationSecurity through obscurity is only problematic if that is the only, or a primary, layer of defense. As an incremental layer of deterrence or delay, it is an absolutely valid tactic, with its primary function being imposing higher costs on the attacker.As such if, as people are postulating post-Mythos, security comes down to which side spends more tokens, it is an even more valid strategy to impose asymmetric costs on the attacker."With enough AI-balls (heheh) all bugs are shallow."From a security perspective, the basic calculus of open versus closed comes down to which you expect to be case for your project: the attention donated by the community outweighs the attention (lowered by openness) invested by attackers, or, the attention from your internal processes outweighs the attention costs (increased by obscurity) on attackers. The only change is that the attention from AI is multifold more effective than from humans, otherwise the calculus is the same.
• erelong
  I'll admit that I agree with a lot of the post but that I can't fully wrap myself around the cybersecurity situation today, is it basically:-if code is open source or closed source, AI bots can still look for exploits-so we need to use AI to develop a checklist program regardless to check for currently known and unknown exploits given our current state of AI tools-we have to just keep running AI tools looking for more security issues as AI models become more powerful, which empowers AI bots attacking but also then AI bots to defensively find exploits and mitigate them-so it's an ongoing effort to work onI understand the logic of closing the source to prevent AI bot scans of the code but also fundamentally people won't trust your closed source code because it could contain harmful code, thus forcing it to be open sourceEdit: Another thing that comes to mind is people are often dunking here on "vibe coding" however can't we just develop "standards / tools" to "harden" vibe coded software and also help guide well for decisions related to architecture of the program, and so on?
• cold_tom
  feels like the real shift is not open vs closed, but reaction time AI attackers don’t need perfect access anymore, just enough surface and time. So the question becomes: can you detect+respond faster than they can iterate in that sense, open source might even help -more eyes reduces time to fix, not just time to find
• linuxhansl
  So Cal.com favors security through obscurity.Open Source was always open to "many eyes" in theory exposing itself to zero-day vulnerabilities. But the "many eyes" go for the good and the bad actors.As far as I am concerned... Way to go Cal.com, and a good reminder to never use your services.
• Menu

  dom96
  Isn’t the real danger now not the ability to find security vulnerabilities, but rather, the ability of anyone to ask an LLM agent to rewrite your open source project in another language and thus work around whatever license your project has?
• cadamsdotcom
  > Security testing has to become an automated, integral part of the CI/CD pipeline. When a developer opens a pull request, an AI agent should immediately attempt to exploit it. When infrastructure changes, an AI should autonomously validate the new attack surface. You do not beat automated attackers by turning off the lights; you beat them by running better automation on the inside.This feels like the core of the article, but it doesn’t prove the need for open source.
• agentifysh
  Pretty overreaching claim about another company's internal decisions and open source in general. There is a lot of incentive to stop open source these days.One of which I am experiencing right now is somebody just copying my repo, not crediting me, didn't even try to change the README. It's pretty discouraging.The other is security reasons, the premise that volunteers will report vulnerabilities really matter if you are big enough for small portion of people to dedicate themselves, for the most part people take open source tool use it and then forget about it, they only want stuff fixed.Lastly, open source development kinda sucks so far. I'v been working on a few different tools and the amount of trolling and just bad faith actors I had to deal with is exhausting. On top of that there is a constant stream of people just demanding stuff to be fixed quickly.
• yc-kraln
  every line of code is a liability. open, closed, doesn't matter. companies will have to treating it that way--which means actual engineering--or they will get burnt, and hard.
• Prunkton
  I'm hopeful the article is right about its prediction, although I'm under the impression the attacker/defender dynamic is asymmetric and the defender on the loosing end. I hope someone can proof me wrong though...Making the assumption that the same amount of money needed to attack a critical vulnerability is also required to find and fix it.Lets say we have a project with 100 modules, and it costs us $100 000 to check these modules for vulnerabilities. What is stopping an attacker from spending the same amount of money to scan, lets say 10 modules but this time with 10x the number of tokens per module than the defender had when hardening the software?
• Divs2890
  Closing your source doesn't close your attack surface,it just closes the community that would have helped you defend it. Security through obscurity is a kind of tradeoff, not a strategy.. i mean that's what I feel.
• pixel_popping
  At the same time, I heavily support open-source and contribute a lot, but I can't necessarily agree that security-through-obfuscation doesn't play a major role in slowing down attacks. Cloudflare have based its whole security being closed-source (for example on its anti-bot mechanism) to be hard to reverse engineer, and they remain leaders as of today with few serious security breaches.Some things just can't be truly secure as well, ddos protection is mostly a guessing/preventive game, exposing your firewall config/scripts will make you more vulnerable than NOT.If your codebase isn't exposed, attackers are constrained by the network and other external restrictions which greatly reduce the number of possible trials, even with a swarm of residential proxies, it's not the same at all from inspecting a codebase in depth with thousand of agents and all models.
• Talderigi
  feels like people are arguing the wrong axis tbh- it’s not open vs closed anymore, it’s more like bug finding going a few devs poking around to basically infinite parallel scanners- so now you don’t get a couple of thoughtful reports, you get a many edge cases and half-real junk. fixing capacity didn’t change though- closing the repo doesn’t really save you, it just switches from white-box to black-box… and that’s getting pretty damn good anywayreal problem is: vuln discovery scaled, patching didn’t. now everything is a backlog game
• 6thbit
  Great PR piece by Strix, but I find mixed messages.Cal.com folks are getting a red team for free, wouldn't that further convince them their closed source software is strong enough?Isn't Strix's business companies paying for scans regardless of whether the software scanned is open source or closed?
• Menu

  shay_ker
  It's a good question - is blackbox hacking as effective as whitebox hacking, for AI agents? I've gotta assume someone at Anthropic is putting together an eval as we speak.
• Menu

  RRRA
  How long before LLM perform perfect disassembly exploitation...
• bzmrgonz
  Strix was so close to being the hero we deserve. I think these blue torches like strix should offer their services for free to open source ships out at sea. There are 3 wins here, GLOBAL GOOD WILL, testimonial and reviews, and market loyalty reward.
• Menu

  simonreiff
  Is there any recent research on whether open or closed-source projects are more secure? I am genuinely curious if anyone has studied the question.
• Menu

  phkahler
  Can any of the AI systems read binary yet? Perhaps generate source code from object file? Is so, that would make access to source redundant for that type of analysis.
• wg0
  > Today, Cal.com announced they are transitioning their core codebase away from open source. The reasoning provided by their CEO, Bailey Pumfleet, is that AI has automated vulnerability discovery at scale, making code scanning and exploitation "near zero-cost". In this new world, they argue, "transparency becomes exposure."Laughable and hilarious. Extremely short sighted. I can show code generated by Claude Opus 4.6 at the highest compute intensity that lacks even basic checks in input validation that was clearly provided in the spec.There's no point in arguing with crypto and AI bros. They are the same tribe. AI crowd however might learn their lessons sooner because the universe isn't forgiving or flexible.Note: I use AI code generators all the time but I take them as very very dumb transpilers no matter how expensive their input/output pricing it and I learned that hard way.PS: Edit to fix typos.
• anon
  undefined
• ChrisArchitect
  Related:Cal.com is going closed sourcehttps://news.ycombinator.com/item?id=47780456
• skal9606
  Seems like flimsy reasoning from the Cal.com CEO. How should we think about Strix vs. foundational model releases like Mythos?
• avivo
  Open source is one of the main reasons I recommended cal.com to everyone — I just did so yesterday again in fact!I'm disappointed to hear this especially since I don't think the rationale makes sense, from what I understand of the security landscape, and it also makes me a little more skeptical of cal.com in general.
• tonymet
  > Closing your source code does not stop an AI from probing your API or finding an authorization bypass in your webhooks.I see this trope a lot in security discussions. “Obscurity isn’t security” or “since you can’t protect against X you may as well do Y”.This is a harmful trope, which discourages perfectly good protections. Sure, closing source is not a perfect protection, but it is a defense against a large band of attacks.Think of the entire field of potential vulnerability probes attackers have. Closing the source closes many of them off, likely most of them.A pen-tester model with implementation will be loads more effective than one with only a black box. And that will give cal.com time to run the pen testing model on the source and address the vulns , hopefully before they are exploited.I tested this myself, first using black box model attacks, secondly using the source code. The model with the source found and exploited the vulns instantly . The model without failed.The lesson is: obscurity is not security ALONE, but it is a component of security.
• themafia
  > In the past, exploiting an application required a highly skilled hacker with years of experience and a significant investment of time to find and exploit vulnerabilities. The reality is that humans don’t have the time, attention, or patience to find everything.I read this as:"We figured no one was looking so we just shipped unsafe garbage for years. We never once did an internal audit, never once paid a hacker to try to exploit our product, never thought we'd get caught with our substandard products."If a guy in his basement with $200 dollars can ruin your company then you were trading on vapor the entire time. I'm sorry you had to find out this way.
• Menu

  funvill
  This is just an excuse to close source their project while blaming AI. Spineless bullshit excuse instead of owning your choices.Shame
• Bridged7756
  It's just an excuse. Classic open source rug pull here.
• Menu

  dzonga
  a lot of the vulnerabilities in web-apps are people trying to be too smart for their own good.use battle-tested frameworks such as Rails, Django then you won't make rookie security mistakes.
• Menu

  reenorap
  All content is going to go behind paywalls.There is zero incentive or reason for content creators to let AI slurp their content for free and distribute it and get all the money from it.Everything new will be licensed and if AI companies want access to it, they will need to pay for it, just like we will.
• Menu

  dangus
  First we blamed AI for layoffs, next we are blaming AI for the AI bait and switch.It's entirely possible this CEO sincerely believes this, but that means you as a potential customer should stay away: now you know that the CEO of this company has no idea how technology works even at an executive level and/or that he doesn't consult his experts before making decisions.
• Menu

  jongjong
  I decided to not open source my latest project but it has nothing to do with security concerns. My code is perfectly secure and bug-free.My concern is mostly financial. Most people would be in a better position to monetize my software than I am... Using AI to obfuscate the origin while appropriating all the key innovations. I wouldn't get any credit.Also, I'm not really interested in humans anymore. I have human fatigue.
• daytonix
  I can't believe we still have people out there buying this baby-brain idea of "If muh code is open than people will find vulns!!" This has been disproven for 20+ years catch up.AI generated bullshit PRs are clearly the bigger issue in the OSS space.
• Menu

  misiti3780
  I have a large open source project and noticed the number of LLM generate PR is making it unmanageable. Every two weeks, I go in, kill all of them and when someone complains or asks why, I realize it was a real person and then I merge it.is anyone else seeing this / fixed this problem ?
• julianozen
  There is another product I use that has a freemium model. They hope to monetize a paid tier for users who use the product a lot.In order to build trust, they open source their product. I forked it, removed the blocks from the freemium feature in 15 minutes using Claude Code. Never published the code to anyone else, just used it myselfUnfortunately, I think it isn’t going to be tenable for systems to be fully open sourced going forward.
• poorcedural
  The idea of tying source code to sustenance will soon be history. We will all remember the days when adding some few thousand smart lines of code meant you could gain notoriety and through cheap viral copy expand those traits to wealth and worth. But software has always just been zeros and ones, the value only happens when interpreted.The future is sharing, you may not believe because your income is tied to being clever. Long term we are all more clever because of the sharing, and your contribution sometimes does not add to your personal success. Asking a company or its individuals to forego their success will not make them add more to our future. But they will add to our future nonetheless, because they all feel like we all do, that adding is what we are all meant to do.
• the_af
  I'm pro FOSS, militantly so. FSF-style.But... playing devil's advocate, if AI makes it very easy to find exploits without the source code, wouldn't it be doubly effective finding them with the source code as well? And why is the dichotomy posed by this blog post "open source with AI reviews by everyone" vs "closed source but only the bad guys use AI"? What if the scenario was: closed source and the authors/security team use every AI tool at their disposal to find bugs? What do the community's eyeballs add to this equation, assuming (big if) AI review of exploits is such a force multiplier?Before any knee-jerk reactions: big fan of open source, I'm not arguing this will kill it, I don't have the faintest idea what Cal.com is and I think a world without FOSS would be a tragedy, I run linux and most of my software on my personal PC (other than games) is FOSS.
• themafia
  > The real solution: fight fire with fireWhich works if you assume that AI can find 100% of your bugs.It can't. So this is a complete waste of your time and will hide actual bugs behind a layer of confidence _and_ obscurity.You're going to actually have to sit down and figure out how to provide real security in your product while earning profits. This is called "work." I understand Silicon Valley would like to earn money and not work. I am eager for these people to get their comeuppance.
• Menu

  shevy-java
  "Open Source Isn't Dead."Well ...Open Source as such will never "die", but we only need to look at what happened in, say, the last 5 or 10 years. Private entities with a commercial interest, have been flexing their muscles. Microsoft - also known as Microslop these days - with Github is probably the most famous example still, but you can see other examples. One that annoys me personally is Shopify's recent influence - rubygems.org is basically just shopifygems.org now. See: https://blog.rubygems.org/2026/04/15/rubygems-org-has-a-publ..."Contributors from both the RubyGems client team and Shopify are already working with us on making native gems a better experience for the Ruby community. "There is a lot more I could add to this (see my complaint about how rubygems.org hijacks gems past the 100.000 download barrier; this was why I retired from using rubygems.org, and then the year afterwards ruby core purged numerous developers. The handwriting is soooooo clear that shopify flexed their muscles here).I think we need to make open source development more accessible to everyone, not just corporations throwing their money to gain influence and leverage. I don't have a great idea to make this model work; economic incentives kind of have to be there too, I get that part, and I am not sure which models could work. But right now we really have a big problem. We can also see this with age sniffing (age verification - see the article that pointed at Meta at orchestrating influence and lobbyism) and many more changes. Something has to change. Hopefully some people cleverer than me can come up with models that are actually sustainable, even if it may not necessarily be a "fund an open source developer for a year". There could be a more wide-spread "achieve xyz" or some other lower finance effort - but again, I don't have a good suggestion here. Hopfully something improves here though, because I am getting really tired of private interests constantly sabotaging and ruining the whole ecosystem while claiming they do "improve" an ecosystem. We have the old "War is peace. Freedom is slavery. Ignorance is strength." going again. Opposite day, every day.
• SingleSourceAI
  [dead]
• Menu

  righthand
  Open source is dead, AI-pundits are applying the wrong lessons. No one has to accept AI or play the game all these AI companies don’t work if everyone stops publishing. Let the AI generated content industry have the publish space, they're very adamant about taking it over and watering it down with slop.I wrote some very nice expressive text for our deployment guide. My project manager took the guide and had Gemini break it down into plain boring bullet points. AI and the pundits can gf themselves in their journey to kill human expression.Here is what I wrote in the guide:"Post Deploy ResponsibilityIf you made it this far, say “Wow I really did it and it was so easy!”Did you say it? Good. Now you are entirely responsible for any issues or bugs that may arise from the newly deployed code. Don’t go anywhere until the deploy has finished (usually takes a few minutes). While an issue or bug may not leave you directly at fault, you are responsible for coordinating any rollbacks or remediations that may be needed until the next deploy."Here is what the product manager slopped it into:"- Post deploy responsibility - You are responsible for performing QA upon deployment - You are responsible for any issues or bugs that may arise from newly deployed code - You are responsible for coordinating any rollbacks or remediations that may be needed until the next deploy" My paragraph wasn't long, hard to understand, or poorly written. I wouldn't have objected to a rewording or some changes but the project manager chose to just copy paste it into Gemini and copy and paste it back. So my take is that they didn't understand what I wrote. Which is a few sentences long and frankly sad if a paragraph is too intense for you to read. When my project manager did this during the meeting I said, "RIP human expression" and their response was a very hasty "no that's not what's happening". This is what all the pundits want to do to everyone and society. Don't believe them that "it's just a tool", that is just a tactic to get you to rollover so they can shove more AI in your face.
• baileypumfleet
  [dead]
• M0HD197
  [dead]
• SyndicateLinks
  [dead]
• realisticid
  [dead]
• tucnak
  [dead]
• theturtletalks
  Enshittification has come for VC backed open-source. As someone on Twittter said, open source has deemed commercial open source obsolete especially when users can point Calude Code to calcom on GitHub and ask it to make them scheduling features directly into their product. That’s what spooked Cal.
• Menu

  Peer_Rich
  cofounder heregoing closed source does not mean we are not fighting fire with firewe are using a handful of internal AI vulnerability scanners for months nowbeing open source simply reduces risk by 5x to 10x according to several security researchers we are working with https://cal.com/blog/continuous-ai-pentesting-vulnerability-...