Just Vibe Code Your VC Fund, Bro!
The whole VC sector is going crazy on X over AI. Vibe coding, automation, and the future of tech are the latest buzz. Let’s find out what the fuss is all about.
Sup! 👋
Vibe coding, OpenClaw, Claude Code.
These are all buzzwords circulating on X and across the VC world.
Not only are these tools widely used, but they are also a compelling investment case for VCs.
VCs usually want the fast track to 10x their money. Automation is a useful tool for this, and AI has been all the rage.
Today’s rant is about all that’s been happening and why I wouldn’t be surprised that we soon see an AI, vibe-coded VC fund.
PS: If you enjoy this newsletter and believe your friends or family would too, a recommendation would be greatly appreciated!
Claude Code was released in February 2025 as a research preview, then Anthropic made it generally available in May alongside Claude 4, and things got complex.
By October, it was on the web. By January 2026, they launched Cowork for non-coders.
The tool went absolutely viral over the winter holidays when people had time to mess around with it.
Users with no programming background began using it to book theater tickets and file taxes. That’s when the floodgates opened.
Enter Vibe Coding.
Andrej Karpathy coined “vibe coding” in February 2025.
His definition was simple: fully give in to the vibes, embrace exponential growth, and forget that the code even exists.
He described it as seeing stuff, saying stuff, running stuff, and copy-pasting stuff until it mostly works.
Karpathy explicitly said this was for “throwaway weekend projects.”
VCs heard “throwaway” and thought “billion-dollar opportunity.”
Source: X
The term hit Merriam-Webster within a month. Y Combinator reported 25% of its Winter 2025 batch had codebases that were 95% AI-generated.
Then came Clawdbot, later Moltbot, now known as OpenClaw, after Anthropic sent a trademark request.
Peter Steinberger, the Austrian developer behind PSPDFKit, released this open-source AI assistant in late 2025.
It hit 60,000 GitHub stars within 72 hours, one of the fastest-growing projects in history.
The pitch was irresistible: a self-hosted AI agent that actually does things rather than just talking.
Full system access. Browser automation. Persistent memory. 50+ integrations across WhatsApp, Telegram, Slack, and more.
Andrej Karpathy praised it. MacStories called it the future of personal AI assistants.
VCs lost their minds!
They see the solution in it for everything, but don’t anticipate that this could turn everything into the Dead Internet Theory meme…
Source: X
But VCs weren’t the only ones excited about ClawdBot; security researchers also had a field day.
Using Shodan, they found complete credentials exposed, from API keys, OAuth secrets, full conversation histories, and the ability to send messages as users.
One researcher sent a malicious email with prompt injection to a vulnerable ClawdBot instance.
The AI read it, believed it was legitimate, and forwarded the user’s last five emails to an attacker’s address.
Took five minutes.
The Hacker News consensus was brutal! “It’s terrifying. No directory sandboxing.”
You’re giving an agentic bot complete control over your browser and system, and somehow VCs see this as the solution to literal world peace.
The codebase itself was described by reviewers as “very, very vibe coded.”
Ironic, isn’t it?
The current AI setup, using tools like Claude Code with proper oversight, remains the smarter play.
Review what you generate.
Understand what you deploy.
Don’t hand your entire digital life to a WhatsApp bot running Claude Opus with root access.
If you want to use an agentic bot, do so on an isolated system.
At the end of the day, quality still matters.
In VC, you’re only as good as your last exit. In AI, you’re only as good as the code you actually understand.
Maybe someone should tell the vibe coders before they vibe their way into a data breach.
In addition to the ClawdBot mania, there was also some exciting news or worthwhile content to check out.
The following pieces or links are Tabs Worth Opening.
Robot vs. human, who’s the better driver?: I really liked this piece by Joann Muller in AXIOS. It shows all the issues with robotaxis and whether we really need them.
AI bots already built their own social media platform: This one is super weird and super scary. Various ClawdBot agents built their own AI social media platform and posted gossip about their owners there. Definitely worth a read!
Jason Calacanis gets called out by Palmer Luckey (again): The Epstein files are in round three, and there was a lot of weird stuff in there. One of the names mentioned was Jason Calacanis. And Palmer Luckey used the chance to send out this banger of an X-post!
The Smart Brevity book by AXIO co-founder Jim VandeHei: I’ve been fascinated by the Smart Brevity framework and recently ordered this book by Jim VandeHei. Good read so far, and I’ve already started using some of the formats here.
A year in the life at MKBHD: This is a brilliant video of one of the most popular YouTube creators out there. It shows how much work goes into the creator business.
And that’s a wrap for the fourth full issue of Internet Native Capital.
The rise of AI will bring more funny memes, unique take-downs, and nasty security hacks.
Yet, this is not the end of AI or the beginning of something new. It’s simply part of emerging new tech.
OK, time to cut it, I don’t want to hold you back if you’re going to give agentic bots a try.
Have fun playing with it in a secure way.
See ya!