Oops! AI Startup Exposes Secrets Like a Bad Reality Show

by

So, picture this: DeepSeek, the latest darling of the AI world, has just pulled a classic “oopsie-daisy” by leaving its database wide open for all the world to see. Yes, you heard it right! It’s like leaving your diary on the bus with all your deepest secrets—except this time, it’s a million log entries of chat history, secret keys, and API secrets. Talk about a major security faux pas! 🙈

Enter Gal Nagli, our knight in shining armor from Wiz, who stumbled upon this digital treasure trove. Apparently, anyone with a computer could waltz right in and take a peek. I mean, who needs a password when you can just have a free-for-all? 🎉

After Wiz gave DeepSeek a friendly nudge (or maybe a full-on shove), they finally decided to fix the issue. But let’s be real—how did they not notice this before? It’s like leaving your front door wide open and wondering why the neighbors are suddenly very interested in your living room decor. 🏡

Now, the database was hosted at oauth2callback.deepseek[.]com:9000 and dev.deepseek[.]com:9000, which sounds like a secret lair for supervillains, but alas, it was just a playground for hackers. And while the vulnerability allowed for total control without authentication, we’re still left wondering if any sneaky hackers had a field day before the doors were finally locked. 🕵️‍♂️

Nagli, the security guru, has a point: AI companies really need to get their act together when it comes to security. I mean, come on, it’s not rocket science! Or maybe it is, and that’s why they’re all so distracted? 🤔

DeepSeek, which has been making waves with its open-source AI models, has had its fair share of drama. After becoming the hottest app on the block, it faced “large-scale attacks” and had to hit the brakes on new sign-ups. So much for a smooth launch! 🚀

Read More

2025-01-30 17:25