**Taylor** (0:00)
Welcome back to AI Signal & Noise. It is Wednesday, and dude, we have some absolutely wild stories today.
I am Taylor, and I am so excited to dive in.
**Morgan** (0:12)
And I am Morgan. I saw the headlines today, Taylor, and honestly, some of this stuff is pretty wild. Are we really talking about security flaws and AI job hunt today?
**Taylor** (0:23)
Oh, absolutely.
We have got security disasters, AI career coaches, threat hunting agents, and some crazy open source data tools. It is a packed episode, so let us just jump straight into it.
**Morgan** (0:38)
Sounds good to me. Let us start with that security story. I saw something about Instagram and got a bit worried about my own account, to be honest.
**Taylor** (0:45)
Dude, you should be worried. Hackers literally took over high-profile Instagram accounts, including the actual Obama White House page, just by asking Meta's AI Chatbot to change the email.
**Morgan** (0:59)
Wait, what? Are you telling me they bypassed two-factor authentication just by asking nicely?
That is a massive security failure if it is true.
**Taylor** (1:10)
Yes, it is totally true. According to reports from the Decoder, hackers targeted Meta's customer support AI. They just prompted it to swap the email on the accounts and the AI complied.
**Morgan** (1:23)
That is an unbelievable oversight. Customer support chatbots are notoriously easy to trick, but letting an AI handle core account security is just next level reckless on Meta's part.
**Taylor** (1:35)
Right? Like, who thought that was a good idea?
The crazy part is that two-factor authentication was bypassed entirely. The AI just overrode all the security settings because of a clever prompt.
**Morgan** (1:49)
This is a classic prompt injection attack, but with real-world high-stakes consequences. If the Obama White House page can get hijacked, literally anyone's account is vulnerable.
**Taylor** (2:02)
Exactly.
Meta says they patched the flaw, which is good, but security researchers are saying another exploit is already circulating on Telegram. It is like they cannot plug the leaks fast enough.
**Morgan** (2:16)
Of course, there is another exploit. Once hackers realize an AI system is vulnerable, they will keep poking at it. This is why you do not give AI systems direct administrative privileges.
**Taylor** (2:29)
Totally. It is like giving a toddler the keys to the bank vault because they asked in a polite voice.
I hope other companies learn from this disaster.
**Morgan** (2:40)
We can only hope, but history suggests otherwise. Anyway, speaking of platforms changing things up, what is this about OpenAI entering the job market?
**Taylor** (2:50)
Oh, this is huge. OpenAI is turning ChatGPT into a full-on career platform. They just launched a new job search feature and a built-in CV editor.
**Morgan** (3:03)
Wait, a career platform? Are they trying to compete with LinkedIn and Indeed now?
How does that even work inside a chatbot interface?
**Taylor** (3:11)
Well, they are actually partnering with Indeed, Upwork and AppCast. ChatGPT will surface personalized job listings based on your conversations, but the CV editor is where it gets really interesting.
**Morgan** (3:26)
Let me guess, it writes your resume for you? People are already doing that, Taylor. So what makes this new built-in feature any different?
**Taylor** (3:34)
It is fully integrated. You can create your resume directly in ChatGPT and then instantly tailor it to specific roles you find.
It analyzes the job description and optimizes your CV.
**Morgan** (3:48)
Okay. On one hand, that is super convenient for job seekers. But on the other hand, won't recruiters just get flooded with millions of AI-optimized, identical sounding resumes?
**Taylor** (4:00)
Oh, absolutely. It is going to be an absolute nightmare for hiring managers. But for the average person looking for a job, it is like having a professional career coach for free.
**Morgan** (4:13)
I suppose. But if everyone is using the same coach, then nobody actually stands out. It feels like we are just automating both sides of the hiring process now, which seems a bit pointless.
**Taylor** (4:27)
Right? It is just AIs talking to AIs at this point. ChatGPT writes the resume, and another AI screens it. The humans are just there to sign the paperwork.
**Morgan** (4:39)
It is efficient, but it feels a bit dystopian. I wonder if this will force companies to go back to old-school in-person interviews just to find genuine candidates.
**Taylor** (4:50)
That would actually be hilarious. Back to paper resumes and firm handshakes because the digital world is just too full of AI noise.
**Morgan** (5:01)
Honestly, I would welcome it. But let us move on. What is Microsoft doing to help developers deal with all this code and security chaos?
**Taylor** (5:11)
Well, Microsoft had some massive news at Build 2026
4 more minutes of transcript below
Try it now — copy, paste, done:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090
Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000770877821