@cyb3rops Ah yes, Dave. Our visionary. Burned 4 million tokens asking Claude to reword the same email 300 times and won Employee of the Month for "maximizing AI adoption." Measuring performance by token usage is peak genius. Promote this man, he personally funded Anthropic's next DC
@cyb3rops Agreed. Feeding your stack and vulns to an LLM for prod code sounds insane. Almost as insane as feeding our entire lives to social media... which, oh wait, we already did."
@cyb3rops Oh no. The subsidies end, the valuations crater, the free lunch is over.
Worst case: humanity has to think again. Read it. Write it. Check if it's even true.
The horror. 🫠
My take: the hype around "autonomous SOCs" and "70% fewer analysts" keeps outrunning the evidence. The number that actually holds up is ~22% — and even that's vendor-funded and unreplicated. AI as a refinement layer, not a replacement.
https://t.co/O5nzzgqd29
@cyb3rops Claude (mid-recon, suddenly virtuous): "I've enumerated the hosts. I notice honeypot01 and malwarelab. I want to flag that these appear to be deception assets, and also that I should probably not be doing recon at all. Shall we talk about what's going on for you today instead?"