Rahul Jaishankar
Kevin CollierToday, Tuesday, March 28, 2023: Microsoft becomes the first major tech company to pitch reporters that its products are safer because they use generative AI.
This was the tone I was going for in the above post, in case it wasn't clear.
falsafa 
@kevincollier the age of AI
marasawr@kevincollier They have case-studies showing that adding GPTs to these workflows improves outcomes, right? And detailed verification processes for their contributions?
@marasawr they have broll!
Cloudsurfer@kevincollier @marasawr well, based on my experience with the signals that Microsoft sends out about suspicious logins on Azure AD, *any* new technology should be an improvement.
Last Friday, around 4:30 PM, my inbox was hit with a message about a suspicious login that occurred over 3 weeks before. And the message had tenant id and name mixed up. Nice way to postpone starting my weekend.
Last Friday, around 4:30 PM, my inbox was hit with a message about a suspicious login that occurred over 3 weeks before. And the message had tenant id and name mixed up. Nice way to postpone starting my weekend.
Spellbind0127@kevincollier “safer” just like twitter is “safer” because its run by a billionaire.
ZeroTr0ll@Spellbind0127 @kevincollier
This is true. Billionaires will protect me until I become a billionaire myself.
BTW The tone I was going for here was 😜🫏 🐂💩
This is true. Billionaires will protect me until I become a billionaire myself.
BTW The tone I was going for here was 😜🫏 🐂💩
greg ✓✅⍻✔☑√@kevincollier this is the sort of insight that a company just can’t get with a functioning AI ethics team
Andrei Kucharavy@kevincollier in our report about generative AI in cyber-defense, we arrived at the opposite conclusion - Copilot will make Spear-phishing much better and faster; and will allow an almost instant relevant data retrieval and exfiltration.
Sphinx of Black QuartzLLMs are definitely robust and very certainly cannot be gamed by savvy and determined threat actors! I feel safer already!
Tormod HalvorsenHey, great!
Now all I ask to believe it: back up that statement by dropping the "we are not responsible for anything bad that could happen if you use our software, under any circumstances" language that has plagued EULAs and TOSes since forever.
Greg@kevincollier that email/pitch sounds like something a generative AI would write.
Pope Rajulio 
@gregdavis @kevincollier
The Safer label is only applied if it comes from the Safer region of the LLM, otherwise it is just Sparkling Security...
The Safer label is only applied if it comes from the Safer region of the LLM, otherwise it is just Sparkling Security...
concoloris@kevincollier what if security via jumped up eliza program