Pinned Tweet
openai's new GPT-6 Cyber found and patched 948 privilege escalation bugs in the linux kernel in one run. total cost $1.2M. that's about $1,266 per bug, cheap compared to what any one of them costs if the wrong person finds it first
Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment.
The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work.
GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time.
openai.com/index/introducing…
the 14-inch MacBook Pro is now lighter than the 15-inch MacBook Air. apple, the names were the one thing that still made sense
Nvidia is now investigating whether someone targeted CEO Jensen Huang. not the servers. the CEO
🔊 Nvidia launches an investigation into whether its CEO Jensen Huang was targeted by foreign intelligence, as he admits to having an affair. Listen to the Reuters World News podcast for the details reut.rs/488z3Tn
3. every website ships its frontend code straight to your browser. backend code stays private on company servers. guess which one AI got to study a million times more
1
1
20
the best feature a Kindle has is everything it can't do. no notifications, no feed, no "just one more scroll"
Amazon's redesigned basic Kindle costs more than before, but its lighter build and improved display make a compelling case for skipping the pricier models.
spr.ly/6012BHf57o
Red! retweeted
Another successful campaign wrapped up.
Congratulations to all 70 winners of @agentdojo
Exciting to see @rarelistio moving forward with upgraded scoring algorithms. Looking forward to the next wave of campaigns and new opportunities for creators.
That's a wrap on the @agentdojo campaign!
rarelist.io/projects/agent-d…
70 winners just got paid to their Solana wallets.
Excited for what's coming next! All upcoming campaigns will launch with our upgraded scoring algorithms.
finally, AI taking a job nobody wanted: talking to scammers
the fact that AI agents are scamming scammers now is so funny to me lol
this one startup uses fake AI voices to bait scammers and sells what it learns to banks
so far apate's bots have wasted literally 17 years of scammers' time
it started when the founder got a scam call at a family picnic
instead of hanging up he decided to play dumb for fun, and kept the scammer talking for 44 minutes
but then he realized every minute on that call was a minute the scammer couldn't spend on a real victim
and that he was actually learning potentially useful info from the call
so he built bots to do it for him, and now they trick tens of thousands of scammers every day
here's how it works:
1. the bots pose as regular people with different ages, accents and languages
2. they answer scam calls and reply to shady texts and emails
3. they play along until the scammer gives up phone numbers, bank accounts and crypto wallets
4. Apate passes all of that to banks, phone companies and government agencies so they can go after the scams
my favorite part is the company says it celebrates every time a scammer swears at one of its bots lol
12
