This may become the single biggest test of insider security threats that will be seen to date.
I applaud the way Anthropic is attempting to handle this - we will see if all the internal teams and selected participants have the same guiding principles.
This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨
Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser...
The numbers are hard to believe:
> $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built
> Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet
> Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit
For context: these are the kinds of findings that previously required elite security researchers working for weeks.
Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning.
The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing:
A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.
@digitalrealty is reimagining the AI deployment process, empowering enterprise customers to reduce time to value and maximize their ROI on AI investments.
Learn more about the Digital Realty Innovation Lab, and hear how Cisco is helping: https://t.co/0241tuH3Xi
Distributed private #AI is at the forefront of a new era in enterprise intelligence. From minimizing exposure to public networks to optimizing performance at the edge.
Read the blog. https://t.co/Vle01z97yb
#TheDataMeetingPlace#WhereTomorrowComesTogether#PlatformDIGITAL
The future of AI isn't about renting tokens—it's about owning and optimizing #AI-as-an-asset.
Distributed private AI is enabling model fine-tuning on owned data sets and use cases for enhanced privacy, latency, and scale. Learn more https://t.co/o57YqsR3Tk
#PlatformDIGITAL
I’m speaking at #CiscoLive Amsterdam next week, if you’re attending let's connect.
We’re excited to share how Digital Realty and Cisco are helping customers build secure, scalable and geo-distributed #AI infrastructure across hybrid multi-cloud environments worldwide.
James Madison had more points in the first 10 minutes against Oregon on the road in Autzen Stadium than Texas Tech did in the entire game at a neutral site.
To be fair, James Madison has scored more points in their CFP game than:
2014 Florida State
2015 Michigan State
2015 Oklahoma
2016 Washington
2016 Ohio State
2017 Clemson
2018 Notre Dame
2019 Ohio State
2020 Notre Dame
2021 Cincinnati
2021 Michigan
2023 Alabama
2024 Indiana
2024 SMU
2024 Clemson
2024 Tennessee
2024 Boise State
2024 Georgia
This paper from Stanford and Harvard explains why most “agentic AI” systems feel impressive in demos and then completely fall apart in real use.
The core argument is simple and uncomfortable: agents don’t fail because they lack intelligence. They fail because they don’t adapt.
The research shows that most agents are built to execute plans, not revise them. They assume the world stays stable. Tools work as expected. Goals remain valid. Once any of that changes, the agent keeps going anyway, confidently making the wrong move over and over.
The authors draw a clear line between execution and adaptation.
Execution is following a plan.
Adaptation is noticing the plan is wrong and changing behavior mid-flight.
Most agents today only do the first.
A few key insights stood out.
Adaptation is not fine-tuning. These agents are not retrained. They adapt by monitoring outcomes, recognizing failure patterns, and updating strategies while the task is still running.
Rigid tool use is a hidden failure mode. Agents that treat tools as fixed options get stuck. Agents that can re-rank, abandon, or switch tools based on feedback perform far better.
Memory beats raw reasoning. Agents that store short, structured lessons from past successes and failures outperform agents that rely on longer chains of reasoning. Remembering what worked matters more than thinking harder.
The takeaway is blunt.
Scaling agentic AI is not about larger models or more complex prompts. It’s about systems that can detect when reality diverges from their assumptions and respond intelligently instead of pushing forward blindly.
Most “autonomous agents” today don’t adapt.
They execute.
And execution without adaptation is just automation with better marketing.
I speak for the WHOLE WORLD when I say we are ALL James Madison fans this weekend.
Can’t imagine rooting for a team just because they are in the same conference. 🙄
If you think Group of 5 teams don’t belong in the College Football Playoff, you don’t actually love college football.
Group of 5 teams belong in the CFP if they earn it.
JMU and Tulane earned their way in.
Anyone trying to gate keep the CFP and keep Group of 5 teams out is carrying water for the wrong people.
Football fans love underdogs. Sports fans love Cinderellas. If they get crushed in the game, that’s okay.
We’ve seen perennial powerhouses get crushed in the College Football playoffs and those same voices didn’t say those teams didn’t belong in consideration for the CFP ever again.
-Alabama lost 44-16 to Clemson in 2019
-Oklahoma lost 63-28 to LSU in 2019
-Ohio State lost 31-0 to Clemson in 2016
-Ohio State lost 52-24 to Alabama in 2021
All are in the CFP this year because they earned it.
We’ve seen Cinderellas beat the big dawgs throughout College Football and College Sports history.
So whether JMU and Tulane get blown out, hold their own, or win their games in the CFP, the result shouldn’t be used against all group of 5 teams to keep them out of the CFP.
An undefeated or 1 loss Group of 5 conference champion earns their way into the CFP.
The size of your brand shouldn’t determine your spot in the CFP.
Winning games matters.
If you’re gonna carry water for anyone, carry it for the PEOPLE OF COLLEGE FOOTBALL. They don’t all live in 2 conferences.
Tonight's live show we will discuss the new phenomenon sweeping the Mainstream Sports Media.
GDS - G5 Derangement Syndrome
Jam packed show, don't miss it!!
Streaming on all platforms at 8pm est.
Paul Finebaum and his media lackeys ⬇️
The countdown to unveil our AMS11: Digital Next has begun!
Paula Cogan, and Chris Sharp, will join our valuable customers, partners, and colleagues to show off our latest AI-ready data center in the middle of Europe on November 20, 2025: https://t.co/Gj9bnpnrOO