Nik Kale
@nik-kale.bsky.social
Building AI systems that don’t break
Principal Engineer @ Cisco
Agentic automation · AI security · In-product AI systems
Patents · Industry awards · Judging
Principal Engineer @ Cisco
Agentic automation · AI security · In-product AI systems
Patents · Industry awards · Judging
TechCrunch has the full breakdown of Discord's rollout and how the verification works: techcrunch.com/2026/02/09/...
Discord to roll out age verification next month for full access to its platform | TechCrunch
All users will be put into a "teen-appropriate experience" by default unless they prove that they are adults.
techcrunch.com
February 11, 2026 at 5:00 AM
TechCrunch has the full breakdown of Discord's rollout and how the verification works: techcrunch.com/2026/02/09/...
And the real risk isn't the check itself. It's what persists afterward. A breached password can be reset. A breached faceprint is compromised forever. You can't rotate your face.
The critical design question isn't whether to verify. It's where the verification happens and what survives it.
The critical design question isn't whether to verify. It's where the verification happens and what survives it.
February 11, 2026 at 5:00 AM
And the real risk isn't the check itself. It's what persists afterward. A breached password can be reset. A breached faceprint is compromised forever. You can't rotate your face.
The critical design question isn't whether to verify. It's where the verification happens and what survives it.
The critical design question isn't whether to verify. It's where the verification happens and what survives it.
Biometric age verification gives platforms plausible deniability more than actual certainty. It lets them say "we checked" without guaranteeing "we know."
February 11, 2026 at 5:00 AM
Biometric age verification gives platforms plausible deniability more than actual certainty. It lets them say "we checked" without guaranteeing "we know."
Read more in my follow-up with Evan Schuman and @CSOonline: www.csoonline.com/article/412...
Think agentic AI is hard to secure today? Just wait a few months
Exponential expansion of autonomous agents in the enterprise may expand enterprise threat surfaces to an almost unmanageable degree — especially given poor foundations for non-human identity oversight.
www.csoonline.com
February 10, 2026 at 8:00 PM
Read more in my follow-up with Evan Schuman and @CSOonline: www.csoonline.com/article/412...
The plan: containment plus a clean slate going forward. Don't try to fix legacy. Assume it's compromised and treat it as hostile territory.
February 10, 2026 at 8:00 PM
The plan: containment plus a clean slate going forward. Don't try to fix legacy. Assume it's compromised and treat it as hostile territory.
Enterprise teams will need both patterns. The question nobody's asking: who governs the agent when it can finish the ticket without you?
February 10, 2026 at 1:00 AM
Enterprise teams will need both patterns. The question nobody's asking: who governs the agent when it can finish the ticket without you?
That's not a competition. That's two different theories of what "AI engineer" means.
One reads the whole repo before touching anything. The other starts running commands and fixes things in motion.
One reads the whole repo before touching anything. The other starts running commands and fixes things in motion.
February 10, 2026 at 1:00 AM
That's not a competition. That's two different theories of what "AI engineer" means.
One reads the whole repo before touching anything. The other starts running commands and fixes things in motion.
One reads the whole repo before touching anything. The other starts running commands and fixes things in motion.
I talked about this with Evan Schuman and @CSOonline: www.csoonline.com/article/410...
Agentic AI already hinting at cybersecurity’s pending identity crisis
An estimated 95% of enterprises have not deployed identity protections for their autonomous agents — not a great start for what some security experts describe as an authentication concern without precedent.
www.csoonline.com
February 9, 2026 at 8:00 PM
I talked about this with Evan Schuman and @CSOonline: www.csoonline.com/article/410...
More in @CSOonline: www.csoonline.com/article/412...
Think agentic AI is hard to secure today? Just wait a few months
Exponential expansion of autonomous agents in the enterprise may expand enterprise threat surfaces to an almost unmanageable degree — especially given poor foundations for non-human identity oversight.
www.csoonline.com
February 7, 2026 at 1:00 AM
More in @CSOonline: www.csoonline.com/article/412...
Most orgs are already undercounting by 2-3x because machine identities are scattered across cloud consoles, repos, config files, and secrets managers nobody's aggregating.
February 7, 2026 at 1:00 AM
Most orgs are already undercounting by 2-3x because machine identities are scattered across cloud consoles, repos, config files, and secrets managers nobody's aggregating.
The engineers who develop this instinct early will be the ones trusted with autonomous systems. The ones who don't will be supervised by the ones who do.
February 6, 2026 at 8:00 PM
The engineers who develop this instinct early will be the ones trusted with autonomous systems. The ones who don't will be supervised by the ones who do.
Knowing when to trust an output. When to verify. When to override. Understanding what audit trail you're creating. Recognizing when a model is operating outside its training distribution.
This isn't a course you take. It's judgment you build from operating real systems and watching them fail.
This isn't a course you take. It's judgment you build from operating real systems and watching them fail.
February 6, 2026 at 8:00 PM
Knowing when to trust an output. When to verify. When to override. Understanding what audit trail you're creating. Recognizing when a model is operating outside its training distribution.
This isn't a course you take. It's judgment you build from operating real systems and watching them fail.
This isn't a course you take. It's judgment you build from operating real systems and watching them fail.
Wrote about this and more: cognitiveworld.com/articles/20...
From Copilots to Embedded Agents: Operationalizing Agentic AI for Service Innovation — COGNITIVE WORLD
Over the past two years, enterprises have rapidly adopted AI copilots, chatbots, and assistants across support, IT, and customer-facing workflows. Early results have been promising. Teams report faster answers, reduced manual effort, and incremental productivity gains.
cognitiveworld.com
February 5, 2026 at 8:00 PM
Wrote about this and more: cognitiveworld.com/articles/20...
More on my quote in Reworked: www.reworked.co/digital-wor...
Why More Workplace Technology Creates More Friction
Employees lose nearly a full workday every week battling fragmented software. Here’s how complexity is crushing productivity, morale and revenue.
www.reworked.co
February 5, 2026 at 1:00 AM
More on my quote in Reworked: www.reworked.co/digital-wor...
Learn More: blogs.cisco.com/innovation/...
What Reviewing 500+ AI System Evaluations Reveals About Enterprise Readiness
Discover what 500+ enterprise AI system evaluations reveal about AI readiness, durability, and trust. Learn the patterns that distinguish successful enterprise AI deployments and drive industry standards.
blogs.cisco.com
February 4, 2026 at 1:00 AM
Learn More: blogs.cisco.com/innovation/...