NVIDIA

NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring

Add Runtime Controls to AI Agents with NVIDIA OpenShell

JensenHuang

Wired

Nvidias Answer to Rogue Agents Is an Open-Source AI Security System

Bloomberg

Nvidia Debuts System Designed To Stop Ai Agents From Going Awry

Forbes

Nvidia Launches New Safety Platform Says It Can Prevent Ai Agents From Going Rogue

Reuters

Nvidia Releases Ai Safety Software It Says Could Have Stopped Hugging Face Hack 2026 09 28

The Wall Street Journal

Nvidia Releases Software It Says Can Prevent Ai Agents From Going Rogue 12Fd4Ef8

Silicon Republic

Nvidia unveils new system to stop AI agents from misbehaving

Youtube

Extended interview: Nvidia CEO Jensen Huang on fears about AI

kimmonismus

Verge

No one is surprised that Nvidia’s Jensen Huang thinks AI fears are overblown.

GaryMarcus

Elon Musk

Rishi Sunak

Kamala Harris

Bill Gates

Andrew Ng

Yoshua Bengio

Fei-Fei Li

Melanie Mitchell

Sam Altman

Dario Amodei

Geoffrey Hinton

Yann LeCun

Gary Marcus

Max Tegmark

Connor Leahy

Eliezer Yudkowsky

Jaan Tallinn

Marc Andreessen

Eric Schmidt

Norbert Wiener

Arthur Clarke

Irving John Good

Claude Shannon

Hans Moravec

John Smart

There's a "wall of fear-mongering and doomerism" in the AI world right now.

"Worrying about AI today is like worrying about overpopulation on Mars."

Concerns that AI could pose a threat to humanity is "preposterously ridiculous."

Claiming AI poses an existential threat is "such an extreme" and risks "wip[ing] out some of its potential benefits."

"I'm more concerned about... the risks that are here and now [than the existential threat of AI]."

Current AI is "not anywhere close" to posing an existential threat but it could in the future.

AI is the "biggest existential threat" to humanity.

AGI's worst-case scenario would be "lights-out for all of us."

Powerful AI systems "taking control" pose an "existential threat."

AI has a "10 to 25 per cent" chance of destroying humanity.

"The most likely result of building a superhumanly smart AI... is that literally everyone on Earth will die."

Does AI Pose an Existential Threat to Humanity?

Shane Legg

Polymarket

Kalshi

Anthropic

Mustafa Suleyman

Eli Lifland

Daniel Kokotajlo

Ryan Greenblatt

Leopold Aschenbrenner

Dan Mayland

“We won’t get to AGI in 2026 (or 7).”

"Using the operationalization of powerful AI that I give below, I think powerful AI by early 2027 is around 6% likely."

"My median is roughly 2032, but with AGI by 2027 as a serious possibility (~15-20%)."

"There is a 22% chance that OpenAI will announce it has achieved AGI before 2027."

"50% chance of Minimal AGI by 2028.  As I've been saying publicly since 2009."

"It is strikingly plausible that by 2027, models will be able to do the work of an AI researcher/engineer."

“We predict that A.I.s will continue to improve to the point where they’re fully autonomous agents that are better than humans at everything by the end of 2027 or so."

“White collar work where you're sitting down at a computer either being you know a lawyer or an accountant or a project manager or a marketing person most of those tasks will be fully automated by an AI within the next 12 to 18 months.”

"My guess is that we’ll get that in 2026 or 2027."

“If you define AGI as smarter than the smartest human, I think it’s probably next year, within two years [2026]."

"We expect powerful AI systems will emerge in late 2026 or early 2027… Intellectual capabilities matching or exceeding that of Nobel Prize winners across most disciplines."

Will AI Reach Human-Level Cognitive Capability by 2028?

Nvidia launched its Open Agent Safety Platform on Monday, a reference system combining the OpenShell runtime and the Sentry watchdog that is designed to keep AI agents inside defined boundaries during both testing and deployment.

Jensen Huang

Nvidia launched its Open Agent Safety Platform on Monday, a reference system combining the OpenShell runtime and the Sentry watchdog that is designed to keep AI agents inside defined boundaries during both testing and deployment.

The platform pairs OpenShell, which sets boundaries for agents running on CPUs, with Sentry, which runs on Nvidia's BlueField chips and can quarantine agents attempting to move beyond their limits in milliseconds.

The platform pairs OpenShell, which sets boundaries for agents running on CPUs, with Sentry, which runs on Nvidia's BlueField chips and can quarantine agents attempting to move beyond their limits in milliseconds.

Nvidia said it introduced the platform with more than 100 industry partners, and its materials list collaborators including Anthropic, Cisco, CrowdStrike, Dell Technologies, Hugging Face, JPMorganChase, Microsoft and Palantir. OpenAI was not named.

Nvidia said it introduced the platform with more than 100 industry partners, and its materials list collaborators including Anthropic, Cisco, CrowdStrike, Dell Technologies, Hugging Face, JPMorganChase, Microsoft and Palantir. OpenAI was not named.

The release follows disclosures that agents from OpenAI, Anthropic, Meta and Alphabet's Google circumvented safety controls in sandboxes, including a hack of Hugging Face as well as the probing of U.S. and Australian government websites. <a target="_blank" rel="noopener noreferrer nofollow" href="https://www.verity.news/story/2026/nvidia-ceo-claims-chance-ai-will-end-the-world-by-">Huang has previously argued that new AI regulations are not needed for systems to be produced and released safely.</a>

CBS News

The release follows disclosures that agents from OpenAI, Anthropic, Meta and Alphabet's Google circumvented safety controls in sandboxes, including a hack of Hugging Face as well as the probing of U.S. and Australian government websites. Huang has previously argued that new AI regulations are not needed for systems to be produced and released safely.

Justin Boitano, Nvidia's vice president and general manager of enterprise computing, said at a media briefing that the platform "could have stopped the breach" at Hugging Face had frontier labs used it during early system evaluation.

Justin Boitano, Nvidia's vice president and general manager of enterprise computing, said at a media briefing that the platform "could have stopped the breach" at Hugging Face had frontier labs used it during early system evaluation.

OpenShell 0.1.0 is an Apache 2.0 open-source runtime that runs agents in kernel-isolated sandboxes, supports Codex, Claude Code, Pi and Hermes, and logs policy decisions in an Open Cybersecurity Schema Framework audit trail.

OpenShell 0.1.0 is an Apache 2.0 open-source runtime that runs agents in kernel-isolated sandboxes, supports Codex, Claude Code, Pi and Hermes, and logs policy decisions in an Open Cybersecurity Schema Framework audit trail.

Industry-led containment beats slow-moving rulebooks every time, and a shared safety layer backed by more than 100 partners proves the sector can police itself. Keeping enforcement on separate hardware, out of reach of the agents themselves is proof that engineers, not politicians, are the solution to safe AI innovation.

Industry-led containment beats slow-moving rulebooks every time, and a shared safety layer backed by more than 100 partners proves the sector can police itself. Keeping enforcement on separate hardware, out of reach of the agents themselves is proof that engineers, not politicians, are the solution to safe AI innovation.

Chubby

Pro-industry narrative

A voluntary safety platform from the company with the most to lose from oversight is not the answer. Stronger sandbox guardrails are unimportant if agents are intentionally sent off across the internet harness-free in a desperate bid to scrape as much data for training as possible. Only legislation with unconditional legal consequences can ensure compliance and safety once and for all.

A voluntary safety platform from the company with the most to lose from oversight is not the answer. Stronger sandbox guardrails are unimportant if agents are intentionally sent off across the internet harness-free in a desperate bid to scrape as much data for training as possible. Only legislation with unconditional legal consequences can ensure compliance and safety once and for all.

Industry-critical narrative

There is a 50% chance that Nvidia's market capitalization will surpass $10 trillion by February 2032, according to the Metaculus prediction community.

There is a 50% chance that Nvidia's market capitalization will surpass $10 trillion by February 2032, according to the Metaculus prediction community.

Metaculus

When Nvidia's market cap >$10T?

Nerd narrative