Techstrong Gang

Can open-weight AI deliver more control without adding new operational risks? What makes an IT team ready for agents, and why is Anthropic setting rules for how people treat Claude?

The October 9 discussion connects open-model access, the foundations needed for agentic AI and the boundaries of Anthropic's Claude conduct policy. The takeaway: separate model access from operational readiness, and platform conduct rules from evidence of machine consciousness.

Host: Mike Vizard. Guests: Jack Poller, Jon Swartz and Jeff Reich.

Watch the October 9 episode on YouTube or LinkedIn.

How can leaders judge AI claims when trust is under pressure? Mike Vizard joins Jon Swartz, Elizabeth Safran, Robert Reeves and Andi Mann to discuss election influence, vulnerability backlogs and SAP’s agent workflow ambitions.

The panel examines the limits of election-infrastructure defenses, reporting on AI propaganda and opinion on AI persuasion. HackerOne’s findings raise a separate operational challenge: discovery is accelerating, but unresolved vulnerability backlogs keep growing despite faster remediation.

SAP’s Joule Work rollout brings agent orchestration into enterprise workflows. The discussion considers what is available now, what is planned and where human oversight belongs.

Story links:

How much integration does industrial AI need, and how much reporting can maintainers absorb? The panel examines Schneider's pending PTC deal and GitHub's new limits on private vulnerability submissions.

Then the discussion turns to agent permissions, policy-as-code and gateway controls from Apple, AWS, Kong, CData and Cribl, alongside California's investigation of OpenAI. The distinction between announced safeguards and demonstrated protection matters.

Host: Mike Vizard. Guests: Kate Scarcella, Sid Nag and Chris Blask.

Story links:

Who gets a say in the AI buildout, and what should enterprises trust agents to do? Mike Vizard, Jack Gold, Chris Blask and Jon Swartz connect the politics of data centers with the practical limits of automated IT operations.

The conversation moves from public opposition and community commitments to AWS's cloud recommendations and the risk of meaning drifting between systems. It closes with Apache Ossie and the push for shared semantic models across data and AI tools.

Host: Mike Vizard. Guests: Jack Gold, Chris Blask and Jon Swartz.

READING FOR THIS EPISODE

Who carries the risk when payments, AI agents and new communities are built by technology companies? On this episode of Techstrong Gang, Mike Vizard and the panel examine three stories about accountability and control.

First, payments giants launch OUSD to challenge Tether and Circle, raising questions about stablecoin economics, distribution and who benefits from the payment network. Next, a nonprofit lawsuit against OpenAI over the Hugging Face incident and an FTC probe of AI agent risks put responsibility for autonomous systems in focus. The lawsuit contains allegations, not a ruling, and the investigation is not a finding of wrongdoing.

The discussion also considers Praxis’s proposed technology-focused community in Uruguay: its governance, practical execution and the distance between a digital community and a functioning city.

Host: Mike Vizard.

Stories discussed:

Futurum Media brings Techstrong, Tech Field Day and Visible Impact into a shared media organization. The panel discusses what that means for technology audiences, then examines AI's effects on occupations, software development and enterprise spending, including cybersecurity priorities.

DeepSeek and Huawei's AI chip software efforts raise questions about NVIDIA's ecosystem advantage. The discussion also considers model-security concerns and keeps allegations of distillation distinct from claims about stolen model weights.

Hosts: Alan Shimel and Mike Vizard. Guests: Jon Swartz, Drew Gutstein, Jaime Hampton, Dan O'Brien, Lee Sellers and Stephen Foskett.

Story links:

Who is responsible for proving that AI is safe and reliable? Alan Shimel, Mike Vizard, Andi Mann, Teri Robinson, JP Morgenthal and Chris Blask examine voluntary safety commitments, proposals for reasoning standards and the implications of AMD’s reported World Labs acquisition agreement.

The September 30, 2026 Techstrong Gang discussion considers accountability, traceability and evidence as enterprise teams evaluate AI systems. It also looks at spatial intelligence, simulation and robotics. The reported acquisition remains subject to closing and regulatory approval.

https://techstrong.ai/agentic-ai/trump-tech-execs-sign-voluntary-super-intelligence-safety-accord-favoring-self-regulation/

When should an AI system receive permission to act? This episode explores release safety, the implications of AMD's World Labs agreement and why automated vulnerability repair still needs rigorous validation. Alan Shimel, Mike Vizard, Stephen Foskett, Kate Scarcella and Chris Blask connect model capability to the operational evidence enterprises need.

Recorded September 29, 2026.

Sources:

Who controls an agent once it starts doing the work? Microsoft is extending Copilot into autonomous workplace tasks, while Docker and Nvidia offer different approaches to governing agent execution.

Alan Shimel, Mike Vizard, Stephen Foskett and Tom Hollingsworth also use Cloud Field Day 26 to explore cloud priorities, automation and buyer diligence. Nvidia's claims about preventing an attack are discussed as company claims, not proven guarantees.

Hosts: Alan Shimel and Mike Vizard. Guests: Stephen Foskett and Tom Hollingsworth.

Story links:

AI wearables promise more useful personal assistants, while the EvilTokens takedown exposes AI-assisted phishing and airport LiDAR raises practical questions about data and privacy. Meta's hardware announcements and Qualcomm's on-device AI strategy set up a discussion about where intelligence belongs, who gets access and how organizations earn trust.

The panel examines Microsoft's disruption of EvilTokens and its use of compromised inboxes to prepare fraud. It also explores Outsight's LiDAR-based airport analytics and the trade-offs involved in managing passenger flows without identifying individuals.

Panel: Jack Poller, Tracy Ragan, Jon Swartz and Jeff Reich. Topics: Mike Vizard. Production: Taylor Yaskulka.

AI at Your Fingertips

EvilTokens Takedown

Travel with LiDAR