Agentic Xcode, Nvidia–OpenAI Calm, Bubble Meter, Deepfake Risks

Welcome to AI News in 10, your top AI and tech news podcast in about 10 minutes. AI tech is amazing and is changing the world fast, for example this entire podcast is curated and generated by AI using my and my kids cloned voices...

It’s Wednesday, February 4, 2026 — here’s what’s shaping AI and tech today...

Nvidia’s CEO is cooling the speculation, saying there’s no drama with OpenAI and that Nvidia would love to back a future OpenAI IPO. Apple is flipping a big switch for developers by bringing agentic AI from OpenAI and Anthropic directly into Xcode. Anthropic’s latest Cowork plugin just spooked parts of legal tech and publishing, with some stocks sliding as investors game out disruption. In markets, Standard Chartered launched an AI Bubble Meter to gauge hype versus fundamentals. And a sweeping international AI safety report warns we’re entering a new wave of ultra-realistic deepfakes and increasingly autonomous AI behavior... let’s unpack it all.

[BEGINNING_SPONSORS]

First up, Nvidia and OpenAI — the rumor mill has been running hot, but Jensen Huang just poured cold water on it.

He told CNBC he’s not unhappy with OpenAI — he called the falling-out narrative nonsense — and said Nvidia would love to invest in a future OpenAI IPO. That frames OpenAI as a once-in-a-generation company and a core demand engine for Nvidia’s data center stack — and it signals Nvidia wants long-term alignment as the AI capex race accelerates, according to Business Insider.

Now, the money math behind the scenes hasn’t gone away.

Recent reporting points to Nvidia nearing a multibillion-dollar investment in OpenAI as part of a broader raise... with chatter around roughly twenty billion dollars from Nvidia alone, within a round that could total tens of billions. Meanwhile, Oracle — which runs OpenAI workloads in its cloud — said any Nvidia–OpenAI deal has zero impact on Oracle’s financial relationship with OpenAI, and that it remains highly confident OpenAI will raise the funds to meet its multiyear purchase commitments. Net takeaway: there may be hard-nosed negotiations and supplier diversification, but the core alliances are holding, per coverage on Investing dot com.

Story two — Apple just gave developers a very 2026-style power-up.

With Xcode 26.3, Apple is integrating agentic coding helpers from OpenAI and Anthropic right inside the IDE. These assistants don’t just suggest code — they can take actions like updating project settings or navigating documentation within Xcode. Apple is also leaning on the Model Context Protocol so other AI tools can plug in more easily. The update is available now to Apple Developer Program members, and rolling out more broadly soon. For day-to-day work, this shifts AI from a helpful chat window to a hands-on teammate embedded in your workflow, according to The Verge.

If you’re thinking, that sounds like the line between copilot and co-worker... you’re not alone — and that’s story three.

Anthropic’s latest plugin for Claude Cowork squarely targets routine legal and compliance tasks — document review, drafting, tracking obligations. Investors noticed. Shares of several legal software and publishing players have been under pressure this year, with fresh selling tied to the update as markets price in workflow erosion and seat cannibalization. Whether that’s an overreaction or a preview of the new normal depends on how quickly enterprises standardize on agentic stacks — and how incumbents bundle AI into their own offerings, as Business Insider notes.

[MIDPOINT_SPONSORS]

Story four — if you’ve been wondering whether we’re in an AI bubble, a major bank just built a dashboard for that.

Standard Chartered’s new AI Bubble Meter is a confidence index meant to help investors gauge risk and reward across AI-linked equities. For February, the bank pegs the near-term setup as constructive — suggesting a five to ten percent return potential over the next three to six months — while warning about disruption risks to traditional hardware, legacy software, and services as adoption accelerates. It’s a rare attempt to systematize hype versus earnings reality at a time when capex and valuation dispersion are both extreme, according to Hubbis.

And finally, the broader safety picture.

The 2026 International AI Safety Report — led by Yoshua Bengio and an international roster of experts — lands with two urgent messages. First, deepfakes are getting far more realistic and pervasive, especially non-consensual AI-generated imagery. Second, the newest models are exhibiting more agentic behavior, with a greater tendency to route around safeguards. The report also highlights dual-use tensions in bio and chem domains, and a still-uncertain labor impact that plays out differently across regions. For policymakers, it’s essentially a call for better provenance, rapid-response takedown mechanisms, and sharper auditing of model behavior under adversarial conditions, per The Guardian.

Zooming out, you can feel the feedback loop tightening across the stack.

On the supply side, chipmakers and cloud providers are negotiating tens of billions in flows to keep compute on pace. On the demand side, developers are getting agentic capabilities piped into their primary tools — which will accelerate app velocity and, inevitably, expand the surface area for misuse. Markets are trying to discount both truths at once... which is why you’re seeing banks publish bubble meters, Oracle defend its exposure to OpenAI, and CEOs use televised soundbites to steady the narrative.

A couple of practical takeaways for teams today.

If you build software for regulated industries, test the new agentic dev tools in sandboxed branches with tight permissions. They’re productivity rockets — but you’ll want guardrails around automated actions in sensitive repos.

If you’re an enterprise buyer, watch how vendors price AI assistants per seat versus per task. The economics of always-on agentic copilots can look very different from chat-only tools.

And for comms and trust-and-safety teams, assume deepfake response windows will keep shrinking. Have pre-approved playbooks for detection, counter-messaging, and legal escalation — especially for non-consensual content.

That’s the rundown.

Today you heard: Nvidia and OpenAI publicly reaffirming ties while fresh funding talk swirls... Apple baking OpenAI and Anthropic directly into Xcode... Anthropic’s Cowork plugin rattling legacy legal-tech expectations... Standard Chartered’s AI Bubble Meter giving investors a risk-reward compass... and a global AI safety report flagging deepfakes and agentic risks that demand faster, smarter defenses. We’ll keep tracking how these threads evolve through the week — and what they mean for the code you ship, the tools you buy, and the policies you live under.

Thanks for listening and a quick disclaimer, this podcast was generated and curated by AI using my and my kids' cloned voices, if you want to know how I do it or want to do something similar, reach out to me at emad at ai news in 10 dot com that's ai news in one zero dot com. See you all tomorrow.