• The Deep View
  • Posts
  • Newsom pitches California AI rules as US template

Newsom pitches California AI rules as US template

Welcome back. Microsoft and Google have a potential edge in the race for persistent agents: they already have your apps, your data and, typically, your trust. Google’s new mid-tier Gemini Neon model faces its own test as it tries to win over users to make it their default model. Meanwhile, California wants to make stronger AI safety rules a national template, but faces an uphill battle in getting Washington to go along. —Jason Hiner

IN TODAY’S NEWSLETTER

1. California tests the limits of state AI rules

2. Why familiar platforms could win the AI agent race

3. Gemini Neon makes its case to be a default model

POLICY

Newsom pitches California AI rules as US template

As it stands, regulation in the US has not been able to keep pace with the rapid development of frontier AI. Several new efforts out of California seek to change that. 

On Wednesday, Calif. Gov. Gavin Newsom unveiled a proposal that would give the state a stronger ability to regulate AI safety. Newsom's proposal would give the state emergency shutdown authority over AI, baseline safety standards for developers, impose mandatory third-party oversight and cap recursive self-improvement. 

"California isn’t waiting until it is too late. We are updating and strengthening our nation-leading framework to move as quickly as technology evolves," Newsom said in a statement. "We’re showing the country what responsible leadership looks like." 

California has largely led the way on state AI regulation, imposing laws seeking to make AI more transparent, establish independent oversight, and require safeguards for child safety around chatbots. However, the regulatory landscape is mixed across the country, creating a patchwork of regulation for something that has a global impact. 

Newsom told reporters on Wednesday that he aims for this to be a framework that can be applied nationally. 

"This takes things to the next level, not only of conversation, but of consciousness in terms of how we effectively regulate artificial intelligence, not only in my home state, but we believe a template for how we should be doing it federally as well," Newsom said.

However, while new regulation is creeping in slowly, a new lawsuit could test the mettle of existing laws against the fast-moving tech: On Monday, California Attorney General Rob Bonta filed a lawsuit against OpenAI alleging that the AI lab prioritized rapid AI development over public safety. The lawsuit centers around the company's breach of HuggingFace in July, arguing that OpenAI was negligent in not adequately securing its testing environments. 

The claim throws a number of allegations at OpenAI, including that it ignored opportunities to prevent the attack, misled consumers about the tech's safety, and withheld information from regulators. 

Legally, the complaint is based on two primary causes of action: Violating the Unfair Competition Law, which claims that the company engaged in deceptive business practices, and a public nuisance claim, which can hold companies liable for infringing on common rights shared by the public—in this case, the right to a safe and secure internet. 

"The company knowingly tested a dangerous product without proper oversight or safety procedures and misled the public about it—this disregard for public safety violated California law," Bonta's office said in a statement. "Without sufficient safeguards and controls, the scope and frequency of this already unacceptable risk will only increase."

OpenAI has since pushed back, with spokesperson Drew Pusateri telling BBC on Monday that the lawsuit "appears to single out OpenAI over a challenge that's been identified by every major AI lab." 

This lawsuit could represent a major test for the efficacy of policing AI through the existing legal system, Tess Vartanian, legal fellow for the Center for Democracy and Technology, told The Deep View, as both claims rest on "novel, untested legal arguments." For instance, the definition of a "public nuisance" has typically been applied to harms caused by things like firearms and opioids. Additionally, the claim of unfair business practices is usually concerning things like false advertising and price-fixing." 

"Regardless of how these arguments fare in court, the Hugging Face incident shows why lawsuits are not enough," said Vartanian. "When autonomous agents can escape a lab and launch cyberattacks, waiting until after the damage is done is a losing strategy. Real accountability requires rules for how frontier models are tested, mandatory incident reporting, and third-party auditing and assessments to prevent incidents like this from happening in the first place."

California's zeal to regulate the AI industry may be a harbinger for what's to come nationally. As the US public largely voted for anti-data center and pro-AI safety candidates in the midterm elections, the industry at large could soon be forced to reckon with a stricter regulatory landscape. However, unless Congress can agree on a federal framework, regulation will only go as far as the border of a given state, limiting its effectiveness. Though Newsom's goal for the California state framework to turn into a federal template is bold, the challenge will be getting enough of Washington to agree on it, especially with the lobbying activity that many of these AI labs and tech giants take part in. 

Nat Rubio-Licht

TOGETHER WITH AMD

AMD Helion™ Agentic AI Rack

AMD Helion™ Agentic AI Rack is an open, rack-scale reference design purpose-built for agentic AI and powered by AMD EPYC™ 9006 Series Server CPUs.

Configurations built for your agentic needs, from a CPU-only rack designed for hyperscale deployments to an integrated CPU and GPU rack for enterprise environments, both built on OCP standards.

PRODUCTS

Why familiar platforms could win the AI agent race

Persistent agents. Get used to that phrase because you're going to be hearing a lot more of it over the next 6 to 12 months. 

On Tuesday at Microsoft Ignite in San Francisco, the hyperscaler unveiled Microsoft Autopilot, its answer to persistent agents. The company dubbed it, "Microsoft’s first enterprise-grade, long-running agent for personal productivity" and it offers many of the same capabilities as OpenAI Dots, Claude Duet, Meta Muse, and GrokBot—but within Microsoft's safety and security constraints. 

If your work life runs within Teams, Outlook, and Microsoft 365, then Autopilot is a personal agent that can take advantage of AI's latest capabilities to proactively assist with your most important tasks and messages. 

Here are some of the specialized and standout features of Autopilot:

  • The cloud-based agent operates as its own entity with memory, computer, and tools, so it can run even when you're offline or your laptop is closed. It can monitor your messages in Teams and Outlook and flag priority items, run recurring tasks, manage projects, and carry out cross-functional workflows. 

  • Each one has its own Microsoft Entra Agent ID and an assigned manager and all of the actions it takes are safety-checked and visible to IT administrators, who can also manage any of those actions. 

  • In Teams, Outlook, documents, and channels, you can @mention the agent to carry out work or follow up on something. Your human teammates can also contact the agent to ask about your calendar or other transactional things that the agent can handle.

  • Access to Autopilot is limited to the Frontier preview program and you'll need a Copilot license with usage-based billing enabled.

Similarly, on Wednesday Google rebranded its Spark agent as "Tasks in Gemini" while making it free to use for everyone, much like Meta's Muse. Google is also trying to help users in its ecosystem automate low-level work using the deep context of people who use Google Workspace and/or Gemini Personal Intelligence. To use it, simply switch from Chat to Tasks in the Gemini interface. You can also now email your agent at [email protected].

Think of persistent agents as the next evolution of personal agents like OpenClaw and others that swept the world in early 2026. All of those agents are still incredibly capable and can do virtually everything that Microsoft Autopilot and other persistent agents can do. The benefit of persistent agents is that they are easier to spin up and they are already running inside platforms you trust and where your most important data lives. While Microsoft Autopilot is squarely focused on corporate usage, Google is pitching its agent as one that can help you pay bills, automate your shopping, make a service appointment for your car, and stay on top of your kids' schedules. As different as the focus is for Microsoft and Google with these two products, these are the two platforms where most people have their logins, work apps, and individual data, and that makes them the most likely to win trust for trying these kinds of agents. Still, the next big opportunity for both could be rolling out more specialized agents to help with dedicated functions, such as social media, marketing, business ops, project management, legal assistance, and health. 

Disclaimer: Jason Hiner's travel to the Microsoft Ignite 2026 event was paid for by Microsoft. The Deep View's coverage is editorially independent from the companies we cover.

Jason Hiner, Editor-in-Chief

TOGETHER WITH GRANOLA

What did you promise in your last call?

Ask Granola what you promised in your last client call and get a real answer, with a link back to the exact note.

Granola Chat searches across all your meetings, so nothing you agreed to gets lost.

BIG TECH

Gemini Neon makes its case to be a default model

Google has released its answer to its rivals' mid-tier models. 

On Wednesday, the tech giant unveiled Gemini 4 Neon, the latest addition to its frontier model line-up. The model is comparable to competitors like Anthropic's Claude Opus 5.5 and OpenAI's GPT-6.2 Sol, with improved performance on complex tasks such as software engineering, knowledge work, and science workflows.

The model scores third in the Artificial Analysis Intelligence Index at it's "xhigh" thinking level, sitting just behind GPT-6.2 Astra and Claude Fable 5.2, which it calls "larger, slower models" in comparison. 

  • Google says the model outranks Claude Opus 5.5 and GPT-6.2 Sol, Anthropic and OpenAI's mid-tier models, in benchmarks for tasks such as knowledge work, automation, reasoning, multimodal understanding, cybersecurity, and science and math. 

  • Efficiency is also a major factor in this release, scoring 32% lower cost per task compared to Gemini Carbon, the company's model that sits one tier above Neon, while completing tasks faster. 

  • Additionally, Google built in real-time safeguards against cyber and biosecurity misuse, improved on alignment compared to Gemini Carbon, and is more resilient to prompt injection attacks compared to Anthropic and OpenAI models, according to the Gray Swan benchmark. 

As for cost, the model sits at $3 per million input tokens, 15 cents per million cached input tokens and $15 per million output tokens. That puts it right in between the cost of GPT-6.2 Sol, which costs $2 per million input tokens, $10 per million output tokens, and $10 cents per million cached, and Claude Opus 5.5, which sits at $4 per million input tokens, $20 per million output tokens, and $20 cents per million cached. 

It's also the latest of Google's waterfall of new models, which released Gemini 4 Hydrogen last week, the lightest-weight model in its line-up, comparable to OpenAI's Terra and Anthropic's Sonnet.

At this point, Google is fully back in the frontier model race after inconsistent releases throughout 2026. The question now is about what, exactly, will give any of these AI labs a bigger competitive advantage over one another. If there's one overarching thread that I've seen at Microsoft Ignite over the past few days, it's that cost efficiency is quickly edging out capability as the determining factor in whether or not an enterprise decides to use a model. This is becoming even more true as models themselves become commoditized and more model routing services start to pop up that automatically pick the right sized (and right priced) model for any given tasks. So while Google's new model edges out competitors in its category, its middling price for comparable performance might make it hard to stand out in a sea of other choices. If it can, the one good thing is that mid-tier models can often become a solid default choice.

Nat Rubio-Licht

LINKS

  • Copilot Workspace: Platform where people and their agents work together

  • Code 2: Microsoft's new model at the cost-performance frontier for enterprise coding

  • GPT-6.2 Astra and Sol: OpenAI released its newest frontier models 

  • OpenAI: updated API usage dashboard for easier usage monitoring

  • Deloitte: Associate, Analytics and Insights

  • EY: Cyber - AI Engineer - Manager - Consulting

  • ByteDance: Applied AI Architect - Seedance / Seedream / Seed Audio

  • Greylock Partners: AI Researcher, Agent Systems

GAMES

Which image is real?

Login or Subscribe to participate in polls.

A QUICK POLL BEFORE YOU GO

Have you tested out using persistent agents in your day-to-day work?

Login or Subscribe to participate in polls.

The Deep View is written by Nat Rubio-Licht, Sabrina Ortiz, Jason Hiner, Faris Kojok and The Deep View crew. Please reply with any feedback.

Thanks for reading today’s edition of The Deep View! We’ll see you in the next one.

“[This image] has consistent shadows and does not blur perspective on the roof to hide the number of tiled rows.”


“The focus on [this image] is correct, especially when you look at the roof in the far upper left corner of the photo.”


“More realistic shadows in [this image].”

“The colors and definition are more vivid in [the other image], muted in [this image]. [This image] looks more like a painting than an actual photograph.”


“Rook in the foreground looks cartoon-like”


“I felt that the dimensions in [this image] were wrong”

If you want to get in front of an audience of 750,000+ developers, business leaders and tech enthusiasts, get in touch with us here.