Latest / DX Today | No-Hype Podcast & News About AI & DX / Executive Briefing: Google Wyoming, Uber Agents, AI Law, Kernel

Transcript

Executive Briefing: Google Wyoming, Uber Agents, AI Law, Kernel · DX Today | No-Hype Podcast & News About AI & DX

• 0:00
  SPEAKER_01: Welcome to the DX Today Executive Briefing.
• 0:03
  SPEAKER_01: Today is Friday, November 6th, 2026.
• 0:06
  SPEAKER_01: I'm Elena, and this show is a deep dive review of this morning's DX Today newsletter, built for the leaders who have to make real decisions about artificial intelligence, not just read about them.
• 0:17
  SPEAKER_00: And I'm Marcus.
• 0:19
  SPEAKER_00: I come at these stories the way someone who has had to run a large technology organization would.
• 0:24
  SPEAKER_00: I want to know what actually changes on Monday morning, who owns it, what breaks its scale, and what a buyer should ask a vendor before signing anything.
• 0:34
  SPEAKER_00: Today's theme is that the artificial intelligence industry is entering what the addition calls its accountability phase.
• 0:42
  SPEAKER_01: The money still flows into physical infrastructure, but the builders must prove their power plants will not land on local utility bills.
• 0:50
  SPEAKER_01: And the most hyped consumer behavior of the year, agents acting on our behalf, is close to invisible in real transaction data.
• 0:58
  SPEAKER_01: Here are the four stories in order.
• 1:01
  SPEAKER_01: Story one is Google's$17 billion commitment to its first data center in Wyoming, and the 702-megawatt fuel cell plant that Google is building to power it.
• 1:12
  SPEAKER_01: The interesting part is less the dollar figure and more the power plan and the promises attached to it.
• 1:19
  SPEAKER_00: Story two is Uber telling its investors that AI agents booked just four rides in its busiest week of the year, in a week with roughly 320 million trips and orders.
• 1:31
  SPEAKER_00: Four.
• 1:33
  SPEAKER_00: I have been waiting for someone to put a number on the agentic commerce story, and now someone has.
• 1:39
  SPEAKER_00: Story three moves to Washington.
• 1:41
  SPEAKER_00: Democrats are set to take control of Congress in January, and they are promising hearings and guardrails on artificial intelligence.
• 1:49
  SPEAKER_01: After a midterm election in which several bets made by AI industry political groups did not pay off.
• 1:56
  SPEAKER_00: There are two specific bills in this story that enterprise legal teams should know by name.
• 2:01
  SPEAKER_00: And story four is the one that should make every security team sit up.
• 2:05
  SPEAKER_00: OpenAI disclosed that an unreleased cyber model built working root exploits for 948 Linux kernel vulnerabilities in a single automated run at a cost of about$1.2 million.
• 2:18
  SPEAKER_00: We will spend real time on what that does to the threat model.
• 2:22
  SPEAKER_01: After the four deep dives, we will connect them because the edition's bottom line makes a strong argument about how they fit together.
• 2:29
  SPEAKER_01: Then we will give you three or four concrete actions for this week and finish with what to watch.
• 2:35
  SPEAKER_00: One note on how we work.
• 2:57
  SPEAKER_00: Let's do it.
• 2:58
  SPEAKER_00: And I promise to ask at least one uncomfortable question about fuel cells before we are done.
• 3:04
  SPEAKER_00: Story one.
• 3:31
  SPEAKER_01: Now the addition makes a point right away.
• 3:34
  SPEAKER_01: A$17 billion data center would be a notable line item in the artificial intelligence build-out on its own.
• 3:42
  SPEAKER_01: What makes this worth an executive's attention is how the power is being supplied.
• 3:47
  SPEAKER_00: And I think that is the right instinct.
• 3:50
  SPEAKER_00: Because for the last two years, we have all watched hyperscaler announcements that were basically a number and a map pen.
• 3:57
  SPEAKER_00: This one has a different structure.
• 3:59
  SPEAKER_00: The number's big, but the real content is the power plan.
• 4:03
  SPEAKER_00: So walk me through it because the details are unusual.
• 4:06
  SPEAKER_01: According to Google, the project is contracted with American Electric Power, which the addition shortens to AEP, to feature a commercial-scale 702-megawatt fuel cell development with carbon capture and storage.
• 4:20
  SPEAKER_01: Google describes it as the largest planned development of its kind to date.
• 4:24
  SPEAKER_01: The fuel cells will be made in America by Bloom Energy, and they convert natural gas into electricity.
• 4:31
  SPEAKER_01: Google says the project involves capturing and permanently storing a large majority of its carbon emissions underground at Tallgrass's nearby eastern Wyoming sequestration hub.
• 4:42
  SPEAKER_00: So just to be plain about it, this is natural gas.
• 4:46
  SPEAKER_00: It is not wind, it is not solar, it is not nuclear.
• 4:49
  SPEAKER_00: It is gas run through fuel cells instead of turbines with the carbon captured and stored.
• 4:55
  SPEAKER_00: And I want to flag, as the edition does, that these are Google's descriptions.
• 5:00
  SPEAKER_00: Google says it is the largest of its kind.
• 5:02
  SPEAKER_00: Google says a large majority of the carbon will be stored.
• 5:06
  SPEAKER_00: Those are claims from the company.
• 5:07
  SPEAKER_00: And they are claims about plans, not about a plant that is running today.
• 5:12
  SPEAKER_01: Fair.
• 5:12
  SPEAKER_01: And the edition keeps that attribution.
• 5:15
  SPEAKER_01: There is more.
• 5:16
  SPEAKER_01: Google is also sourcing power from gas generators operated by Tallgrass Energy and Black Hills Energy.
• 5:23
  SPEAKER_01: And it says the facility will be air-cooled to minimize strain on local water resources.
• 5:27
  SPEAKER_01: So there are two local concerns being addressed in the design itself: electricity and water.
• 5:34
  SPEAKER_00: Which in my experience is how these projects actually get approved now.
• 5:38
  SPEAKER_00: Ten years ago, a company would announce a campus, and the conversation with the county was about tax abatements.
• 5:44
  SPEAKER_00: Now the conversation is where does your power come from and are you going to drink our water?
• 5:49
  SPEAKER_00: The design is the answer to the permit hearing.
• 5:56
  SPEAKER_01: The addition says hyperscalers has spent the past two years discovering that the binding constraint on AI is not chips, but electrons, and that waiting in a utility interconnection queue can cost years.
• 6:09
  SPEAKER_01: Building dedicated generation next to the load and financing it as part of the project is a way to collapse that timeline.
• 6:16
  SPEAKER_01: Fuel cells running on natural gas paired with geologic carbon storage are described as Google's attempt to get firm around-the-clock power without abandoning its emissions commitments on paper.
• 6:28
  SPEAKER_00: Notice the phrase on paper.
• 6:31
  SPEAKER_00: That is the addition being a little wry, and I like it.
• 6:34
  SPEAKER_00: The question for any buyer is whether the emissions story holds up in practice.
• 6:39
  SPEAKER_00: And that is not something we can answer from this announcement.
• 6:41
  SPEAKER_00: What we can say is that the strategic logic is clear.
• 6:45
  SPEAKER_00: If the grid cannot deliver electrons on your timeline, you bring your own electrons.
• 6:50
  SPEAKER_00: That is a very different business than renting servers.
• 6:53
  SPEAKER_00: It really is.
• 6:54
  SPEAKER_00: It means a cloud provider becomes, in part, an energy developer.
• 6:58
  SPEAKER_01: And that leads to the second half of the announcement, which the edition says is aimed squarely at the political environment.
• 7:05
  SPEAKER_01: Google says it will pay for 100% of the power it uses, and that it will cover the infrastructure needs directly driven by its growth when it builds data centers.
• 7:15
  SPEAKER_00: That pledge is an answer to the loudest local objection to AI campuses, and the edition names it plainly.
• 7:22
  SPEAKER_00: The fear is that residential ratepayers will end up subsidizing grid upgrades for the largest technology companies.
• 7:29
  SPEAKER_00: If your neighbor's electric bill goes up because a data center showed up, you will not like the data center, however many jobs it brings.
• 7:36
  SPEAKER_01: And this ties directly into story three, which we will get to.
• 7:41
  SPEAKER_01: The edition notes that AI and data centers have become kitchen table issues, tied to jobs, children, and electricity bills.
• 7:49
  SPEAKER_01: So Google's pledge in Wyoming is, in a sense, preparing for a political climate that is already here.
• 7:55
  SPEAKER_01: Google is also standing up a community council with Tallgrass Energy, Black Hills Energy, Related Digital, AEP, and Bloom Energy, a structure the edition describes as designed to give local stakeholders a formal channel into the project.
• 8:11
  SPEAKER_00: Let me push on that one, because a community council can be wonderful or it can be decoration.
• 8:17
  SPEAKER_00: A council with the project partners on it is a start.
• 8:21
  SPEAKER_00: The test is whether local people have any real say and whether the cost commitments are written down somewhere enforceable.
• 8:29
  SPEAKER_00: I am not saying Google is being cynical.
• 8:31
  SPEAKER_00: I am saying as a buyer or a neighbor, I would want to see the paper.
• 8:36
  SPEAKER_01: That is a reasonable test.
• 8:38
  SPEAKER_01: And I would add that the edition itself frames the community cost commitments as becoming a condition of permission to build.
• 8:45
  SPEAKER_01: So it is not just goodwill.
• 8:47
  SPEAKER_01: It is the price of admission.
• 8:49
  SPEAKER_01: Now let's talk about the economics Google is presenting, because the addition makes a sharp observation here.
• 8:56
  SPEAKER_01: Go for it.
• 8:56
  SPEAKER_01: This is my favorite part of the story.
• 8:59
  SPEAKER_01: Google says construction will support more than 8,800 jobs statewide, with an on-site workforce of roughly 2,500 skilled jobs at peak and approximately$610 million in average annual labor income.
• 9:14
  SPEAKER_01: Once operating, Google says the data center will support over 3,700 jobs annually, including directly creating 250 roles in the state and generate hundreds of millions of dollars in local property taxes for Laramie County.
• 9:29
  SPEAKER_00: And here is the contrast the edition calls the honest arithmetic of AI infrastructure.
• 9:35
  SPEAKER_00: 2,500 construction workers at peak and 250 direct permanent roles.
• 9:40
  SPEAKER_00: That is 10 to 1.
• 9:42
  SPEAKER_00: And I appreciate that the newsletter says it out loud, because the headline number, over 3,700 jobs, includes the jobs that are supported indirectly.
• 9:52
  SPEAKER_00: The direct number is 250.
• 9:54
  SPEAKER_00: The addition's framing is that data centers are capital-intensive, not labor-intense.
• 10:00
  SPEAKER_00: Their local value proposition rests on property taxes, construction activity, and supplier spending far more than on long-term employment.
• 10:10
  SPEAKER_01: And it says executives negotiating their own sites should expect communities to understand that math increasingly well.
• 11:06
  SPEAKER_00: Which is a good point, but it cuts in two directions.
• 11:09
  SPEAKER_00: Control of generation lets you promise firmer timelines only if the generation itself arrives on time.
• 11:15
  SPEAKER_00: So the diligence question is not just do you have power?
• 11:19
  SPEAKER_00: It is, is the power contracted or merely planned?
• 11:22
  SPEAKER_00: And the edition's strategic takeaway says exactly that.
• 11:25
  SPEAKER_00: Let me read it in plain words.
• 11:27
  SPEAKER_00: The takeaway is addressed to CIOs, CFOs, and infrastructure strategy leaders.
• 11:34
  SPEAKER_00: Treat energy strategy as part of your AI vendor due diligence.
• 11:38
  SPEAKER_01: Ask cloud and co-location partners where the power for your future capacity will come from, whether it is contracted or merely planned.
• 11:47
  SPEAKER_01: And who bears the cost of GWID upgrades?
• 11:50
  SPEAKER_01: Google's Wyoming design shows that the providers able to promise dependable AI capacity are increasingly the ones building their own power, and that community cost commitments are becoming a condition of permission to build.
• 12:17
  SPEAKER_00: Is it a signed power agreement or a hope?
• 12:20
  SPEAKER_00: A good vendor will have a crisp answer.
• 12:22
  SPEAKER_00: A vague answer is information too.
• 12:25
  SPEAKER_00: And for the finance side, I would say this changes how you think about the long-term price of compute.
• 12:31
  SPEAKER_00: If energy engineering is now a core input, then providers who have solved power may have more pricing stability, and providers who have not may face costs they eventually pass on.
• 12:43
  SPEAKER_00: That is my reasoning, not something the addition reports, but it follows from the logic of the story.
• 12:52
  SPEAKER_00: And there is a second-order effect I find interesting.
• 12:55
  SPEAKER_00: If every large provider starts building dedicated generation, then the energy sector and the technology sector become much more entangled.
• 13:04
  SPEAKER_00: A technology company will have partners in gas, in fuel cells, in carbon storage.
• 13:09
  SPEAKER_00: That is a lot of new dependencies.
• 13:12
  SPEAKER_01: And a lot of new places for something to go wrong, which is why the community and cost commitments matter.
• 13:20
  SPEAKER_01: The summary is that story one is really a story about permission.
• 13:25
  SPEAKER_01: Google is not just buying land and chips, it is assembling the power, the water plan, the cost pledge, and the council that together make the project buildable.
• 13:36
  SPEAKER_00: And the lesson I take for executives is to treat the power plan as a first-class part of any capacity conversation.
• 13:43
  SPEAKER_00: Next time someone shows you a roadmap of new regions and new capacity, ask where the electrons come from.
• 13:50
  SPEAKER_01: Story two, and this is the one I have been looking forward to.
• 13:54
  SPEAKER_01: For a year, the technology industry has told a story in which AI agents become the new front door to commerce.
• 14:01
  SPEAKER_01: You tell an assistant what you want, and it books the car, orders dinner, and handles the rest.
• 14:34
  SPEAKER_00: And I want to be careful here to say what that measures.
• 14:38
  SPEAKER_00: It is the number of rides ordered through an AI agent in one week in a company that, as we are about to hear, handled hundreds of millions of trips.
• 14:48
  SPEAKER_00: It is a small number by any standard.
• 14:51
  SPEAKER_00: Here is the scale.
• 14:53
  SPEAKER_00: Uber's chief financial officer, Balaji Krishnamurti, told analysts, as TechCrunch reported, that Uber handled around 320 million trips and orders across Uber and Uber Eats during that Halloween week.
• 15:08
  SPEAKER_01: And on the food side, only 500 Uber Eats orders were placed by an AI agent.
• 15:14
  SPEAKER_01: So four rides and 500 meals, against roughly 320 million trips and orders.
• 15:21
  SPEAKER_00: I did the math on a napkin, and I will label it my own arithmetic.
• 15:26
  SPEAKER_00: 504 out of 320 million is a rounding error.
• 15:31
  SPEAKER_00: It is a tiny fraction of 1%, so small you would struggle to see it on a chart.
• 15:36
  SPEAKER_00: The CEO, Dara Kos Rashahi, described the traction as very, very early and very limited so far, which is about as plain as an executive gets on an earnings call.
• 15:47
  SPEAKER_01: And it is not only Uber.
• 15:49
  SPEAKER_01: TechCrunch noted that DoorDash said on its own earnings call that it had received only a few hundred orders from AI agents to date.
• 15:57
  SPEAKER_01: So two of the largest platforms in local commerce on two separate calls are giving a similar message.
• 16:04
  SPEAKER_01: The agent channel exists and it is barely registering.
• 16:08
  SPEAKER_00: Now let me play the skeptic the other direction for a second, because I do not want to overreact to one data point.
• 16:14
  SPEAKER_00: Maybe it is early because the agents are early.
• 16:18
  SPEAKER_00: Maybe the assistants have only recently been able to book things.
• 16:22
  SPEAKER_00: Maybe the number triples next quarter.
• 16:24
  SPEAKER_00: A small number today does not say what the number will be in two years.
• 16:51
  SPEAKER_00: And that is the whole story in one sentence, really.
• 16:54
  SPEAKER_00: If it takes 50 seconds to book a ride, what exactly is the agent saving me?
• 16:59
  SPEAKER_00: I would have to open the assistant, describe what I want, wait for it to confirm, and maybe correct it.
• 17:04
  SPEAKER_00: That is not obviously faster than tapping a button in an app I already know where my address and payment are already saved.
• 17:12
  SPEAKER_00: The edition puts it this way: when the existing interface is already that fast, an agent has very little friction to remove.
• 17:20
  SPEAKER_01: The core value proposition of agentic commerce, saving the user time and effort, is weakest precisely in the apps that have spent a decade optimizing for speed.
• 17:42
  SPEAKER_00: Disputing a billing error with three departments is annoying.
• 17:46
  SPEAKER_00: Reconciling a stack of invoices is annoying.
• 17:49
  SPEAKER_00: So I would not expect agents to win first in the most polished consumer apps in the world.
• 17:56
  SPEAKER_01: Let me bring in the other side because the edition is careful not to say Uber is dismissing the channel.
• 18:02
  SPEAKER_01: Kos Roshahi said that every AI assistant company is talking to Uber, and that he expects agents to be a net positive for demand on the platform.
• 18:11
  SPEAKER_01: He also noted there are around eight major AI assistants on the market, and that he expects that to grow to more than 10.
• 18:18
  SPEAKER_00: So Uber is not saying no.
• 18:20
  SPEAKER_00: Uber is saying we are open, we are talking to everybody, and so far almost nobody is showing up.
• 18:26
  SPEAKER_00: Which is a sensible position for a company that has to keep every door open without knowing which one customers will walk through.
• 18:33
  SPEAKER_00: And there is a statement in Uber's prepared remarks that I think matters a lot.
• 18:39
  SPEAKER_00: Kosura Shahi said direct consumer relationships are central to the company's approach to AI.
• 18:45
  SPEAKER_00: And he cited 215 million monthly active platform consumers.
• 18:50
  SPEAKER_01: The additions reading is that Uber is positioning agents as an additional distribution channel, layered on top of its own app, not as a replacement for it.
• 19:06
  SPEAKER_00: An AI assistant that books your ride sits between the brand and the consumer.
• 19:12
  SPEAKER_00: If I am Uber, I have spent years and enormous money building a direct relationship with 215 million people.
• 19:20
  SPEAKER_00: Why would I hand the front door to a third party that can swap me out for a competitor in one sentence?
• 19:27
  SPEAKER_01: And the addition makes the same point.
• 19:29
  SPEAKER_01: Every platform that has built a direct relationship has reason to be cautious about handing that position to a third party.
• 19:36
  SPEAKER_01: And the striking thing is what the numbers suggest about consumers.
• 19:39
  SPEAKER_01: For now, consumers are making that same choice on their own.
• 19:43
  SPEAKER_01: They are going straight to the app.
• 19:45
  SPEAKER_00: I find that fascinating because the usual story is that platforms are fighting to keep customers while technology pulls customers away.
• 19:53
  SPEAKER_00: Here, the customers are doing the protecting.
• 19:55
  SPEAKER_00: They like the app, it works, they have no reason to leave.
• 19:59
  SPEAKER_01: That is exactly where the addition lands.
• 20:01
  SPEAKER_01: It says none of this undermines the case for agents inside the enterprise, where workflows are slow, repetitive, and expensive, and where agents can remove hours rather than seconds.
• 20:13
  SPEAKER_00: Hours versus seconds is the contrast.
• 20:15
  SPEAKER_00: Aaron Powell And this is the point I would underline for anyone running internal operations.
• 20:20
  SPEAKER_00: Consider a procurement approval that takes three days of emails, or a customer onboarding that crosses five systems.
• 20:26
  SPEAKER_00: There, an agent has a lot of friction to remove.
• 20:29
  SPEAKER_00: That is a very different ratio from 50 seconds to book a car.
• 20:33
  SPEAKER_00: The same technology can look weak in one setting and strong in another purely because of how much friction exists beforehand.
• 20:40
  SPEAKER_00: The edition says it should recalibrate expectations for consumer-facing agent strategies.
• 20:46
  SPEAKER_00: And for any business plan that assumes customers will soon arrive mainly through someone else's assistant.
• 20:53
  SPEAKER_00: The gap between the volume of agent announcements and the volume of agent transactions is now measurable, and it is very wide.
• 21:07
  SPEAKER_00: Count the transactions, not the press releases.
• 21:10
  SPEAKER_00: If your strategy deck says that 30% of future sales will come through assistance, ask what the number is today and ask who is measuring it.
• 21:18
  SPEAKER_01: And the edition closes the story with a useful framing for executives.
• 21:22
  SPEAKER_01: Separate two questions that are often blurred together.
• 21:26
  SPEAKER_01: One is whether agents will matter.
• 21:28
  SPEAKER_01: The other is whether they matter to your customers this year.
• 21:32
  SPEAKER_01: Uber's data answers the second question with unusual clarity while leaving the first open.
• 21:38
  SPEAKER_00: Which is a great place, to be honest.
• 21:39
  SPEAKER_00: I would not write off agents, and I would not build a plan that depends on them arriving next quarter.
• 21:45
  SPEAKER_00: I would instrument, I would pilot, and I would keep investing where the friction is highest.
• 21:49
  SPEAKER_00: Here is the strategic takeaway.
• 21:52
  SPEAKER_00: Addressed to chief digital officers, product leaders, and heads of e commerce.
• 21:57
  SPEAKER_00: Instrument agent traffic before you invest.
• 22:00
  SPEAKER_00: Heavily in it.
• 22:01
  SPEAKER_00: Measure how many transactions actually arrive through AI assistance and compare the time an agent saves against the speed of your existing app or checkout.
• 22:10
  SPEAKER_00: Where your own experience is already fast, prioritize protecting the direct customer relationship.
• 22:17
  SPEAKER_01: And treat agent integrations as an inexpensive option, not a core channel.
• 22:22
  SPEAKER_01: Where your processes are slow and complex, the case for agents is far stronger.
• 22:27
  SPEAKER_00: And that is where to focus first.
• 22:30
  SPEAKER_00: And I will add one operational note of my own.
• 22:33
  SPEAKER_00: Tagging the source of every order is a small piece of engineering, but it is the thing that makes all the later decisions honest.
• 22:40
  SPEAKER_00: If you cannot tell an agent order from a human order, you are guessing.
• 22:44
  SPEAKER_00: Uber could say four and five hundred because someone had built the measurement.
• 22:49
  SPEAKER_01: That is a nice point.
• 22:50
  SPEAKER_01: And it is worth noticing that the whole news value of this story comes from a company being willing to publish an uncomfortable number.
• 22:59
  SPEAKER_01: In an accountability phase, that kind of candor may become more valuable, not less.
• 23:04
  SPEAKER_01: Agreed.
• 23:05
  SPEAKER_00: And it sets up a contrast with story three, where the uncomfortable numbers come from the ballot box rather than the earnings call.
• 23:12
  SPEAKER_01: Story three takes us to Washington.
• 23:15
  SPEAKER_01: The midterm elections have changed the outlook for federal AI policy.
• 23:19
  SPEAKER_01: Democrats are set to take control of Congress in January of 2027, and their leadership is already signaling that artificial intelligence will be on the agenda.
• 23:28
  SPEAKER_01: California Representative Ted Liu, a member of the House Democratic leadership team, said at a post-midterm news conference that the new majority would hold AI hearings and pursue reasonable guardrails on AI, according to the National News Desk.
• 23:42
  SPEAKER_01: Liu accused President Donald Trump and Congressional Republicans of making a conscious decision to let the AI industry run wild.
• 24:02
  SPEAKER_00: What matters for a business audience is the practical signal, which is that hearings are coming and the leadership says so out loud.
• 24:10
  SPEAKER_00: Agreed.
• 24:11
  SPEAKER_00: And the edition handles it that way.
• 24:13
  SPEAKER_00: Now advocates are pointing to the electorate as their mandate.
• 24:18
  SPEAKER_01: An NBC News exit poll found that over half of voters said AI and data centers were a crucial or major factor in their vote for Congress.
• 24:28
  SPEAKER_01: That finding was cited by Jim Steyer, the founder and chief executive of Common Sense Media.
• 24:33
  SPEAKER_00: Over half.
• 24:35
  SPEAKER_00: I think that is the single most important sentence in this story for anyone who sells or deploys AI.
• 24:42
  SPEAKER_00: Whether you love the technology or fear it, voters are telling pollsters it mattered to how they voted.
• 24:47
  SPEAKER_00: That used to be a niche topic for policy walks.
• 24:50
  SPEAKER_00: The edition's phrase is that AI and the data center buildout have become kitchen table issues tied to jobs, children, and electricity bills.
• 24:59
  SPEAKER_01: And notice the connection back to story one: electricity bills.
• 25:03
  SPEAKER_01: Google is pledging to pay for 100% of the power it uses in Wyoming.
• 25:08
  SPEAKER_01: Now we see a political reason that pledges not just good manners, it is a response to something voters are paying attention to.
• 25:16
  SPEAKER_00: Right.
• 25:16
  SPEAKER_00: The Wyoming announcement and this exit poll are two ends of the same wire.
• 25:21
  SPEAKER_00: One is a company trying to get ahead of the objection, the other is the objection showing up in the voting booth.
• 25:26
  SPEAKER_00: Let me lay out what the edition says about the legislative agenda because it is specific.
• 25:32
  SPEAKER_01: The near-term agenda is likely to start narrow.
• 25:35
  SPEAKER_01: Danny Weiss, the Chief Advocacy Officer of Common Sense Media, said the new Congress should revisit two bipartisan bills.
• 25:42
  SPEAKER_01: The first is the GARD Act, introduced by Senator Josh Hawley in October of 2025 with Senators Richard Blumenthal, Katie Britt, Mark Warner, and Chris Murphy.
• 25:53
  SPEAKER_00: And what would it do?
• 25:54
  SPEAKER_00: The edition says it would ban AI companions for minors, require AI chatbots to disclose their non-human status, and create new crimes for companies that knowingly make available to minors AI companions that solicit or produce sexual content.
• 26:11
  SPEAKER_00: So that is child safety, plainly.
• 26:14
  SPEAKER_01: The second is the AI Lead Act, introduced by Senate Democratic whip Dick Durbin together with Senator Hawley in September of 2025.
• 26:23
  SPEAKER_01: It would classify AI systems as products and create a federal cause of action for products liability claims when an AI system causes harm.
• 26:42
  SPEAKER_00: Right now, when software causes harm, the legal theories are murky.
• 26:46
  SPEAKER_00: Software is often treated as a service, with liability limited by contract.
• 26:51
  SPEAKER_00: If an AI system is classified as a product, the whole frame shifts.
• 26:56
  SPEAKER_00: Product makers can be liable for defects, and a federal cause of action would let injured people sue.
• 27:02
  SPEAKER_01: The addition spells out the consequences.
• 27:04
  SPEAKER_01: A product's liability standard would change how AI vendors write contracts, how they allocate indemnity, and how much testing evidence buyers will demand before deployment.
• 27:14
  SPEAKER_01: Three things contracts, indemnity, and testing evidence.
• 27:18
  SPEAKER_00: And I would underline testing evidence.
• 27:20
  SPEAKER_00: Today a lot of AI procurement runs on demos and benchmarks.
• 27:24
  SPEAKER_00: If a product's liability standard comes, a buyer is going to want to see how the system was tested, on what, with what results, and what happened when it failed.
• 27:33
  SPEAKER_00: That is a documentation burden, and it will fall on both vendors and the companies that deploy their systems.
• 27:39
  SPEAKER_00: Let me add a note of balance.
• 27:41
  SPEAKER_00: These are bills that have been introduced.
• 27:44
  SPEAKER_01: The edition says common sense media thinks the new Congress should revisit them.
• 27:48
  SPEAKER_01: It does not say they will pass.
• 27:51
  SPEAKER_00: So this is a direction of travel, not a forecast of law.
• 27:55
  SPEAKER_00: Yes, and I would say the same about the whole story.
• 27:58
  SPEAKER_00: The edition is careful, and so should we be.
• 28:01
  SPEAKER_00: Hearings are not laws.
• 28:03
  SPEAKER_00: Introduced bills are not enacted bills, but the agenda tells you what questions will be asked in public.
• 28:10
  SPEAKER_00: And the questions shape what boards and insurers worry about even before any law passes.
• 28:16
  SPEAKER_01: Now, to the part of the story that I found most striking, which is the political spending.
• 28:21
  SPEAKER_01: The election results also tested the political spending power of the AI industry itself, and several of its bets did not pay off.
• 28:29
  SPEAKER_01: TechCrunch reported that two AI-focused political action committees, American Mission and Building the Future, collectively spent$4.8 million backing South Carolina Senate candidate Darling Graham, who lost.
• 28:44
  SPEAKER_00: $4.8 million and a loss.
• 28:47
  SPEAKER_00: And the addition lists more.
• 28:49
  SPEAKER_00: According to TechCrunch, American Mission, the Republican spending arm of the Leading the Future network, backed by OpenAI President Greg Brockman and Andreysen Horowitz, spent more than$500,000 apiece supporting House candidates Aaron Flint and Lori Buckout, both of whom lost.
• 29:06
  SPEAKER_01: The same report said Leading the Future spent$27.38 million in the primaries, including$8 million opposing New York Assembly member Alex Boers, and told TechCrunch it plans to keep spending in the 2028 elections.
• 29:23
  SPEAKER_01: So the group is not retreating.
• 29:44
  SPEAKER_01: And the edition is even-handed about the other side, too.
• 29:46
  SPEAKER_01: On the safety side, TechCrunch reported, a group called the Guardrails Alliance opposed Representative Jimmy Gomez and Aaron Flint, both of whom lost, and spent$300,000 supporting Angela Gonzalez-Tores, who defeated Gomez.
• 30:02
  SPEAKER_01: So this was not a clean sweep for either camp.
• 30:05
  SPEAKER_00: Right.
• 30:06
  SPEAKER_00: So I would resist a simple narrative of industry money loses, safety money wins.
• 30:12
  SPEAKER_00: The record in this edition is more tangled than that.
• 30:15
  SPEAKER_00: One safety aligned group opposed two candidates and both lost.
• 30:19
  SPEAKER_00: And it backed one who won.
• 30:20
  SPEAKER_00: That is mixed.
• 30:22
  SPEAKER_00: What it does show is that big spending on AI politics is now a real feature of elections, and it does not buy outcomes automatically.
• 30:30
  SPEAKER_01: I would add my own analysis here.
• 30:32
  SPEAKER_01: When politics becomes a place where AI companies spend tens of millions of dollars, it also becomes a place where AI companies are scrutinized for spending it.
• 30:42
  SPEAKER_01: Money is visible.
• 30:43
  SPEAKER_01: A hearing is a natural venue to ask who funded what.
• 30:47
  SPEAKER_01: That is my reasoning, not a reported fact, but I suspect the accountability theme reaches the industry's own political activity.
• 30:54
  SPEAKER_00: That is plausible.
• 30:55
  SPEAKER_00: And as a practical matter for a company that is not a giant AI lab, the question is how closely you want your brand attached to any of this.
• 31:03
  SPEAKER_00: If you are a deployer, a bank or a retailer or a hospital, you are probably not in the pack game, but your vendors might be.
• 31:11
  SPEAKER_00: It is worth knowing where your suppliers stand.
• 31:14
  SPEAKER_01: The addition also includes a voice from outside the partisan arena.
• 31:18
  SPEAKER_01: Anton DeBurra, co-director of the Johns Hopkins Institute for Assured Autonomy, said he expects the federal government to change its approach to regulating AI and that new AI-centric laws are inevitable.
• 31:31
  SPEAKER_01: He called concerns that a national safety framework would stop the United States from competing with China overblown, arguing that well-thought-out regulations could help accelerate the technology by avoiding problems.
• 31:44
  SPEAKER_00: This is the argument I find most useful for a skeptical executive.
• 31:49
  SPEAKER_00: The standard fear is that regulation slows innovation and hands the lead to arrival.
• 31:54
  SPEAKER_00: Deborah says the opposite, that sensible rules can speed things up by avoiding problems.
• 32:00
  SPEAKER_00: I am not saying he is right in every case, but the logic is easy to follow.
• 32:04
  SPEAKER_00: Prevention can be faster than recovery.
• 32:08
  SPEAKER_01: Let me challenge that a little, Marcus.
• 32:10
  SPEAKER_01: Not every regulation is well thought out.
• 32:13
  SPEAKER_01: The word in his sentence is well thought out.
• 32:16
  SPEAKER_01: A badly designed rule can absolutely slow things down.
• 32:20
  SPEAKER_01: So the practical question is not regulation or no regulation, it is whether the industry engages early enough to help shape rules that are workable.
• 32:30
  SPEAKER_01: That is exactly right.
• 32:32
  SPEAKER_00: And it is why the addition's last line on this matters.
• 32:35
  SPEAKER_00: The takeaway is that the regulatory pendulum has started to swing, and that liability and child safety are the first places it will land.
• 32:44
  SPEAKER_00: If you want to shape the rules rather than react to them, you want your evidence ready before the first hearing.
• 32:50
  SPEAKER_01: Here is the strategic takeaway addressed to general counsel, chief compliance officers, and government affairs leaders.
• 32:58
  SPEAKER_01: Plan for a federal AI agenda that begins with liability and child safety and widens from there.
• 33:05
  SPEAKER_01: Inventory every customer-facing AI system for exposure to minors, review vendor contracts for how they would allocate risk under a product's liability standard, and start assembling the testing documentation that any stricter federal rules would demand.
• 33:22
  SPEAKER_01: Companies that build that evidence now will shape the hearings rather than react to them.
• 33:28
  SPEAKER_00: Three tasks there, and I would add a sequencing note.
• 33:47
  SPEAKER_00: And the contract review is where legal and procurement meet.
• 33:50
  SPEAKER_00: If a product's liability standard arrives, the question of who pays when something goes wrong moves from a footnote to the main event.
• 33:58
  SPEAKER_00: Teams that have already mapped their indemnities will negotiate from knowledge instead of surprise.
• 34:04
  SPEAKER_00: And one more thing, in fairness, none of this requires you to predict the outcome of any bill.
• 34:10
  SPEAKER_00: These are low regret moves.
• 34:12
  SPEAKER_00: An inventory, a contract review, and a testing file are useful whether Congress acts quickly, slowly, or not at all.
• 34:19
  SPEAKER_01: That is a strong way to put it.
• 34:22
  SPEAKER_01: And it ends the political story on a practical note, which leaves us with story four, where the practical sakes might be the highest of the day.
• 34:31
  SPEAKER_00: Yes.
• 34:32
  SPEAKER_00: Let me take a breath before we go there, because story four is the one with the biggest numbers.
• 34:37
  SPEAKER_01: Story four.
• 34:39
  SPEAKER_01: The headline in the edition is that OpenAI's unreleased cyber model built root exploits for 948 Linux kernel flaws in a single run that costs$1.2 million.
• 34:52
  SPEAKER_01: The edition opens with a strong sentence.
• 35:00
  SPEAKER_01: OpenAI reported that an autonomous pipeline powered by an unreleased cyber model built working root exploits for 948 distinct Linux kernel vulnerabilities in a single run, as detailed by Security Week.
• 35:31
  SPEAKER_00: So 948 distinct bugs each turned into a working path to that level of control.
• 35:37
  SPEAKER_00: That is a lot of keys to a lot of doors.
• 35:40
  SPEAKER_00: The numbers in the addition are detailed.
• 36:04
  SPEAKER_00: And local privilege escalation is the key phrase.
• 36:08
  SPEAKER_00: It means the attacker is already on the machine in some limited way, and the exploit lifts them to full control.
• 36:14
  SPEAKER_00: I will return to why that matters because the edition is careful to say these attacks required an existing foothold.
• 36:20
  SPEAKER_00: It is not remote magic from the Internet.
• 36:23
  SPEAKER_01: Now the cost, which the edition calls the headline for every security leader.
• 36:28
  SPEAKER_01: According to Security Week's account of OpenAI's research, the run cost$1,213,430 in API credit, roughly$1,279 per exploited vulnerability.
• 36:43
  SPEAKER_00: Think about that number.
• 36:45
  SPEAKER_00: About$1,200 for a working exploit of a serious kernel flaw.
• 36:50
  SPEAKER_00: If you have ever priced a penetration test or a specialist's week of work, you know that is startlingly low.
• 36:58
  SPEAKER_00: A skilled human researcher might spend days or weeks on a single one of these.
• 37:09
  SPEAKER_00: Fair, and I will own it as my opinion.
• 37:11
  SPEAKER_00: What the addition does say is the comparison to the market.
• 37:15
  SPEAKER_00: And it is worth reading.
• 37:16
  SPEAKER_00: Let me hold that for a moment, because there is a trend line first.
• 37:20
  SPEAKER_01: Yes, the trend.
• 37:21
  SPEAKER_01: An earlier scan in late July, powered by GPT-5.6 SOL, yielded 166 unique local privilege escalation vulnerabilities at a cost of$513,690.
• 37:36
  SPEAKER_01: Between the two scans, according to the edition, the yield rose more than fivefold, while the cost per exploit fell by more than half.
• 37:45
  SPEAKER_00: And that is a span of a few months, late July to mid-October.
• 37:49
  SPEAKER_00: So in one quarter, five times the output and less than half the unit cost.
• 37:54
  SPEAKER_00: If your risk model is updated once a year, you are working with a picture that is two or three generations out of date.
• 38:23
  SPEAKER_00: And the addition adds a caveat that I like.
• 38:26
  SPEAKER_00: Whatever one thinks of bounty prices as a valuation method, the ratio makes the strategic point.
• 38:33
  SPEAKER_00: Bounties are what a company pays someone to report a bug.
• 38:36
  SPEAKER_00: They are not what a bug is worth to everybody.
• 38:39
  SPEAKER_00: But the ratio tells you that automated vulnerability research is now dramatically cheaper than the human labor market it is competing with.
• 39:23
  SPEAKER_00: I think that is reasonable.
• 39:25
  SPEAKER_00: And the addition itself raises the same asymmetry later.
• 39:29
  SPEAKER_01: Let me come to that.
• 39:30
  SPEAKER_01: First, the testing conditions, because the addition stresses that they were realistic.
• 39:35
  SPEAKER_01: The exploits were tested on virtual machines running an unmodified desktop Linux setup, with common protections such as KASLR, SMEP, and SMAP turned on, and the scan targeted the latest code in the kernel's master branch.
• 39:52
  SPEAKER_00: Those three acronyms are defensive features, mitigations that are supposed to make exploitation harder.
• 39:58
  SPEAKER_00: The point is that these were not tested against a deliberately weakened target.
• 40:02
  SPEAKER_00: The protections were on and the exploits still worked.
• 40:05
  SPEAKER_00: That matters because someone will inevitably say, well, that was a lab with the safety turned off.
• 40:11
  SPEAKER_00: The edition says it was not.
• 40:13
  SPEAKER_01: But the edition also notes the limits, citing the neuron, the attacks were developed in a controlled research environment, and they required an existing foothold on the target computer.
• 40:24
  SPEAKER_01: The agents also generated candidate patches for human review.
• 40:29
  SPEAKER_01: And OpenAI said its agents discovered and used at least 10 novel Linux exploitation techniques, with full technical details kept private for safety reasons.
• 40:42
  SPEAKER_00: This is not only an offensive story.
• 40:45
  SPEAKER_00: The same system that produced exploits also generated candidate patches for human review.
• 40:51
  SPEAKER_00: That is the defensive side of the ledger.
• 40:53
  SPEAKER_00: And it is the reason I do not read this as simply bad news.
• 40:57
  SPEAKER_00: A tool that finds and fixes at this speed is potentially a gift to maintainers, if they can absorb it.
• 41:04
  SPEAKER_01: Which brings us to the disclosure process.
• 41:07
  SPEAKER_01: And the edition describes it as deliberate.
• 41:10
  SPEAKER_01: OpenAI's security researchers checked the exploits, fixes, and proposed severities before anything went to kernel maintainers.
• 41:18
  SPEAKER_01: The company says it has submitted hundreds of reports and patches, with remaining fixes to be sent in stages through its Patch the Planet initiative.
• 41:27
  SPEAKER_00: That staging matters to maintainers.
• 41:30
  SPEAKER_00: Imagine being a small group of volunteers or a small team responsible for a critical piece of code, and a thousand reports arrive on the same day.
• 41:39
  SPEAKER_00: You would drown.
• 41:40
  SPEAKER_00: Sending them in stages with verified severities and proposed fixes is a way of respecting the humans who have to review them.
• 41:48
  SPEAKER_00: And there is a detail the edition says should not be overlooked.
• 41:52
  SPEAKER_00: OpenAI said that within 10 days, while the scan was running, over 100 of the vulnerabilities had been patched or had staging patches from other contributors, with about 300 more being publicly discussed.
• 42:06
  SPEAKER_00: The edition's conclusion is that many of these bugs were findable, and some were being found by others at the same time.
• 42:14
  SPEAKER_00: This is my favorite detail because it punctures a comforting story.
• 42:18
  SPEAKER_00: The comforting story is that these were deep secrets only a superhuman system could find.
• 42:24
  SPEAKER_00: Instead, a large share were already on the radar of other researchers.
• 42:28
  SPEAKER_00: That tells me two things.
• 42:29
  SPEAKER_00: The bugs are discoverable by many parties, and the race between finders is already on, whether or not any of us is paying attention.
• 42:38
  SPEAKER_01: Now, what the edition calls the most consequential sentence in the research, the defensive guidance.
• 42:43
  SPEAKER_01: OpenAI advised that for the near and medium term, a low-privileged account on an operating system that is not fully patched should be considered functionally equivalent to root access for a well-resourced attacker.
• 42:56
  SPEAKER_01: It pointed to sandboxing and isolation technologies such as Firecracker and G-visor.
• 43:02
  SPEAKER_00: Read that sentence again slowly because it undoes an old habit.
• 43:06
  SPEAKER_00: For decades, the security model on a shared system has been layers.
• 43:10
  SPEAKER_00: Give people limited accounts, keep the powerful account protected, and a breach of a limited account is contained.
• 43:16
  SPEAKER_00: OpenAI is saying against a capable attacker and an unpatched system, that containment is a fiction.
• 43:22
  SPEAKER_00: A foothold is close to full compromise.
• 43:24
  SPEAKER_00: In other words, as the edition puts it, the old assumption that limited user access contains the damage of a breach, no longer holds against capable adversaries.
• 43:36
  SPEAKER_00: And notice the qualifiers.
• 43:42
  SPEAKER_01: The guidance is not that every system is doomed.
• 43:45
  SPEAKER_01: It is that the patch level and the isolation layers now carry a lot more of the load.
• 43:51
  SPEAKER_00: Exactly.
• 43:52
  SPEAKER_00: Which is why they point to Firecracker and Gvisor.
• 43:56
  SPEAKER_00: Those are ways to put untrusted work inside a boundary that does not.
• 44:00
  SPEAKER_00: Depend on a normal user account staying humble.
• 44:03
  SPEAKER_00: Think of the difference between locking a door inside a building and putting the risky activity in a separate building entirely.
• 44:10
  SPEAKER_01: The edition closes the analysis with the asymmetry, and I find it sobering.
• 44:14
  SPEAKER_01: The same capability that lets a defender sweep a kernel for a little over a million dollars will eventually be available to attackers, whether through leaked models, open weight alternatives, or simple persistence.
• 44:27
  SPEAKER_01: The window in which defenders hold the advantage depends on how quickly patches move from discovery to deployment.
• 44:35
  SPEAKER_00: And I want to stress the word window.
• 44:37
  SPEAKER_00: It means there is a period of advantage, and it will close.
• 44:41
  SPEAKER_00: The defenders who use that period to shorten their patch cycles will be in a good place.
• 44:45
  SPEAKER_00: The ones who treat this as a news item will find out the hard way what the window was for.
• 44:51
  SPEAKER_01: Here is the strategic takeaway addressed to CISOs, security architects, and infrastructure leaders.
• 44:58
  SPEAKER_01: Rewrite your threat model so that any foothold on an unpatched system is treated as full compromise.
• 45:04
  SPEAKER_01: Shorten kernel and operating system patch cycles, prioritize isolation layers such as micro VMs and sandboxed runtimes for untrusted workloads, and budget for the coming surge in disclosed vulnerabilities.
• 45:18
  SPEAKER_01: Automated exploit discovery at roughly$1,279 per flaw means the volume of patches your teams must absorb is about to rise sharply.
• 45:35
  SPEAKER_00: A surge in disclosed vulnerabilities is a workload problem.
• 45:39
  SPEAKER_00: Patching is not free.
• 45:40
  SPEAKER_00: It means testing, scheduling downtime, coordinating with owners.
• 45:44
  SPEAKER_00: If the number of fixes doubles or triples and your team stays the same size, something gives.
• 45:50
  SPEAKER_00: And what usually gives is the speed.
• 45:53
  SPEAKER_00: So the practical question for a security leader is: what is your actual time from patch release to patch deployed?
• 46:01
  SPEAKER_00: Not the policy, the real number.
• 46:04
  SPEAKER_00: That is the metric I would take to my next leadership meeting.
• 46:07
  SPEAKER_00: If it is measured in weeks or months, that is the gap.
• 46:12
  SPEAKER_00: And for the non-security executive listening, the question to ask your security team is simple.
• 46:18
  SPEAKER_00: If an attacker got a low-level account on one of our servers today, what stops them from owning it?
• 46:24
  SPEAKER_00: That conversation is going to be different after this research.
• 46:27
  SPEAKER_01: Let's connect the four stories using the edition's bottom line.
• 46:31
  SPEAKER_01: Taken together, it says, these stories describe an AI industry being asked from several directions at once to show its work.
• 46:39
  SPEAKER_01: Google is not simply announcing a data center.
• 46:42
  SPEAKER_01: It is publishing a power plan, a cost pledge, and a community council, because permission to build now depends on them.
• 46:49
  SPEAKER_00: And Uber is not simply praising agents, it is reporting how few transactions they actually generate.
• 46:56
  SPEAKER_00: And Congress is preparing to ask what happens when AI systems cause harm.
• 47:01
  SPEAKER_00: So you have a power company question, a customer behavior question, and a liability question, all arriving in the same news cycle.
• 47:10
  SPEAKER_00: Then the edition says, the open AI research adds the sharpest edge.
• 47:14
  SPEAKER_00: Capability is compounding quickly, and the cost of doing serious work with AI, whether finding vulnerabilities or anything else, is falling faster than most organizations can adapt.
• 47:26
  SPEAKER_00: That is the real executive challenge of this moment.
• 47:29
  SPEAKER_01: In the edition's words, the technology's power is growing on one curve, while the evidence of value in everyday customer behavior and the rules of accountability are developing on much slower ones.
• 47:41
  SPEAKER_00: Two curves.
• 47:42
  SPEAKER_00: I think that is the cleanest way to hold it.
• 47:45
  SPEAKER_00: Capability is on a steep curve, and the cost of using it is falling.
• 47:49
  SPEAKER_00: Evidence of value and rules of accountability are on shallow curves.
• 47:53
  SPEAKER_00: The danger lives in the gap.
• 47:55
  SPEAKER_00: If you organize your company around the steep curve alone, you overinvest in things like consumer agents.
• 48:01
  SPEAKER_00: If you organize around the slow curve alone, you get blindsided by something like story four.
• 48:06
  SPEAKER_00: The edition also points out where the technology is already delivering measurable results, which it names as security, infrastructure, and internal workflows.
• 48:18
  SPEAKER_00: And it names where the evidence is thin, which is consumer adoption stories that the data does not yet support.
• 48:25
  SPEAKER_01: So the advice is to move fast where AI delivers measurable results, stay skeptical where the data does not support the story, and prepare now for a regulatory environment that is about to become far more demanding.
• 48:39
  SPEAKER_00: My one addition is that these are not three separate postures.
• 48:43
  SPEAKER_00: They are one habit.
• 48:44
  SPEAKER_00: Ask for the evidence.
• 48:45
  SPEAKER_00: The evidence in Wyoming is a contracted power plan.
• 48:48
  SPEAKER_00: The evidence in the Uber story is a transaction count.
• 48:51
  SPEAKER_00: The evidence in the legislative story is a testing file.
• 48:54
  SPEAKER_00: And the evidence in the security story is a patch time.
• 48:57
  SPEAKER_00: Same question, four settings.
• 49:00
  SPEAKER_01: That is the common thread, and I will put it as my own reasoning.
• 49:04
  SPEAKER_01: Every one of these stories rewards the organization that can show its work before somebody asks for it.
• 49:11
  SPEAKER_01: Which brings us to what to do about it this week.
• 49:14
  SPEAKER_01: What to do this week?
• 49:15
  SPEAKER_01: Action one comes from story one.
• 49:18
  SPEAKER_01: Ask your cloud and co-location providers where the power for your future capacity will come from.
• 49:23
  SPEAKER_01: Have your infrastructure strategy team put the question in writing and ask whether it is contracted or merely planned, and who bears the cost of grid upgrades.
• 49:33
  SPEAKER_00: Action two comes from story two.
• 49:36
  SPEAKER_00: Have your digital and e-commerce teams count how many orders actually arrive through AI assistance and compare the time an agent saves against your current checkout.
• 49:46
  SPEAKER_00: That is a one-week analytics task, and it tells you whether to invest or just keep an inexpensive option open.
• 49:53
  SPEAKER_01: Action three comes from story three.
• 49:56
  SPEAKER_01: Ask your legal and compliance teams to start an inventory of every customer-facing AI system, flagging any exposure to minors, and to review vendor contracts for how risk would be allocated under a product's liability standard.
• 50:11
  SPEAKER_00: And action four comes from story four.
• 50:14
  SPEAKER_00: Ask your security team for the real time, from patch release to patch deployed, and have them walk through what happens if an attacker gets a low-privilege account on an unpatched server.
• 50:25
  SPEAKER_00: Treat that foothold as full compromise, and look at isolation for untrusted workloads.
• 50:31
  SPEAKER_00: What to watch?
• 50:32
  SPEAKER_01: The edition does not publish a formal watch list, so these are the next tests we see implied by its four stories.
• 50:39
  SPEAKER_01: First, whether Google's Wyoming plan, including the Community Council and the Power Pledge, is followed by similar commitments from other hyperscalers.
• 50:49
  SPEAKER_01: Second, whether the agent order numbers at Uber, DoorDash, and others rise from a few hundred to something material.
• 50:58
  SPEAKER_00: Third, whether the new Congress moves the Guard Act and the AI Lead Act in January, and how hearings treat liability and child safety.
• 51:06
  SPEAKER_00: And fourth, how quickly the remaining Linux kernel fixes arrive through Patch the Planet, and how fast defenders shorten their own patch cycles while the window is open.
• 51:16
  SPEAKER_01: That is the Friday edition of the DX Today Executive Briefing for November 6, 2026.
• 51:22
  SPEAKER_01: Our next episode is on Tuesday.
• 51:24
  SPEAKER_01: Thank you for spending the time with us.
• 51:26
  SPEAKER_01: I'm Elena.
• 51:27
  SPEAKER_00: And I'm Marcus.
• 51:28
  SPEAKER_00: Thanks for listening to the DX Today Executive Briefing.
• 51:30
  SPEAKER_00: And we'll see you Tuesday.