WEBVTT

1
00:00:10.000 --> 00:00:24.656
Welcome to the Global Advisors A.I Voices podcast. I'm Maya.
Today is Monday, the sixteenth of November 2026.
This is your morning catch-up on A.I. Please like and follow the show for your daily A.I update.

2
00:00:26.705 --> 00:00:46.140
A recent Capgemini survey found that fifty-three percent of senior executives cite A.I costs as a major concern. The report highlights a significant blind spot: many leaders struggle to confidently estimate indirect costs like human oversight, and few organizations calculate what they call 'A.I Yield.'

3
00:00:46.490 --> 00:01:08.186
To manage this unexpected spend, businesses are tightening controls and exploring multi-model strategies. Capgemini forecasts that the primary metric will shift from total expenditure against budget to first-pass success rate and cost per completed task. This moves the focus from simply tracking spend to measuring successful outcomes.

4
00:01:08.536 --> 00:01:27.160
Nate B Jones argues that companies should evaluate models based on the full cost of review and correction, not just the sticker price. He notes that if a cheaper model passes specific tests for the work being bought, it provides evidence for trading down, which changes who has pricing power in negotiations.

5
00:01:29.559 --> 00:01:50.210
While technical milestones continue, Nate B Jones argues that better A.I intelligence and broader adoption are not sufficient on their own to prove a financial case. He points to data showing A.I spending has been effectively flat for three straight months, raising questions about where value is actually accruing in the industry.

6
00:01:50.560 --> 00:02:12.874
This gap is echoed by Torsten Slok, an economist at Apollo, who states that A.I has delivered 'zero financial payoff' to corporate America outside of the tech industry despite massive investments. Jones suggests that customers are becoming better at identifying 'good enough' solutions, which may drive usage growth without a linear increase in costs.

7
00:02:13.224 --> 00:02:30.696
The source video notes that while frontier models account for a large share of spending, task saturation means cheaper models often meet business requirements. This dynamic is increasing price sensitivity, forcing a shift in focus from raw capability to value chain economics.

8
00:02:33.095 --> 00:02:56.114
Bankers are reportedly considering pricing Anthropic's IPO above the seventy to seventy-eight dollar range, potentially reaching a one point five trillion dollar valuation. The Financial Times, citing people familiar with the matter, notes that Anthropic's backers prioritize raising capital for long-term computing requirements over maximizing valuation.

9
00:02:56.464 --> 00:03:14.597
Demand appears strong, with Bloomberg reporting the offering was multiple times oversubscribed by large institutional orders. Seven tech giants have committed to purchase twenty-six billion dollars in Anthropic shares in a private placement alongside the IPO, according to the Motley Fool.

10
00:03:14.947 --> 00:03:36.067
FinanceFeeds commentary suggests that higher pricing increases proceeds but also raises the valuation that future revenue must justify. The Motley Fool reports Anthropic had an eighty-one billion dollar annualized revenue run rate as of October, highlighting the scale of growth required to support such a high market cap.

11
00:03:38.467 --> 00:03:54.744
On the MAD Podcast, Ryan Greenblatt argues that people currently hold power to a substantial degree because they have the ability to work. He warns that if human labor loses its value due to A.I, it could facilitate easier consolidation of power.

12
00:03:55.094 --> 00:04:11.798
Greenblatt compares this potential dynamic to how oil wealth enables dictatorial control in non-productive economies. He suggests it might be much easier for someone to consolidate power if the economic base is no longer broadly distributed through human labor.

13
00:04:12.148 --> 00:04:27.295
However, Greenblatt notes that it is unclear exactly how the devaluation of labor works. He argues it depends on whether people have intrinsic preferences to employ humans even if A.I could do the job much better, leaving the outcome uncertain.

14
00:04:29.694 --> 00:04:48.744
Nvidia has been a major player in the 'A.I circular economy' by providing infrastructure and funding for numerous A.I projects. The company accounts for roughly eighty percent of the A.I GPU market and has reported ninety-nine billion dollars in investments and twenty-five billion dollars in commitments.

15
00:04:49.094 --> 00:05:10.556
Jensen Huang resigned following reports of an affair with a former executive at a customer with alleged ties to Chinese intelligence. C F O Collette Kress was appointed as interim C E O. Rebecca Bellan says this appointment is a signal to manage expenses and stabilize the company rather than pursue aggressive expansion.

16
00:05:10.906 --> 00:05:29.551
Anthony Ha reports that analysts suggest if Nvidia enters a 'lockdown' phase or pulls back on investments due to leadership uncertainty, it could have dramatic repercussions for the A.I industry. The stakes are high because so much of the current economy is riding on Nvidia's stability.

17
00:05:31.950 --> 00:05:49.721
Indian chip startup Bharat Semi has raised twenty-four million dollars in seed funding to develop A.I-enabled image sensors. The round was backed by General Atomics, a United States defense technology company, and Kaynes Semicon, an Indian contract manufacturer.

18
00:05:50.071 --> 00:06:09.079
The startup focuses on high-reliability image sensors and sensing platforms ready for edge A.I, prioritizing compound semiconductors. National defense applications are a primary focus, and the company traces its roots to 3rdiTech, launched in 2019 at IIT Delhi.

19
00:06:09.429 --> 00:06:27.626
Bharat Semi says it has secured design wins with companies including General Atomics and is targeting its first commercial production run for 2027. However, Diego Almada Lopez notes that the product is still being built, indicating that commercial success is not yet guaranteed.

20
00:06:30.025 --> 00:06:45.769
Spotify reports that eighty percent of listeners return to the DJ feature weekly, establishing it as a popular Premium tool. To address user feedback, the update adds new English and Spanish voices, plus a toggle to turn off commentary between songs.

21
00:06:46.119 --> 00:07:00.050
The commentary toggle is designed for scenarios like dinner parties where users want a seamless soundtrack without chatter. These new features began rolling out to Premium users in existing DJ markets on November 10.

22
00:07:00.400 --> 00:07:19.877
WIRED highlights user frustration with A.I DJs, specifically mentioning YouTube Music's 'Beyond the Beat' and Spotify's variant as disruptive to the listening experience. A user comment in WIRED states they tried the Spotify A.I DJ variant once and never again, citing a dislike for talking DJs.

23
00:07:22.276 --> 00:07:39.684
A Carnegie Mellon student has developed tiny E Ink gadgets called OATs that give everyday objects expressive personalities to nudge users toward better habits. For example, the device provides visual feedback on a plant-shaped pet display when a user forgets to water a plant.

24
00:07:40.034 --> 00:08:01.176
Users can attach OATs to objects like dumbbells to track behaviors, using the A.I pet's disappointment as motivation for regular exercise. The creator, Kim, restricts OATs' interactions to expressive emotions on E Ink displays to mitigate risks of distorted moral judgments and over-trust associated with anthropomorphism.

25
00:08:01.526 --> 00:08:17.654
Kim argues that while anthropomorphism can help change behaviors, designers must be careful not to abuse the relationship between users and A.I. This approach aims to leverage emotional connection for habit formation while maintaining ethical boundaries.

26
00:08:20.053 --> 00:08:38.464
OpenAI has unveiled a new science blog to publish A.I-enabled research findings, methods, and researcher stories. The research team announced the launch on November 15, 2026, aiming to showcase fresh findings and first-hand narratives from scientists using A.I.

27
00:08:38.814 --> 00:08:55.198
The blog covers a broad range of scientific fields including chemistry, biology, physics, and climate science, without limiting scope to specific disciplines. Articles will include links to code repositories and datasets to support reproducibility.

28
00:08:55.548 --> 00:09:13.126
Notably, the editorial policy allows contributions using non-OpenAI models if the methodology is reproducible. This signals an openness to diverse tools in the scientific community, focusing on the validity of the research rather than the specific vendor used.

29
00:09:13.476 --> 00:09:29.924
That's A.I Voices for today. I'm Maya. Thanks for listening.
Follow the show on Spotify, YouTube, Apple Podcasts, Amazon Music, or your favourite podcast platform, and explore the original voices and sources in the episode notes. See you tomorrow.

