WSJ Tech News Briefing: AI Boom Creates Blind Spot in Big Tech Accounting

Date: February 17, 2026
Host: Julie Chang (A)
Guests:

• Jonathan Weil, Heard on the Street columnist, WSJ (B)
• Erik Brynjolfsson, Professor, Stanford Institute for Human-Centered AI & Co-founder, Workhelix (E)
• Wendy Bounds, WSJ Leadership Institute (D, interviewer)

Episode Overview

This episode explores how the ongoing boom in artificial intelligence (AI) is fundamentally complicating the financial transparency of major tech companies, especially regarding how capital investments and depreciation expenses are reported. It also delves into how AI is reshaping the labor market, featuring insights from Stanford’s Erik Brynjolfsson on AI's impact on hiring, job augmentation vs. automation, and why focusing on “augmenting” rather than “replacing” human workers is crucial.

Key Discussion Points & Insights

1. The Challenge of Accounting for AI-Driven Capital Spending

• Major Tech Spending: Alphabet, Amazon, Meta, Microsoft, and Oracle are expected to spend a combined $3 trillion on property and equipment in four years, largely data center infrastructure for AI (00:35).
• Investor Blind Spots: The rapid growth generates soaring depreciation expenses, but investors can’t see these clearly on income statements because there’s no standardized disclosure.

Depreciation Expenses Explained

• Definition (01:29):

  "Depreciation is the expense that companies report as part of their earnings when they go out and build huge plants or they go out and buy lots of equipment... they gradually write down the cost of that over time."
  — Jonathan Weil (01:29)

• Investors struggle to see how depreciation is allocated (cost of revenue, R&D, SG&A), making it hard to project future costs and judge profitability.

Comparisons with Other Industries

• Transparency Comparison (02:45):

  "There are exceptions... railroads, a classic asset heavy industry. They will tell you the depreciation expense on the face of the income statement."
  — Jonathan Weil (02:45)

• Most tech companies bury depreciation in broad expense categories, unlike some industries, making analysis challenging.

Forthcoming Financial Disclosure Rule Changes

• GAAP Rule Change (03:20):
  New rules will disaggregate income statement expense categories into five buckets, improving disclosure of depreciation's impact—but not until financial statements filed in 2028.
• Company Adoption Lag: Big tech companies have not signaled willingness to implement these changes early.

Implications for the "AI Bubble" Narrative

• Market Frothiness (04:25):

  "What it does, it gives us insight into blind spots that would keep investors from being able to evaluate the companies and how frothy things may be."
  — Jonathan Weil (04:25)

• Lack of transparency hampers investors’ ability to validate or refute claims of a bubble.

2. AI’s Influence on the Labor Market: From the AI Summit

Highlight Interview: Wendy Bounds (WSJ) with Erik Brynjolfsson (Stanford/Workhelix)

The Power Law of AI Adoption Within Companies

• Super Users Drive Results (06:51):

  "In basically every company... there are a few like super users that are doing great and then there's this long tail of the typical users."
  — Erik Brynjolfsson (07:13)

• Identifying what super users do can help "level up" others through company-specific templates and continual peer learning.

Automation vs. Augmentation – Entry-Level Worker Risks

• Declines for Young, Exposed Workers (08:40): By analyzing ADP payroll data:

  • Young workers (22–26) in most AI-exposed jobs saw employment decline by ~16%.
  • Older/more experienced workers in the same fields less impacted.

• Automation vs. Augmentation:

  • Workers using AI to automate tasks experience sharper declines in employment.
  • Workers leveraging AI to augment (extend their skills) see job growth—though this group is a minority.

• Notable Quote (09:25):

  "The employment growth was only in that group that was using the LLMs to augment what they were doing."
  — Erik Brynjolfsson

The “Turing Trap”: Why Replicating Human Tasks is Limiting

• The Trap Explained (10:21):

  "Now I think [the Turing Test] is a terrible idea. It is not what you want AI to be doing, and it leads to a trap... We want the machines to be able to enable us to do new things that we never could have done before."
  — Erik Brynjolfsson

• Societal Implications:

  • Aiming to replace humans with AI reduces wages and discourages employee innovation.
  • Augmenting human capability ("centaur benchmarks") encourages prosperity, shared gains, and uplifts both employees and organizations.

• Advice for Industry:

  "Think about ways we can use AI to extend what we can do instead of having these black box benchmarks... I want them to use what I call Centaur benchmarks, half human, half machine..."
  — Erik Brynjolfsson (11:26)

Notable Quotes & Memorable Moments

| Timestamp | Speaker | Quote | |-----------|----------------|------------------------------------------------------------------------------------------------------------| | 01:29 | Jonathan Weil | "Depreciation is the expense that companies report as part of their earnings when they go out and build huge plants... and then they gradually write down the cost of that over time." | | 02:45 | Jonathan Weil | "Railroads, a classic asset heavy industry. They will tell you the depreciation expense on the face of the income statement." | | 04:25 | Jonathan Weil | "It gives us insight into blind spots that would keep investors from being able to evaluate the companies and how frothy things may be." | | 07:13 | Erik Brynjolfsson | "There are a few like super users that are doing great and then there's this long tail of the typical users." | | 09:25 | Erik Brynjolfsson | "The employment growth was only in that group that was using the LLMs to augment what they were doing." | | 10:21 | Erik Brynjolfsson | "Now I think [the Turing Test] is a terrible idea. It is not what you want AI to be doing, and it leads to a trap..." | | 11:26 | Erik Brynjolfsson | "I want them to use what I call Centaur benchmarks, half human, half machine, where they look at how well the team can do to solve some new problem." |

Timestamps for Key Segments

• 00:35 — Introduction to capital spending surge in Big Tech and resulting accounting issues
• 01:29 — What is depreciation? (Jonathan Weil explains)
• 02:45 — How other industries disclose depreciation
• 03:20 — Upcoming GAAP rule changes and timeframe
• 04:25 — Depreciation opacity and implications for "AI bubble" concerns
• 06:08 — Segment shift: how AI is changing the workplace
• 06:51 — Wendy Bounds and Erik Brynjolfsson: "Power law adoption" and getting more employees on board
• 08:40 — Data on AI's labor market impact, especially for young/entry-level workers
• 10:15 — Explaining the "Turing Trap" and risks of focusing solely on human imitation
• 11:26 — Brynjolfsson's advocacy for "centaur benchmarks" in AI evaluation

Episode Tone & Takeaways

The episode blends analytical, investigative reporting on opaque corporate finances with forward-looking, research-driven insight on technology’s labor impacts. Jonathan Weil’s clarity about accounting’s limitations is illuminated by Brynjolfsson’s practical optimism for how to harness AI for mutual human and organizational benefit—if applied thoughtfully.

For further exploration: The episode references the full conversation with Erik Brynjolfsson and more resources in the show notes.