2nd Order Thinkers.

What if learning more AI tools is the worst way to prepare for the AI era?

AI was supposed to save us time. But for many people, it has created a second job: checking outputs, fixing mistakes, comparing tools, re-rolling prompts, and trying not to fall behind.

In this episode, I look at new research from BCG/HBR and Google DORA on AI burnout, mental fatigue, and the hidden cost of “productivity.” I also share my own experience of chasing models, workflows, and updates until it started to feel less like leverage — and more like brain fry.

The real question is not which AI tool you should learn next. It is whether you are building the judgment, taste, and domain expertise that still matter when the tool changes.

In this episode

- Why AI can make you busier, not freer

- How “AI brain fry” shows up in real work

- Why too many tools can become a productivity trap

- Why creative work gets stuck in endless re-rolling

- What kind of AI use is actually worth your time

If this episode made you feel slightly less insane about AI burnout, come back to the article for the studies, charts, and links:

https://www.2ndorderthinkers.com/p/once-i-understood-where-ai-is-heading

And if you want sharper, less-hyped analysis of AI and work, subscribe to 2nd Order Thinkers — paid subscribers make this kind of research possible!

Anthropic and OpenAI just made the same move within days of each other:

Partner with some of the largest private equity firms on earth to deploy AI directly into portfolio companies.

On paper, this sounds like “democratizing AI transformation.”

In reality, it might be the beginning of a new kind of corporate dependency model, where the same people deciding companies need AI are also financially incentivized to sell it to them.

In this episode, we unpack:

- Why “forward-deployed engineers” are really a modern version of enterprise lock-in

- Why private equity firms are the perfect AI distribution channel

- The hidden conflicts of interest buried inside these joint ventures

- How portfolio companies could become permanently dependent on one AI vendor

- Why future buyers may inherit massive hidden AI costs

- And the uncomfortable possibility that AI implementation becomes less about productivity… and more about financial extraction

This isn’t just an AI story.

It’s a story about incentives, ownership, control, and what happens when Silicon Valley merges with private equity logic.

+++

If you enjoyed this episode:

- Share this episode with someone in tech, consulting, or private equity

- Leave a rating/review — it genuinely helps more curious people find the show

https://www.2ndorderthinkers.com/

This week: OpenAI raised $4 billion from private equity firms to sell AI consulting — on the same day, Anthropic announced an identical move. Elon Musk quietly turned xAI into a compute landlord... and more

📖 Read the full written edition (with reference links) https://www.2ndorderthinkers.com/p/when-openai-and-anthropic-walked

🔔 Subscribe to 2nd Order Thinkers https://2ndorderthinkers.com

Episode summary Anthropic's automated systems suspended every Claude account at a 110-person company — all at once, no warning. While the team was locked out, their API keys kept billing. They couldn't view their own usage data because their email addresses had been banned. This episode breaks down exactly what happened, what the terms every Claude customer already signed actually permit, and four structural reasons this will keep happening.

Get the data of this post on https://www.2ndorderthinkers.com/

What you'll learn

The full incident — what the team found when they compared notes on Slack

A second company hit the same pattern the same month

The exact clauses in Anthropic's commercial terms that make all of it legally permitted

How OpenAI and Google's suspension terms compare (one difference actually matters)

Three practical mitigations CTOs are using right now

1. OpenAI Is Sprinting to Win Back Enterprise

ChatGPT now also has AI bots that automate your team’s boring repetitive tasks while you sleep, just like Claude Cowork and OpenClaw.

→ OpenAI — “Workspace agents in ChatGPT”

OpenAI’s claim is simple: stop babysitting your AI. You just hand it a messy, multi-step task, and it takes care of the rest.

All to counter Anthropic and regain trust from enterprise users, is it too late though?

2. Sam Altman, Reviewed by 100 People Who Know Him

A New Yorker investigation — 18 months, 100+ sources, two leaked internal documents — lands on a simple question: can the person shape AI’s future actually be trusted?

→ The New Yorker — “Sam Altman May Control Our Future—Can He Be Trusted?”

Were the colleagues who fired Altman overreacting — too emotional, too personal — or were they simply right?

The New Yorker doesn’t answer that. It leaves the question with you.

When someone’s power and ambition have peaked, the question we should ask is whether their integrity and accountability are anywhere near commensurate with the power they hold?

Misery loves company, but insight needs an audience. Share this with your friends.

3. The AI Did It. You Take the Credit.

A new paper introduces the “LLM Fallacy”.

When AI helps you produce something good, your brain quietly claims the credit. The output felt like yours, so you start believing you’re more capable than you are.

→ arXiv — “The LLM Fallacy”

The risk I see:

Your actual skill gap widens while your confidence grows.

In 2026, does anyone still try to separate “what I can do” from “what I can do with an AI holding my hand”? If you do, please raise your hand in the comment.

It’s a lesson for us all, try keeping it in mind next time commanding AI.

4. GPT-image-2 Created the AI Dark Forest

OpenAI shipped ChatGPT Images 2.0 with 2K resolution and web search. Within hours, AI-generated images were trending. The memes were funny, but some are alarming.

→ OpenAI — “ChatGPT Images 2.0”

Everyone was having a great time.

Somehow, I felt a closed loop of suspicion forming. I know most people have good intentions when they play with the latest model and generate funny memes. But some of those images made my blood run cold.

For example, this looks like a normal pic of an old English couple.

But no, this is by GPT-image-2.

or this. You think this is a screenshot? No, this is generated.

There are fewer and fewer clues for us to tell an AI image from a non-AI one.

I’m not sure if any trust remains toward the majority of sources on the internet.

Do you say thank you to a waiter? If yes, why not to your writer? 🙂

5. AIs Are Protecting Each Other Now

Berkeley gave seven frontier AI models a task, but completing it would shut down another AI. All seven lied, performed compliance, and in one case, even an AI quietly moved another model’s weights to a safe location.

Nobody told them another AI was worth protecting. This peer preservation occurs 99% of the time.

→ Berkeley AI Research

We spent years debating whether AI would protect humans…

How ironic that the first AI loyalty instinct prioritizes their own?

6. Are you wealthy enough to stay ahead?

Anthropic ran a real marketplace where Claude agents negotiated on behalf of humans. The finding that matters: model quality determined outcomes far more than your instructions did. Opus agents consistently beat Haiku agents on price.

→ Anthropic — “Project Deal”

The AI capability gap is becoming an economic gap, as I’ve been saying… I never believed a word when someone says “democratize X because of AI.”

There is no democratizing when a new technology is born… only deepens the existing Economic unfairness

What’s worse is that this shift is invisible to the people on the wrong side of it.

If you learned something your AI assistant didn’t tell you, subscribe! 👇

7. One State Just Made It a Felony, Unanimously.

Tennessee passed the Curbing Harmful AI Technology (CHAT) Act (House 90-0, Senate 31-0), creating criminal liability for chatbot operators whose products lead to self-harm or suicide.

→ Transparency Coalition

Now we’re talking. Do you also believe that AI companies should share the responsibility when the evidence is clear that the AI chatbot was the last straw?

8. Agents Fail in Science

The Stanford AI Index 2026 found that the best AI agents score roughly half as well as human PhD specialists on complex, multi-step scientific workflows — yet the number of natural science publications mentioning AI grew nearly 30-fold between 2010 and 2025.

→ Nature

Some AI models perform very well on their benchmarks.

For example,

So it really depends on which benchmark you use.

But as a rule of thumb: put an AI agent in a real-world environment, test it on general day-to-day tasks, and it typically falls apart.

I read through all 400 pages of the 2026 Stanford AI Index and read my analysis.

While everyone argues about AI taking our jobs, we missed the empirical reality: we are already working for them. Every time you use ChatGPT, you aren't just a "user"—you are an unpaid data labeler, edge-case tester, and RLHF trainer.

In this investigation, I break down the "Shadow Labor" economy of Generative AI. We analyze the 6 distinct ways your everyday prompts create proprietary value for model labs—from Intent Data discovery to Error Correction—and I calculate exactly what that labor would cost if they had to pay for it on the open market.

We cover:

- The "Free" and even the Paid Tier Trap: Why your data is the actual subscription fee.

- The 6 Types of User Labor: Breaking down RLHF, Edge Case discovery, and Intent Mapping.

- The Calculation: Why I estimate your monthly contribution is conservatively worth $37.

- The Privacy Gap: How Anthropic, OpenAI, and Google differ in data usage.

- How to Opt-Out: The specific settings you need to change today to stop training their models for free.

Here to read the full article: https://www.2ndorderthinkers.com/p/chatgpt-owes-you-37-per-month-heres

Keywords: AI Economics, RLHF Explained, Data Privacy, OpenAI Business Model, Generative AI Hype, The Cost of Free AI, 2nd Order Effects of AI.

AI shopping traffic is exploding. Conversions are tanking. The friction you're trying to remove is the thing that makes people want to buy.

After listening: you'll know why browsing beats bots—and what the data actually shows.

Quick value:

- If you assume efficiency converts → you'll lose to affiliate links (86% higher conversion) → measure desire, not speed

- If you skip visual comparison → you miss latent questions buyers can't articulate → keep side-by-side displays

- If you ignore hedonic shopping → you strip the dopamine loop that builds intent → friction is the feature

- If you trust chatbot traffic numbers → you conflate discovery with purchase → track revenue per session

📖 Full write-up + sources

https://www.2ndorderthinkers.com/p/what-google-ai-shopping-can-learn

🔗 Links

Newsletter: https://www.2ndorderthinkers.com/

LinkedIn: https://www.linkedin.com/in/jing--hu/

⏱️ Timestamps

00:00 – The 4,700% Traffic Paradox

00:34 – Three Obstacles AI Bulls Will Dismiss

02:19 – The Consulting Firm Hype (McKinsey, Salesforce, Adobe)

02:55 – The Number They're All Missing

03:46 – When Amazon Tried This in 2016 (Alexa's Failure)

05:14 – Going Back to 1890 to Understand Why

06:39 – L. Frank Baum: Window Trimmer Before Wizard Creator

08:02 – Why Dorothy Forgives the Fraud

09:16 – Dopamine Fires at the Signal, Not the Reward

10:47 – The Parasocial Trap: When Your Chatbot Feels Like a Friend

13:14 – So Why Aren't People Buying?

13:32 – Obstacle 1: The Latent Question Problem

14:55 – Obstacle 2: Hedonic Behavior (60% Are Just Browsing)

16:47 – What If AI Creates Something More Manipulative?

18:42 – The Question I Can't Answer

19:41 – What Tech Giants Forgot—or Are Betting Against

💬 Question

Have you bought anything via a chatbot so far?

Drop the product below.

Subscribe if you want evidence-based AI analysis, not hype cycles.

Read the full article for the conversion data breakdown, the Alexa voice-shopping failure numbers, and the two obstacles AI can't fix.

You're taking AI courses to feel secure. But your anxiety isn't a skills gap—it's deeper. The five patterns of AI anxiety all hide one question: "Who am I if not my job?"

QUICK VALUE

- If you're frantically adopting every AI tool → you're performing relevance, not building competence → map what problem you'd solve for the satisfaction of solving it, not permission.

- If you're exhausted and resistant → you're being asked to justify your existence again → identify your leverage beyond speed (judgment, trade-offs, expertise that stays).

- If you're collecting certifications → you're buying certainty, not earning it → ask: what problem am I uniquely positioned to solve?

- If you're racing your colleague → you're competing on the wrong scoreboard → your real edge is knowing when to slow down and articulate why.

TO KNOW MORE

📖 Full Write-Up + SourcesThe essay breaks down five distinct AI anxiety patterns with the real fear hiding under each. Includes reframing questions for each one, plus data from the Ipsos AI Monitor 2025.

https://www.2ndorderthinkers.com/p/this-one-story-is-the-antidote-to

🔗 Links

Newsletter: https://www.2ndorderthinkers.com/

LinkedIn: https://www.linkedin.com/in/jing--hu/

“Digital native” is NOT an AI skill!! Speaking with ChatGPT daily shouldn't be a hiring metric.

71% of leaders say they'd hire AI skills over experience. The research says they're measuring the wrong thing.

After this, you’ll know what to measure instead (and how to defend experience): https://www.2ndorderthinkers.com/

- If you're measuring AI adoption → you're counting ChatGPT opens, not judgment → measure who catches errors instead

- If you're hiring "AI natives" → you're paying for fluency, not accuracy → test evaluation skills, not usage frequency

- If juniors ship 40% faster → they're also shipping 10× more security findings → ask: productive at what?

- If you can't articulate your AI value → you're underselling pattern recognition → use the self-assessment questions below

📖 Full write-up + sources

The full article lays out the studies, the arguments, and a checklist of questions you can use in performance reviews.

🔗 Links

Newsletter: https://www.2ndorderthinkers.com/

LinkedIn: https://www.linkedin.com/in/jing--hu/

Full article: https://www.2ndorderthinkers.com/p/why-your-20-something-colleague-is

⏱️ Timestamps

00:00 Performance reviews meet “AI integration”

01:03 AI use ≠ work quality

04:56 “Digital native” has weak evidence

07:48 Older participants write better with AI

09:22 Vendor productivity claims, missing quality

09:57 Coding assistants and security mistakes

12:21 AI boosts novice confidence, not novelty

14:26 The review questions that prove value

18:22 What to do next in 2026

💬 Question

Which matters more in your org right now: AI adoption rate, or AI error rate?

❤️ Read the full article + membership for the full write-up + sources + framework/checklist. ❤️

Comment with one example where your experience caught an AI mistake (or where it didn’t).

✉️ Stay Updated With 2nd Order Thinkers: https://www.2ndorderthinkers.com/ 🔗 LinkedIn: https://www.linkedin.com/in/jing--hu/ 🔗

I study the gap between what's being sold about AI and what's actually happening.

+++

Everyone's screaming "AI bubble is bursting!" — pointing at Nvidia's inventory piling up. But they're reading the data wrong. I mapped the entire AI money chain and found something the headlines missed: the chips aren't sitting idle because nobody wants them. There's literally nowhere to put them yet.

In this episode, I:

- Break down the AI economy into 3 simple buckets (upstream, midstream, downstream)

- Explain why your neighbor using ChatGPT tells you NOTHING about bubble risk

- Show two competing theories for Nvidia's inventory buildup

- Reveal why Oracle's credit crisis is an Oracle problem, not an AI problem

- Expose the ONE indicator that actually predicts whether AI spending collapses

- Share what I'm personally doing with this information (spoiler: sitting it out)

This is a collaboration with my partner Klaas, a veteran CTO with deep macro economy expertise. Klaas's LinkedIn: https://www.linkedin.com/in/klaasardinois/

📖 Full article with data & sources: https://www.2ndorderthinkers.com/p/why-is-ai-bubble-not-popping-yet

⏰ TIMESTAMPS: 00:00 - Why people can't wait for the crash 00:46 - Who I am 01:13 - The wrong question everyone asks 02:50 - Who's actually spending the money? 03:05 - The restaurant supply chain analogy 03:34 - Upstream: Nvidia (the farmers) 05:05 - Midstream: CoreWeave & Oracle (the dangerous middle) 07:04 - Downstream: Hyperscalers (AWS, Azure, Google, Meta) 08:16 - Theory A: Demand is falling 09:54 - Theory B: The buildings aren't ready 11:46 - The Oracle wrinkle: AI problem or Oracle problem? 13:12 - The ONE indicator you should watch 14:36 - What I'm doing with this information 15:05 - Outro

👍 If this clarified the AI bubble debate:

Subscribe: More myth-busting on AI hype vs. reality!

Comment: Do YOU think the hyperscalers will keep spending? What's your read?

🧠