Sign in

Cassie Kozyrkov

@decisionleader.bsky.social
154 followers 2 following 457 posts

CEO, AI Advisor, Keynote Speaker, fmr Chief Decision Scientist at Google Newsletter: decision.substack.com

PostsRepliesMedia
Cassie Kozyrkov @decisionleader.bsky.social · 11/06/2026
“The LLM decided” is not an explanation. It's a gap where an explanation should be. Every automated outcome has a human lineage: someone chose the objective, set the constraints, and signed off on deployment.
100
Cassie Kozyrkov @decisionleader.bsky.social · 06/06/2026
These so called "AI layoffs" are most likely to happen at companies who are doing the least AI. 🧐 Here's my deep dive: Who's on team Medium? bit.ly/4g0zJzb Who's on team Substack? bit.ly/quaesita_lay...
010
Cassie Kozyrkov @decisionleader.bsky.social · 05/06/2026
AI is here to stay! Learn how to use it to make your decision-making better, not worse. My Decision-Making with ChatGPT course is running on demand and you'll have your opportunity to ask me anything you like during the 2h live Q&A on Monday. Enroll here: decisiongptcourse.com
100
Cassie Kozyrkov @decisionleader.bsky.social · 04/06/2026
AI sees what you did, not why you did it. It captures patterns in behavior, not the intent behind decisions. Patterns are useful. Purpose still belongs to humans. We’ll talk about this in the latest cohort of my course. Live Q&A is at 8 AM ET on Monday. Don’t miss it! decisiongptcourse.com
010
Cassie Kozyrkov @decisionleader.bsky.social · 03/06/2026
Last year, 4.5% of US layoffs were "due to AI" and now this year in March, it's 25%. In April, the number is 26%.* I smell a rat. 🐀 🐀 🐀 🐀 Here's my take on what's happening: bit.ly/quaesita_layoffs
140
Cassie Kozyrkov @decisionleader.bsky.social · 28/04/2026
The more powerful AI gets, the more valuable restraint becomes. All these new developments can sometimes trigger the wrong instinct: Automate all the things. Easy there 😄 Just because you can delegate to a machine doesn’t mean you’ve found leverage.
100
Cassie Kozyrkov @decisionleader.bsky.social · 24/04/2026
At Meta, tens of thousands of employees were ranked by how many tokens they burned. The top user logged hundreds of billions in a month. Across companies, similar patterns are emerging. Read the full piece here: bit.ly/quaesita_tok...
001
Cassie Kozyrkov @decisionleader.bsky.social · 17/04/2026
AI-first isn’t a tech mandate. It’s a decision-making posture. If something works well without AI, that’s often a feature, not a failure. If you want to get better at making those calls: decisiongptcourse.com 📅 April 21, Tuesday, 2 PM ET #DecisionIntelligence #AILeadership #StrategicThinking
000
Cassie Kozyrkov @decisionleader.bsky.social · 14/04/2026
Starting with AI is often how teams get it wrong. Start with the decision. Then the outcome. Then the constraints. Then the way of working. Then, with the tool. Sometimes that tool is AI. Sometimes it isn’t. We’ll talk about how to tell the difference: decisiongptcourse.com 📅 April 21, 2 PM ET
011
Cassie Kozyrkov @decisionleader.bsky.social · 14/04/2026
If seams tell you where a system breaks, objectives and incentives tell you why it breaks.
000
Cassie Kozyrkov @decisionleader.bsky.social · 10/04/2026
Anthropic is holding monster model Claude Mythos back from public release and sharing it instead with the organizations in its Project Glasswing security consortium. A lot of people seem to want this story to be one of 3 things:
210
Cassie Kozyrkov @decisionleader.bsky.social · 10/04/2026
How many of these 25 #AI #headlines did you catch recently? Tell me the one you find most interesting in the comments. For the curious, my newsletter digs into each one: decision.substack.com
100
Cassie Kozyrkov @decisionleader.bsky.social · 09/04/2026
Which of these true news stories were on your 2026 bingo card? ⚡️ Anthropic launched Mythos, a model too dangerous to release publicly.
100
Cassie Kozyrkov @decisionleader.bsky.social · 07/04/2026
"Yay!! Oh. Oh, no." ...is the themesong of agentic AI.
010
Cassie Kozyrkov @decisionleader.bsky.social · 06/04/2026
NVIDIA CEO Jensen Huang described AI as a five-layer cake. 🍰 Chips, models, applications, energy, data centers. A full stack. A complete system. A solved problem. Delicious. 🍴 Except there’s a layer missing: youtu.be/MOTEZqdzyRA
youtu.be
The Real Bottleneck for AI: Human Attention
NVIDIA CEO Jensen Huang described AI as a five-layer cake. 🍰 Chips, models, applications, energy, data centers. A full stack. A complete system. A solved problem. Delicious. 🍴 Except there’s a…
110
Cassie Kozyrkov @decisionleader.bsky.social · 31/03/2026
AI should serve people; yet many organizations start with the tool. When you start with tools, humans serve tools. When you start with decisions, tools serve humans. Automation after leadership, not instead of it. Join my live AMA: decisiongptcourse.com 📅 Apr 21, Tuesday, 2 PM ET #AILeadership
000
Cassie Kozyrkov @decisionleader.bsky.social · 29/03/2026
😠 😡 😤 What do you think of Claude Code's answer to a question I asked while in a Sunday morning vibe coding session? Agree? Disagree? Throw your phone at the wall in rage? ME 👩‍💻 ❯ How common is systems thinking among developers?
100
Cassie Kozyrkov @decisionleader.bsky.social · 26/03/2026
How many of these 16 #AI #headlines did you catch recently? Tell me the one you find most interesting in the comments. For the curious, my newsletter digs into each one: decision.substack.com 🗞️ 1 China is giving away apartments to anyone with a computer and an AI agent
100
Cassie Kozyrkov @decisionleader.bsky.social · 22/03/2026
My AI agents only helped me halfway with making this image, but halfway is better than no way. Who else is optimizing for time saved wherever they can get it? Always trying to find ways to cram more life into my life. How I'm doing things:
100
Cassie Kozyrkov @decisionleader.bsky.social · 10/03/2026
AI adoption is expensive. But so is avoiding it. Everyone wants the upside of AI. Fewer want to foot the bill. Avoiding that cost doesn't make it disappear, it just shows up later.🤷‍♀️ I'll be answering questions live in my Maven course's latest cohort: decisiongptcourse.com
000
Cassie Kozyrkov @decisionleader.bsky.social · 05/03/2026
If you don’t define the default action, you still have one. More insight, less noise → decision.substack.com #DecisionIntelligence #Governance #Leadership
010
Cassie Kozyrkov @decisionleader.bsky.social · 04/03/2026
How many of these 14 #AI #headlines did you catch recently? Tell me the one you find most interesting! 🗞️ 1 OpenAI secures Pentagon deal as Anthropic is cut off over AI limits 🗞️ 2 93% of developers use AI for code, but productivity gains stall at 10%
210
Cassie Kozyrkov @decisionleader.bsky.social · 03/03/2026
AI writes the code. The leverage is in supervision. Less time coding. More time specifying and verifying. Before code: make the AI restate what it thinks you want. After code: have a different AI explain what was built and how it could fail.
100
Cassie Kozyrkov @decisionleader.bsky.social · 27/02/2026
A hobbyist tried pairing his DJI Romo robot vacuum with a PlayStation controller. Instead, he accessed nearly 7,000 devices worldwide.... including live camera feeds inside homes. Augmented curiosity is a new category of risk.
110
Cassie Kozyrkov @decisionleader.bsky.social · 25/02/2026
AI makes answers cheap. That makes judgment expensive. More insight, less noise → decision.substack.com #DecisionIntelligence #AILeadership #FutureOfWork
000
Cassie Kozyrkov @decisionleader.bsky.social · 24/02/2026
Before asking for more data, ask: What would it take to change my mind? No commitment, no clarity. No clarity, no decision. Sharpen your decision-making skills here: decisiongptcourse.com #DecisionIntelligence #Leadership #AI
010
Cassie Kozyrkov @decisionleader.bsky.social · 19/02/2026
How many of these 16 #AI #headlines did you catch recently? For the curious, my newsletter digs into each one: decision.substack.com 🗞️ 1 Lunar New Year becomes launchpad for China’s AI price war 🗞️ 2 The AI cost-cutting trap: why “time saved” may be the wrong metric
100
Cassie Kozyrkov @decisionleader.bsky.social · 18/02/2026
Most people think the barrier to automation is technical skill. It isn't. The real barrier is knowing what you want badly enough to bother building it. AI is wasted on people who don't know how to want. Let me explain: Mhttps://bit.ly/quaesita_startvibing
000
Cassie Kozyrkov @decisionleader.bsky.social · 17/02/2026
One reassuring thing about teaching leaders to work with AI: They've been building AI skills their whole careers... without realizing it. Judgment, context, and institutional memory don’t disappear just because the tools change. They become more valuable. decision.substack.com
000
Cassie Kozyrkov @decisionleader.bsky.social · 12/02/2026
What’s the most valuable skill in the AI era? ⛔ It’s not coding. ⛔ It’s not prompt tricks. ⛔ It’s not “keeping up with the latest model.” 💪 It’s judgment. Whenever a new #AI capability drops, debate centers on how powerful it is. That's the wrong question. Right question: bit.ly/quaesita_spe...
000
Cassie Kozyrkov @decisionleader.bsky.social · 12/02/2026
A subtle shift I’m seeing in strong leaders: They’re less interested in what AI can do and more interested in what it shouldn’t. That instinct turns out to be a competitive advantage. This distinction is central to how I teach Decision-Making with ChatGPT. Enroll here 👉 decisiongptcourse.com
000
Cassie Kozyrkov @decisionleader.bsky.social · 10/02/2026
Change doesn’t happen because leaders commit harder. It happens when the way decisions are made changes. That’s the real leverage in an age of AI and agents: Decisions. Enroll here and learn more: decisiongptcourse.com
decisiongptcourse.com
Decision-Making with ChatGPT by Cassie Kozyrkov on Maven
Make better decisions in the Age of AI. Turn ChatGPT into your strategic thinking partner.
000
Cassie Kozyrkov @decisionleader.bsky.social · 10/02/2026
One thing AI failures expose quickly: implicit assumptions. With people, shared context fills the gaps. With machines, silence gets interpreted as permission. That difference explains a lot of “AI went wrong” stories. #StrategicThinking #ExecutiveAI #Leadership
100
Cassie Kozyrkov @decisionleader.bsky.social · 09/02/2026
Data doesn’t drive decisions. People do. This gap between AI capability and decision design is why I built Decision-Making with ChatGPT. Enroll here: decisiongptcourse.com #DecisionIntelligence #AILeadership #Leadership
000
Cassie Kozyrkov @decisionleader.bsky.social · 06/02/2026
People who were *excellent* at Google are often surprisingly bad at working with AI. Not because they’re dong anything wrong, but because search trained us to compress, and AI needs context. Search rewarded brevity. AI rewards context. Read more here: bit.ly/quaesita_sho...
bit.ly
Google Search Has Been Quietly Poisoning Your AI Skills
The habit that made you good at searching is making you terrible at prompting
100
Cassie Kozyrkov @decisionleader.bsky.social · 29/01/2026
Does AI training double worker productivity? What's your reaction to this study from Microsoft? Follow for more.
000
Cassie Kozyrkov @decisionleader.bsky.social · 22/01/2026
🐓🐓🐓 Calling all early birds! If you're awake enough to read this, then consider joining the rise and shine edition of my Decision-Making with ChatGPT course today. The live Q&A with me starts in just over 2h (8 AM ET today) and the rest is on demand. 🐓 Enroll here: decisiongptcourse.com
decisiongptcourse.com
Decision-Making with ChatGPT by Cassie Kozyrkov on Maven
Make better decisions in the Age of AI. Turn ChatGPT into your strategic thinking partner.
000
Cassie Kozyrkov @decisionleader.bsky.social · 21/01/2026
Alignment beats accuracy - every time. (That’s not a metric. It’s a leadership choice.) This is the core principle behind my Maven course, Decision-Making with ChatGPT. Cohorts 4 and 5, which both include a live Q&A with me, coming up on Jan 22 and Feb 12! Enroll here: decisiongptcourse.com
010
Cassie Kozyrkov @decisionleader.bsky.social · 20/01/2026
Most decision failures aren’t due to lack of data, they’re due to unclear judgment. Cohort 4 of Decision-Making with ChatGPT is now open: 👉 maven.com/cassie-kozyr... #StrategicThinking
000
Cassie Kozyrkov @decisionleader.bsky.social · 14/01/2026
So much wasted potential! Most people use ChatGPT for emails and summaries. Forget busywork, let's focus on the things that make a real impact. Enroll in my Decision-Making with ChatGPT course to use AI to move the needle on what matters most -- your decisions: decisionGPTcourse.com
000
Cassie Kozyrkov @decisionleader.bsky.social · 13/01/2026
Who's surprised by this finding from Gartner that CEOs have less faith in their C-Level colleagues being AI-savvy in 2025/26 than they had in those same colleagues being tech savvy in 2019/20?
000
Cassie Kozyrkov @decisionleader.bsky.social · 12/01/2026
The more powerful the tool, the more the human matters. AI can't replace what's most essential. Not everyone agrees. Do you? decisiongptcourse.com
000
Cassie Kozyrkov @decisionleader.bsky.social · 10/01/2026
Since several folks have complained to me recently about ChatGPT's sycophantic tone, I figured it's time for that PSA again: 🤯 YOU CAN CHANGE CHATGPT'S TONE. 🤯 I recommend "Efficient" - that's the one I use. Click, click, done.
021
Cassie Kozyrkov @decisionleader.bsky.social · 09/01/2026
Last year, 364 students in 56 cities learned from me on @Maven. Will you take one of my AI courses for leaders, aspiring leaders, and curious self-improvers this year? The next cohort of Decision-Making with ChatGPT starts on Jan 14! Enroll at decisiongptcourse.com
000
Cassie Kozyrkov @decisionleader.bsky.social · 06/01/2026
In case you missed it, NVIDIA just showed us the new Vera Rubin chip. 💪
000
Cassie Kozyrkov @decisionleader.bsky.social · 29/12/2025
Stop fighting yourself, starting winning at your self-improvement resolutions! I have a New Year's gift for you all: an enormous compendium of all the resolutions and motivation advice I’ve ever written. Now in illustrated form bit.ly/advancedmanual
000
Cassie Kozyrkov @decisionleader.bsky.social · 26/12/2025
It’s Dec 26. In five days, millions will set resolutions fueled by holiday guilt. By February, 80% will have failed. The mistake? Using willpower to solve a design problem. Willpower is finite; behavioral architecture is infinite. A thread on hacking your 2026. 🧵
bit.ly
✨The Advanced Manual of Self-Improvement✨
Decision science tips to help you stick to your New Year’s resolutions
100
Cassie Kozyrkov @decisionleader.bsky.social · 24/12/2025
You outgrew your teddy bear because it got boring. AI toys don't get boring. They scale with your kid forever.
100
Cassie Kozyrkov @decisionleader.bsky.social · 23/12/2025
Did you hear the one about the AI-powered teddy bear that engaged in sex talk with kids? Just last month, it got yanked from shelves after teaching children about kitchen knives, fire starting, and BDSM. Maybe check what's in your kids' stocking this year. Full story: decision.substack.com
000
Cassie Kozyrkov @decisionleader.bsky.social · 17/12/2025
Most decision failures aren’t due to lack of data, they’re due to unclear judgment. Cohort 3 of Decision-Making with ChatGPT is now open: 👉 maven.com/cassie-kozyr... #StrategicThinking
000