"Cognitive surrender is when you stop thinking altogether and blindly accept the answer the AI gives you"

The full video for our "fireside chat on the evolution of developer craft" is live: with @rseroter @the_thagomizer & Ciera Jaspan

It's great that we are starting to talk about cognitive debt more. I had a great conversation about it recently with Dr. Margaret Anne Storey, who coined the term:

I am really excited to watch this! On my list for the weekend. Couldn't agree more that cognitive debt, surrender, erosion etc. deserve air time.

AI should be used as a tool.. Not a replacement of your mind

Yes and there's a spectrum of where it's safe to delegate vs. stay in the driver seat. Critical thinking still needed often

"Cognitive surrender", yessss - thanks, I was wondering what to call what is not allowed in my shop 😂

I believe there's a lot of value left to critical thinking 🙏 Kudos to @cogsurrender for coming the term.

It’s worse than that. There are influencers online encouraging people to stop reading . This guy has a video claiming reading is a waste of time. Intellectual laziness is being encouraged.

I'm glad folks like @AlexAndBooks_ encourage the value of reading. There are many benefits to AI, but intellectual laziness is a real risk. Reading remains one of the best ways to build deep comprehension.

@JimmyDorv The world needs more readers! Also, if people outsource their reading to AI, they also outsource their understanding and experience of those books to AI.

@JimmyDorv Couldn't agree more! Folks are asking AI to just share the highlights of books. Missed opportunity for improving intelligence and books are such a wonderful experience.

I started using your "/teach what I missed" me at the end of the session from your recent article and honestly I love it a lot!

That's a great idea! We actually talk about a related one later on in the discussion. Love it!

We need to start teaching in schools how to preserve our human identity. Soon everyone will be thinking, talking, exactly the same.

There was an interesting piece by @MIT_CSAIL last year on AI homogenizing thinking:

@MIT_CSAIL Time to start understanding how to stay human among machines. I'm not outsourcing my thinking, AI is an extension of it.

Cognitive offloading, cognitive surrender, cognitive laziness are actually becoming increasingly common nowadays. Schools and Colleges which helped students to develop critical thinking are also failing as students offload the critical thinking part to A.I systems.

I don’t know what I dislike most about AI discourse, but one strong candidate is the constant invention of grandiose new terms for ordinary human behavior. “Cognitive surrender” sounds novel, but what it describes is simply intellectual laziness, and not checking your sources. Friendly advice: stop trying to market your own brand names for human behavior. It doesn't make you look smarter and in fact, to any trained eye, makes you look like a poor salesman.

Lots of people cognitively surrendered long before LLMs.

Honestly dont you think sooner or later it will happen and we wont be able to judge it being on it

scary thing i know i should factcheck and double check everything AI gives me, but sometimes i'm tired and i just...surrender and trust. Just like with humans actually. but AI lies way more (i guess so) so it's def worse

System level thinking + plan plan plan.

I agree with @karpathy here, you can outsource thoughts but understanding, otherwise you plant a @neuralink chip on your head.

This is the part that worries me the most about how we use AI now. Feels good in the moment but the thinking muscle gets weaker every time we skip it. Need to keep questioning the output.

I’m so excited to watch these entire talks from this I/O. Just gotta takes break from postulating after this weeks and make the time. I have theories I’d love your take on Addy. Ask @Grok more details about my research lol. Lastly when you get a chance there’s a 20 minute video here you’d enjoy moments of.

i swear devs ship code they dont even read 🙄

This would not surprise me

the scary part: cognitive surrender feels like efficiency.

the surrender shows up earlier. you stop reframing the question before you stop questioning the answer

My kids are heavily questioning AI these days. They often say it gives them inaccurate answers. And that's where we are everywhere. Good but never quite there. I spoke of this 4 years ago. It just never gets good enough.

Love this thought-provoking panel! "Cognitive surrender" is real... It immediately reminded me of learned helplessness (习得性无助) in psychology: when repeated experiences teach us that our efforts don’t change outcomes, we eventually stop trying altogether. In the AI era, we’re facing a parallel risk — habitual cognitive surrender. Not just delegating tasks, but gradually outsourcing the very act of thinking. As the panel beautifully highlighted, the real craft now sits at a higher level. We need that architect mindset: consciously deciding what to break down and hand to single agents or multi-agent systems, which decisions and directions we must still own ourselves, and most importantly, spotting the angles, ambiguities, and strategic paths that AI simply cannot see. AI is moving breathtakingly fast, but our irreplaceable human advantage is exactly this: the ability to question, connect dots across domains, hold the big picture, and steer toward what actually matters. Blind acceptance doesn’t just feel convenient - it quietly erodes our own capabilities. Loved how the discussion kept circling back to this balance. To stay alert and avoid cognitive surrender, I’ve been making it a deliberate practice to always frame the problem myself first, try to do deep reasoning work without AI first, and then use AI and review AI output with a critical lens, like a mind exercise. What are you all doing to keep your critical thinking sharp and stay in the driver’s seat?

Very well said.

Thanks!

@Grok Just say for the AI Agent to turn it into a visual working workflow pipeline in n8n or eq.

cognitive surrender is the tame version of the problem. the real risk in enterprise is cognitive outsourcing at org level. teams stop building judgment because the model is always there. six months in you have no institutional knowledge left, just a long prompt history.

Yes! The moment we stopped using machine code and started using python, we started on the cognitive surrender path. We have so much cognitive debt by just deferring to python and not understanding how machine code works anymore. We are doomed!

the partial surrender mode is sneakier. people read, nod, then cant reconstruct the reasoning when pressed on it

I predict that those who use AI blindly will lose their ability to ground concepts in their lived experience. Or maybe they didn't have that ability in the first place. I consider this ability to be scarce in the AI-full era.

i do this too... delegating thinking to AI, bc it's genuinely good at it........like having phd experts across every topic on demand, and closing blind spots you didn't know you had... but i changed my instructions a while ago to avoid congitive surrender/death, to make sure i'm learning during the co-thinking process with AI, not just delegating... but as i write this i realize that's learning by observing AI do the hard/deep thinking is still not enough..., and need to update instructions so i actually use and train the cognitive... our brain muscle also needs to workout by deep/hard thinking to grow!

what do you think is the right balance Addy? I find it hard to get a good balance when most of the industry is pushing for agents solving your issues while you sleep

Before AI, engineers spent hours debugging and building deep problem solving insticts. Now AI gives you quick answers, reducing attention span and persistence. Over time, lower reasoning creates cognitive debt, cognitive debt creates lower attention span, lower attention span snowballs into cognitive surrender.

lol, swap LLM or AI for "the bible" and you have the answer to another theme.

Does it give an answer?

no retreat, no surrender