What If AI Is Making You Dumber?

Here’s an uncomfortable question: What if the tool that’s supposed to make you more productive is actually eroding the very skills that made you good at your job in the first place?

I know, I know. That sounds like the kind of technophobic nonsense your uncle posts on Facebook. But bear with me, because the neuroscience tells a story that most AI evangelists don’t want you to hear.

I caught myself last week doing something I’d never noticed before: I was reaching for my phone to solve a simple decision I already knew how to handle. I had an idea. I knew the approach. But some instinct made me think, What if AI has a better solution?

And here’s the thing—sometimes it did. Sometimes it just rephrased what I was already thinking, but more eloquently. Unless you’re a professional writer with a distinctive voice that’s hard to replicate, AI genuinely does produce cleaner prose with faster throughput than most of us typing away.

So what’s the problem?

The problem isn’t that AI writes better. It’s that I’d stopped trusting my own thinking enough to even try first.

And I’m not alone. If you’re a knowledge worker right now, you’re probably feeling it too—that creeping sense that AI is both saving you time and somehow… stealing something from you. Your attention. Your creativity. Maybe even your ability to think deeply about complex problems.

The promise was simple: AI would handle the boring stuff so we could focus on the meaningful work. But here’s what nobody told you: your brain doesn’t distinguish between “boring work” and “meaningful work” the way you think it does. And by outsourcing the boring stuff, you might be outsourcing your ability to do the meaningful stuff too.

Let me show you what I mean.


Act 1: The Dopamine Trap—Why AI Feels So Good (And Why That’s a Problem)

You know that feeling when you’re working on something hard, and you just… can’t quite figure it out? So you click over to social media. Just for a second. Just to clear your head.

Except it’s never just a second, is it?

Here’s what’s happening in your brain: Social media platforms have essentially “druggified” human connection through dopamine-driven reward mechanisms.[1] They use the same psychological tricks as slot machines—variable reinforcement schedules that create habit-forming behavior.[2] Every scroll gives you a tiny hit of dopamine, building tolerance over time and requiring more stimulation to get the same effect.[3]

And here’s the kicker: AI tools work the same way.

When you’re stuck on a problem, clicking over to ChatGPT or Claude gives you that same dopamine hit. You type in your half-formed thought, hit enter, and boom—instant gratification. An answer. A solution. Relief from the discomfort of not knowing.

Each click-refine-click cycle provides a dopamine reward, creating a continuous feedback loop.[4] This mirrors the reward uncertainty that makes gambling addictive. The emotional reinforcement precedes logical evaluation, hijacking your decision-making.[5]

fMRI studies show that problematic social media use activates the same brain regions as substance addiction—specifically the ventral striatum and nucleus accumbens.[6] Your brain literally can’t tell the difference between checking Instagram and doing cocaine. Well, almost.

But here’s where AI gets more insidious than social media: At least with social media, you know you’re procrastinating. With AI, you convince yourself you’re being productive. You’re “working smarter, not harder.” You’re “leveraging tools.” You’re “optimizing your workflow.”

What you’re actually doing is training your brain to avoid the discomfort of deep thinking.

Here’s what I’ve discovered through trial and error: AI-assisted work versus doing something manually produces dramatically different results in retention and understanding. But—and this is critical—it depends entirely on how you prompt it.

When I use careful prompt engineering—where I gather contextual information first, decide what specific points need to be outlined rather than randomly generated, and get AI to show me its thinking process through an outline before drafting—the accuracy, intent, and throughput massively increases compared to thinking from scratch.

The key distinction: AI doing the heavy lifting while you direct it, versus AI directing you when it feels like it.

This isn’t about AI being bad. It’s about most people using it thoughtlessly. They treat it like a magic box that spits out answers. What they should be treating it like is a research assistant that needs clear direction about what you’re trying to accomplish.


Act 2: The Creativity Crisis—What Happens When Your Brain Never Gets Bored

Let’s talk about something you probably never thought you’d hear: You need to be more bored.

I know how that sounds in 2025, when productivity gurus are telling you to optimize every minute of your day. But hear me out, because the neuroscience is pretty clear on this.

Boredom isn’t a bug in your brain—it’s a feature. It’s actually a signal that your brain needs stimulation of a different kind.[7] When you let yourself be bored, your brain activates something called the Default Mode Network (DMN), which is crucial for creativity and problem-solving.[8]

The DMN is like your brain’s background processing system. It integrates spontaneous neural activity with learned knowledge, producing novel, meaningful insights.[9] This is where your best ideas come from—not from frantically consuming content, but from letting your mind wander.

Creativity emerges from slow, spontaneous fluctuations in resting-state brain activity.[9] Boredom increases cognitive flexibility by allowing your mind to explore alternative perspectives.[10] Novelty-seeking, which is driven by boredom, is the foundation of both creativity and curiosity.[11]

But when was the last time you let yourself be truly bored?

I’m guessing it’s been a while. Because the moment you feel that little twinge of boredom—waiting in line, sitting in a waiting room, having a slow moment at work—you reach for your phone. And increasingly, you reach for an AI tool.

Need to write something? AI. Need to research something? AI. Need to brainstorm? AI.

You’re never giving your Default Mode Network a chance to do its job.

Recent neuroscience research confirms that spontaneous, off-task thought is essential for mental health and problem-solving.[12] Mind-wandering and intentional boredom allow the hippocampus to consolidate memories and replay experiences for learning.[13]

I’ve noticed something interesting: my best insights come during activities that require complete attention but no screen. Running. Walking. Any task that pulls you away from direct computer engagement.

Why? Because it’s harder to mindlessly scroll social media when you need to focus on not tripping over a curb. This type of insight feels more creatively raw—less processed, more genuine.

Now, to be fair, you can still get inspiration by throwing random ideas into ChatGPT. But it’s hit or miss if you’re trying to capture a specific feel. Without proper prompt engineering, you’re getting general AI slop. On the positive side, AI has completely shattered writer’s block when used constructively—giving you that initial momentum to overcome the blank page.

But AI is stealing those moments from you. Every time you reach for a tool to fill that cognitive gap, you’re robbing your brain of the space it needs to make those connections on its own.


Act 3: The Outsourcing Spiral—The Hidden Cost of Convenience

Let’s get real for a second: AI is exhausting you in ways you probably don’t even recognize.

You thought AI was supposed to reduce your cognitive load, right? Just hand off the boring stuff and focus on the strategic thinking. Except that’s not what’s happening.

Here’s what’s actually happening: AI usage creates what researchers call “cognitive offloading,” which leads to skill erosion when overused.[14] People become willing to offload cognitive tasks to AI when they’re under high cognitive load, but this creates a dependency.[15]

And here’s the paradox: The more you use AI to reduce cognitive load, the more cognitively loaded you become.

Why? Because now you have to evaluate AI outputs, which requires understanding something you didn’t actually create. Researchers call this the “efficiency-accountability tradeoff”—prioritizing AI efficiency erodes human comprehension and creates over-reliance.[16]

Think about it: How many times have you asked AI to write something, then spent twenty minutes editing it, trying to figure out if it’s actually saying what you want it to say? You’re not saving time—you’re just shifting the cognitive burden from creation to evaluation. And evaluation without comprehension is exhausting.

Studies show that AI anxiety and attitudes toward AI significantly increase decision fatigue in daily technology use.[17]Cognitive load actually increases when AI provides answers without explanation, because you still have to evaluate outputs without understanding them.[18]

The research has a term for this: “burned out by technology.” It’s the overwhelming psychological and physiological fatigue from constant tech engagement.[19]

Here’s where it gets interesting: I’ve noticed AI can be both energizing and exhausting, depending on how you engage with it.

When I’m overly engrossed in AI-assisted research, I tend to go in learning a ton and leave with more questions. This creates a perpetual feedback loop that some might call obsessive—though I’d argue it’s more like passionate curiosity. Your mileage may vary.

In the early days of ChatGPT, when it wasn’t as polished, getting a bespoke response without high-level prompt engineering was genuinely frustrating. Notice the theme here: AI directed with human thinking versus AI used thoughtlessly without thinking.

For me personally, AI is mostly energizing. But I’ve learned to recognize when taking a break is the right call. The exhaustion comes not from using AI, but from using it badly—letting it drive when you should be steering.

Here’s the uncomfortable truth: When you offload thinking to AI, you often don’t understand the outputs you’re using. You become a pass-through device. Research confirms that people willing to use AI with minimal evaluation create outputs they cannot explain or defend.

This creates skill atrophy. The cognitive muscles needed for original thought weaken from disuse.[14]

And the scariest part? You don’t notice it happening. It’s not like you wake up one day and can’t think anymore. It’s gradual. Insidious. You just start reaching for AI a little more often. Trusting your own thinking a little less.


Act 4: The Mental Bandwidth Tax—Why You Can’t Think Anymore

Let’s talk about something you’ve probably noticed but couldn’t quite name: your attention is fragmented. Like, reallyfragmented.

You sit down to work on something important, and within minutes, you’ve:

  • Checked Slack
  • Glanced at your email
  • Opened ChatGPT “just to see if AI could help”
  • Clicked over to Twitter to “clear your head”
  • Come back to your work, lost your train of thought, and started the cycle again

Sound familiar?

Digital overwhelm leads to attention fragmentation and degraded social interaction.[1] Constant digital engagement creates mental distraction, reducing your capacity for deep work.[20]

You’re not just distracted—you’re paying what I call a mental bandwidth tax. Every app you have open, every notification you haven’t cleared, every AI tool running in the background—they’re all taking up mental real estate. Your brain is running too many background processes, and it’s slowing down your primary processor.

Here’s what’s wild: The solution has been staring us in the face the whole time, but we’ve been too busy optimizing our tools to notice.

Research on digital detox is unambiguous: Deliberate digital withdrawal enhances “eudaimonic well-being”—that’s a fancy term for purposeful living and personal growth.[21] Digital detox improves attention, reduces stress, and enhances self-reflection.[21]

But it’s not just about unplugging. It’s about recognizing that your brain needs rest breaks to function optimally. Restoration of attention through rest breaks significantly improves performance in multitasking environments.[22]

I practice something simple but powerful: I don’t check my phone first thing in the morning until I’m fully up and awake, and until I’ve completed most of my morning routines.

The impact has been significant. I feel more present and intentional with how I spend personal time in the morning. The research backs this up—exposure to digital stimuli first thing in the morning induces stress that may not be beneficial right out of the gate.

This isn’t about rejecting technology. It’s about being deliberate with when you let it into your headspace. The morning is when your mind is freshest, most creative, least cluttered. Why would you immediately fill it with other people’s priorities, notifications, and algorithmic feeds?


Act 5: The Solution—The Radical Act of Using Your Hands

Okay, here’s where I’m going to lose some of you. But stick with me, because this is where the science gets really interesting.

Ready? Start writing things by hand.

I know, I know. It sounds absurdly low-tech. Practically Luddite. But the neuroscience is overwhelming on this point, and it reveals something profound about how we think.

Handwriting activates widespread brain connectivity across sensory, motor, and cognitive regions.[23] The motor component of handwriting creates stronger memory encoding than typing.[24] Students who take handwritten notes perform better on conceptual questions than laptop note-takers.[25]

But it’s not just about memory. It’s about thinking.

When you type, your fingers are mostly on autopilot. You’re transcribing thoughts that already exist in your head. But when you write by hand, something different happens. The slower pace forces you to think more carefully about what you’re writing. You can’t copy-paste. You can’t let AI finish your sentences. You have to think through each word, each phrase, each idea.

I use a hybrid approach to drafting that’s evolved with how I work: I use what’s on hand at the time.

I used to carry my laptop everywhere. That’s simply not feasible anymore. Now I carry a pocket notebook wherever I go, using writing to capture notes and ideas on the fly. If I’m pressed for time, I’ll do an audio recording and then engage AI processes to transcribe and organize my thinking.

The key insight: different tools for different moments and different types of thinking.

Here’s my philosophy on this: There’s a time and place for both approaches. Use the best tools available to you in the given moment.

Why? Because inaction has an expensive penalty. Speed is almost always favorable over the need to solely depend on either analog or digital methods. That said, I’ve found that for certain applications, one method clearly outperforms the other.

The key is recognizing which is which—and not defaulting to AI just because it’s faster, or avoiding it out of principle when it would genuinely accelerate your work without compromising your thinking.

This is the real solution to the AI overwhelm problem: Not rejecting AI entirely, but being intentional about when you use your own cognitive muscles and when you reach for tools.

Think of it like physical fitness. You wouldn’t use an elevator for every single flight of stairs and expect to stay in shape. Your muscles need resistance to stay strong. Your brain is no different.

When you always reach for AI, you’re using the cognitive elevator. And just like your body gets weaker when you never take the stairs, your thinking gets weaker when you never engage in the hard work of figuring things out yourself.


So What Do You Actually Do?

Look, I’m not suggesting you throw your laptop in a lake and become a monk. AI is powerful, and it’s not going anywhere.

But here’s what I am suggesting:

Create friction points. Before you reach for AI, ask yourself: “Am I using this because I genuinely need help, or because I’m avoiding the discomfort of thinking?”

Protect your boredom. Schedule time for “analog thinking”—walks, showers, drives—where you’re not consuming anything. Just letting your mind wander.

Use your hands. Draft important thinking by hand first. Let AI help with execution, but not with thinking.

Notice the pattern. Pay attention to when AI actually helps versus when it just makes you feel productive while making you dumber.

Reclaim your mental bandwidth. You can’t think deeply when you’re juggling seventeen browser tabs and three AI tools. Close some tabs. Create space.

The uncomfortable truth is this: AI is only as good as the thinking you bring to it. If you outsource all your thinking to AI, you’ll eventually have nothing left to bring.

The goal isn’t to become anti-technology. It’s to become pro-thinking. To recognize that the most valuable thing you have to offer as a knowledge worker isn’t your ability to produce more content faster—it’s your ability to think clearly about complex problems.

And you can’t do that if you’re always reaching for tools to do the thinking for you.

So here’s my challenge to you: Pick one task today—just one—and do it the hard way. No AI. No tools. Just you and the problem.

You might be surprised by what your brain can do when you actually let it work.

– Kai T.


Research Citations

This article draws on peer-reviewed neuroscience, psychology, and behavioral research across multiple domains. Key studies include:

Dopamine & Digital Addiction:

Boredom & Creativity:

Cognitive Load & AI Dependency:

Digital Minimalism:

Handwriting & Memory:

DeepSeek: The AI Disruptor Shaking Up Silicon Valley

In a seismic shift that sent shockwaves through the tech industry, Chinese AI startup DeepSeek has emerged as a formidable challenger to Silicon Valley’s AI dominance. The company’s latest AI model, R1, has accomplished what many thought impossible: matching the performance of leading U.S.-based AI systems while slashing costs to a fraction of industry standards.

The $6 Million Revolution

Perhaps the most striking aspect of DeepSeek’s breakthrough is its price tag. The company developed its R1 model for under $6 million—a figure that seems almost impossibly low compared to the billions typically invested by tech giants like OpenAI and Meta. This cost efficiency hasn’t come at the expense of performance; DeepSeek’s R1 has matched or exceeded its competitors across various benchmarks, including sophisticated mathematics and coding challenges.

Market Impact: The DeepSeek Effect

The market’s reaction to DeepSeek’s emergence was swift and dramatic. On January 27, 2025, Nvidia—the chip giant that has been riding the AI wave—saw its stock plummet by nearly 17%, erasing over $400 billion in market value. The ripple effects extended beyond Nvidia, triggering a broader tech sell-off that saw the Nasdaq Composite fall over 3%.

Technical Innovation: Doing More with Less

What sets DeepSeek apart is its innovative approach to AI development. The company’s R1 model employs a “Mixture of Experts” architecture that selectively activates specific neural networks for different tasks, significantly reducing computational overhead. This efficiency means the model can run on less powerful hardware, including older GPU models, making advanced AI capabilities more accessible to a broader range of organizations.

Performance That Speaks Volumes

DeepSeek’s R1 model has demonstrated impressive capabilities across various benchmarks:

  • Achieved a 52.5% pass rate on the American Invitational Mathematics Examination, surpassing established competitors
  • Scored 1450 on Codeforces coding challenges, demonstrating strong programming capabilities
  • Excelled in reasoning-heavy tasks and complex problem-solving scenarios
  • Maintained high performance in multilingual applications and contextual understanding

Industry-Wide Implications

DeepSeek’s success has far-reaching implications for multiple sectors:

  • Manufacturing companies can now implement sophisticated AI solutions at lower costs
  • Healthcare providers have access to powerful diagnostic and research tools
  • Financial institutions can leverage advanced market analysis capabilities
  • Retail businesses can enhance customer experiences through improved AI-driven personalization
  • Biotechnology firms can accelerate research and development processes

The Road Ahead

While DeepSeek’s emergence has disrupted the status quo, questions remain about the long-term implications for the AI industry. The company’s success challenges the assumption that developing cutting-edge AI requires massive capital investments and premium hardware. This democratization of AI technology could accelerate innovation across sectors while forcing established players to rethink their strategies.

The tech industry’s reaction to DeepSeek serves as a reminder that innovation can come from unexpected places and that the future of AI may not be as predetermined as many had thought. As the dust settles from this market shock, one thing is clear: DeepSeek has ushered in a new era of cost-effective, high-performance AI that could reshape the technological landscape for years to come.

– Kai T.

Silicon Valley’s Power Play: Amazon’s $4B Bet on Anthropic

There’s an old saying in the tech world: go big or go home. Amazon, it seems, has chosen to go astronomical with its latest $4 billion investment in Anthropic, marking a seismic shift in the AI landscape that few saw coming.

In what can only be described as a watershed moment in the annals of technological advancement, Amazon’s $4 billion investment in Anthropic represents far more than mere financial maneuvering. It’s a calculated recognition of an undeniable truth: artificial intelligence isn’t just the future – it’s the present’s most pressing imperative.

Let’s delve into the machinery beneath this monumental partnership. At its core, Amazon’s custom silicon – the Trainium and Inferentia chips – represents a fundamental shift in how AI computations are processed. Unlike traditional GPU-based processing, these custom chips are architected specifically for machine learning workloads, offering up to 40% better price performance than comparable GPU-based instances. This isn’t just about processing power; it’s about reimagining the very infrastructure that powers our AI future.

The implications for healthcare alone are staggering. Claude’s natural language processing capabilities are already being deployed in medical research settings, where they’re capable of analyzing complex clinical trials data in hours rather than weeks. Imagine a system that can cross-reference millions of medical journals, identify patterns in patient data, and suggest treatment protocols – all while maintaining the nuanced understanding that healthcare demands. This isn’t science fiction; it’s happening in hospitals and research facilities right now.

In the financial sector, the partnership’s impact becomes even more intriguing. Claude’s ability to detect fraudulent patterns in financial transactions operates at a scale that human analysts simply cannot match. We’re talking about systems that can analyze millions of transactions per second, identifying suspicious patterns while maintaining false positive rates below 0.1%. This level of precision wasn’t possible even a few years ago.

The competitive dynamics at play here deserve closer scrutiny. While Microsoft’s $13 billion investment in OpenAI grabbed headlines, Amazon’s strategic approach with Anthropic might prove more significant in the long run. Why? Because Amazon isn’t just buying into AI technology – they’re building the very infrastructure that will power it. The AWS platform, combined with custom silicon and Anthropic’s models, creates a vertically integrated AI stack that could prove more efficient and cost-effective than competing solutions.

Consider the mathematics of this investment: Amazon’s total $8 billion commitment to Anthropic represents roughly 2% of their annual revenue, yet it positions them to compete in a market projected to generate $15.7 trillion in global economic value by 2030. This isn’t just good business – it’s technological prescience of the highest order.

The technical architecture of this partnership reveals even more fascinating details. Anthropic’s Claude models, when running on AWS’s infrastructure, can process up to 100,000 tokens per second – a rate that makes real-time language processing not just possible but practical for enterprise applications. This level of performance, combined with AWS’s global infrastructure, means AI capabilities can be deployed at the edge, reducing latency and improving user experience.

But perhaps the most intriguing aspect of this partnership is its potential impact on AI development itself. The collaboration between AWS and Anthropic on future chip development suggests we’re moving toward a new paradigm in AI hardware design. Traditional von Neumann architecture, which has served computing well for decades, may give way to new designs specifically optimized for AI workloads. This could lead to exponential improvements in both performance and energy efficiency.

For the business world, this partnership represents a new blueprint for AI integration. Companies can now access enterprise-grade AI capabilities through AWS, with the option to fine-tune Claude models for specific use cases. This democratization of AI technology could accelerate innovation across industries, from manufacturing to creative services.

The ethical implications shouldn’t be overlooked either. Anthropic’s approach to AI safety, combined with Amazon’s global reach, could help establish new standards for responsible AI deployment. This isn’t just about preventing misuse; it’s about building AI systems that are inherently aligned with human values and interests.

Looking at the competitive landscape, this investment positions Amazon uniquely. While Google focuses on consumer-facing AI with Bard, and Microsoft leverages OpenAI for software integration, Amazon is building the foundational infrastructure that could power the next generation of AI applications. It’s a different game entirely – one where the prize isn’t just market share, but the very future of computing itself.

This partnership represents more than just another tech industry investment; it’s a glimpse into a future where AI isn’t just a tool, but a fundamental layer of technological infrastructure. As we stand on the brink of this new era, Amazon’s investment in Anthropic may well be remembered as the moment when AI truly began its transition from promising technology to ubiquitous utility.

– Kai T.