David Linthicum

Microsoft pitched Copilot as the "everyday AI companion" that would turbocharge Windows. Then it landed like bloatware with a badge. This video breaks down how a hype-first launch, a forced Windows update rollout, and messy "everywhere at once" integration turned curiosity into backlash. When an assistant shows up in your taskbar, your search box, and your apps without asking, the question isn't "Is it cool?"—it's "Who's in control?" Add uneven answers, hallucinations, and the constant need to double-check, and the productivity promise collapses into friction. We'll look at why the average user couldn't see a clear win, why power users felt their workflow was being hijacked, and why trust debates (privacy, data, OS-level presence) hit harder than any feature list.

Finally, we'll cover the collateral damage: PC makers betting on "AI PC" narratives while Microsoft keeps shifting the goalposts. Copilot didn't just need to be smart. It needed to be optional, predictable, and genuinely useful. Microsoft shipped loud. Users demanded value. We'll replay the marketing, compare it to real-world tasks, and explain why "integrated" became "intrusive." If you've ever wondered how a feature can be both everywhere and nowhere at the same time, this is the autopsy. No sugarcoating—just receipts.

AI is rapidly turning modern marketing into a surveillance-and-optimization machine. What started with loyalty cards and basic customer databases has evolved into always-on tracking across apps, websites, and devices—feeding models that learn what people want, when they're vulnerable to buying, and how to push them toward a decision. In this video, we break down how "surveillance marketing" works in plain language: companies collect massive amounts of behavioral data, stitch it together with identity graphs and third-party sources, and use AI to target messages in real time.

Then comes the next step: dynamic pricing. Instead of one price for everyone, algorithms can adjust prices on the fly based on demand, timing, channel, device signals, and past behavior—essentially guessing what you're willing to pay. That may boost revenue, but it also creates real risks: bias, unfair outcomes, privacy exposure, and a growing "trust debt" when customers realize the system is opaque.

We'll also cover why the vendor ecosystem matters—data brokers, ad platforms, CDPs, and personalization engines—and why governance is lagging behind. The takeaway: this isn't going away, but it must be architected responsibly, with limits, audits, fairness testing, and transparency.

The launch of the Amazon McKinsey Group (AMG), a high-profile partnership between AWS and McKinsey, is being presented as a game-changing initiative for enterprise-scale digital transformation. They tout end-to-end value, integrated teams, and billion-dollar business impact. But let's cut through the advertising: AMG's very existence highlights the growing desperation among cloud providers and consulting giants faced with the slow, challenging rollout of artificial intelligence across enterprise landscapes. These firms, each with their own vested interests, are combining efforts not to offer truly objective solutions, but to solidify their own revenue streams by locking clients into their own infrastructures and consulting engagements.

This approach severely undermines impartiality, as AWS's core focus is expanding its cloud footprint, and McKinsey is monetizing its transformation playbooks. What gets sold as an "airtight" solution is, in reality, a packaged commercial bundle—engineered to deliver outcomes that benefit the sellers as much as, if not more than, their customers. Before buying into the AMG hype, business leaders need to recognize these partnerships as clever marketing moves rather than unbiased, best-of-breed solutions. Falling for these vendor-driven alliances can cost organizations flexibility, objectivity, and, ultimately, the agility required to thrive in an era of rapid technological change.

I've spent decades watching enterprises adopt technology, and the pattern is always the same: innovation only creates growth when it reduces friction and increases trust. The automotive industry is pushing AI into the cabin as if "more intelligence" automatically means "more demand." But buyers don't purchase abstractions—they purchase outcomes. Right now, much of in-car AI adds complexity to routine tasks, introduces unpredictable behavior, and shifts capabilities behind subscriptions and post-sale updates. That's not a value story; it's a risk story.

What's worse is that automakers are treating the car like a software platform while customers still expect a durable product. If a UI changes every few months, or a feature degrades when connectivity is weak, the car feels less reliable—even if the drivetrain is excellent. And when the AI fails in basic moments—navigation, calling, climate control—people don't think "early adopter." They think "I overpaid."

So instead of creating incremental sales, today's AI often inflates cost, increases buyer hesitation, and drives shoppers toward simpler alternatives. The industry needs fewer demos and more dependable, measurable utility.

In this video, I'm going to take you back to 1985, when building AI meant rolling up your sleeves and encoding expertise by hand. I'll tell a personal story from that era—using Prolog, Lisp, and Borland M1 to create rule-based systems that could make decisions in the real world, long before the cloud and GPUs made "intelligence" feel instant. Then we'll jump to 2026, where AI is defined by foundation models, tool-using agents, and systems that learn from enormous datasets rather than just following explicit rules. You'll see what we gained—speed, scale, and the ability to work with messy language and unstructured information—and what we gave up, including some of the determinism and straightforward explainability of classic expert systems. Finally, I'll lay out a practical view of where the industry is headed: the most valuable architectures don't pick sides, they combine modern models with governance, evaluation, and good old-fashioned business logic to deliver outcomes you can trust.

Windows 11 can run local AI, but in real day-to-day use it often feels like it's working against you—especially once you start stacking multiple AI tools, projects, and installs. What I found is that Linux generally delivers a smoother "AI desktop" experience: setups are more straightforward, common AI instructions match what you're actually running, and GPU-accelerated apps tend to behave more consistently. The result is less time spent troubleshooting and more time getting outputs.

On Windows 11, the biggest pain points showed up around friction and interruptions—extra steps during installs, more chances for version mismatches, and occasional driver or update moments that break a working setup. Even when performance is similar, the overall workflow can feel slower because you're dealing with more overhead and more "little problems" that add up.

Linux stood out for predictability: once things were working, they stayed working. Tools were easier to manage, projects were easier to separate, and the system felt more responsive while running AI tasks alongside normal desktop work.

If you're building a desktop or laptop mainly to run AI locally, this video explains why Linux often ends up being the more reliable, less stressful choice—and how to decide if it makes sense for you.

AI agents are the new buzzword in enterprise tech, but the real question isn't "can we build them?"—it's "can we actually sell them in a way enterprises will trust and fund?" This video, from a Dave Linthicum-style vantage point, cuts through the marketing gloss and treats agents as what they really are: autonomous software systems wired into messy, mission-critical environments. We unpack what an AI agent actually is, how it differs from a simple chatbot or workflow, and what it takes architecturally to move from a cool demo to a production-grade capability.

From there, we tackle the harder part: the business model. Who's really buying agents, and what are they expecting to get—labor replacement, outcome guarantees, or just experimental toys? We look at why agent marketplaces are overhyped, why domain specificity and deep integration matter more than model choice, and why trust, governance, and accountability will determine who makes money in this space. If you're wondering whether "AI agents" are the next big product category or just another repackaged services play, this video gives you a brutally honest, enterprise-centric take.

Windows 11 wasn't "the future"—it was a forced pivot. Microsoft took an OS people relied on for speed, flexibility, and control, then locked the door behind TPM 2.0, Secure Boot, and arbitrary CPU lists that stranded millions of perfectly good PCs. And for what? A redesigned UI that's less customizable, a Start Menu that feels like a billboard, and a setup flow that tries to drag Home users into an always-online Microsoft account whether they want it or not.

Then comes the real point: Windows 11 increasingly feels like a platform for Microsoft's priorities—cloud services, Edge, Copilot, and AI-first features—instead of a tool built around the user. Privacy concerns haven't eased either, with telemetry worries and controversial features like Recall raising the question: how much of your desktop is yours, and how much is being watched, indexed, and monetized?

Meanwhile, everyday annoyances stack up: File Explorer weirdness, inconsistent performance, and "updates" that often add friction instead of fixing fundamentals. In this video, we break down why Windows 11's biggest failure isn't one bug or one design choice—it's the message behind it: comply, subscribe, and get out of the way.

Everyone's hyped about AI breakthroughs—but almost nobody is talking about the bill that's being handed to normal people.

In this video, we break down how the AI gold rush is quietly driving up the price of basic hardware: GPUs, RAM, SSDs, and even CPUs. Hyperscalers are signing multi‑billion‑dollar contracts and buying entire foundry runs of chips, and that doesn't just drain supply—it resets the global price floor for everyone else. The result? Gamers, PC builders, small IT shops, and indie ML labs are all forced to pay more for the same components they used to buy a few years ago.

We'll look at how this "micro‑inflation" shows up as an extra 10–20% on a part here and there, and how that snowballs into a serious hidden tax across a full build or refresh cycle. This is the new two‑tier hardware market: trillion‑dollar AI giants at the top, and everyone else fighting over overpriced scraps.

If you care about hardware, open ecosystems, or just not getting screwed by invisible market dynamics, this one's for you.

Is AI coming for your job… or not yet? In this video, I break down a simple test to understand how exposed your role really is, based on the patterns of work AI is best at replacing. Instead of vague hype, we'll look at concrete signals that your job might be in the danger zone.

You'll see the core pattern behind high‑risk roles: predictable, rules-based tasks, high repetition, and "good enough" outputs where speed and cost matter more than originality. I'll walk through real examples like customer support, basic content production, document processing, and reporting roles—and explain why they're so easy for AI to swallow.

Then we'll zoom out: how many hours of your week look like this? How much of your workflow could AI already do end-to-end if your company really pushed it?

By the end, you'll know whether AI is likely to replace big chunks of your job, merely assist you, or force you to move up the value chain. Use this as a wake-up call—not to panic, but to start deliberately shifting your skills toward judgment, relationships, and work that doesn't look like a template.