6 big tech trends for 2025 that you can’t ignore

Keith Shaw: The world of technology continues to disrupt businesses across the globe, and 2024 was no different. As we look ahead to 2025, what will be the big technology trends that shake up companies? We're going to speak with our favorite futurist on this episode of Today in Tech.

Hi everybody.

Welcome to Today in Tech. I'm Keith Shaw. Joining me on the show today is Mike Bechtel. He is the Chief Futurist at Deloitte and a regular guest on the show. Welcome back, Mike. Happy New Year. Mike Bechtel: Oh, Happy New Year, Keith. Thanks for having me, brother.

Keith: All right. So let's talk about Tech Trends 2025. This is a big report. I think you guys have been doing this for 16 years now? Mike: Yeah, yeah. Deloitte has been chronicling what's new and what's next for a full generation of enterprise. Keith: Wow.

We last spoke about the 2024 report in February. This time, we’re closer to the actual release — this is going out in January. Give me a brief overview of the report and then talk about the six "buckets" you use to group the trends.

There’s an overarching framework, but each bucket contains specific trends we're going to explore.

Mike: The thing about being a futurist — and you can imagine, right? We’ve touched on this in the past — is people ask, “What do you have, a crystal ball or something?” And I say, “No.” In fact, there's a distinct lack of clairvoyance or astrology involved.

To be a credible futurist, you actually need to be a bit of a historian and take a wide-angle view. One of the benefits of doing Deloitte Tech Trends for 16 years is that you develop a pretty good pattern detector — separating needles from hay, signal from noise.

The six areas that seem to matter most, both in our historical data and current work, are: Human-computer interaction – the ways people work with tech. The trend is toward simplicity, intuitiveness, and clarity. Information layer – formerly called business logic and algorithms; essentially decision-making.

The trend here is toward greater intelligence — smarter systems. Computation – processing power and the ability to do more with less. The trend here is abundance or strength.

Now, if you're thinking this all sounds optimistic and lofty, you're right. These are what we call the elevating forces — areas that drive growth. Simpler, smarter, stronger systems attract investment and attention. But our clients are enterprises — household-name organizations that have reputations to protect.

So we also consider grounding forces, or what I like to call "adulting forces," that temper that optimism: The business of IT – encompassing culture, talent, and finance. As the saying goes, "Culture eats strategy for breakfast." Cybersecurity – because there are always bad actors out there.

You’ve got to protect the house. Core modernization – it might sound old-school, but 70% of enterprise IT budgets still go to maintaining legacy systems. "Nursing Old Bessie along," as we say. Each year, we explore what’s new in these six areas. They're the treasure hunts that matter.

Keith: This is a brilliant way to look at the future. You’ve got three trends that represent lofty goals, and three that act almost like gravity — obstacles and challenges. So, is there a push and pull dynamic?

And my second question is — actually, maybe it’s the first — did you come up with this framework at the beginning, or did the six trends evolve over time? Mike: Great questions.

I’ll answer in reverse, as any proper geek would say — LIFO stack. This framework emerged from first principles. Sixteen years ago, it was just a flood of novelties.

We asked, "What truly matters?" Through back-testing, we found that the elevating trends — simplicity, intelligence, and computational strength — were key indicators of staying power.

As for the grounding forces — business, cybersecurity, legacy tech — those also emerged from patterns. If a company wasn’t addressing these, they usually didn’t last. Keith: Before we dive into the 2025 trends, do you ever go back and grade the 2024 ones?

Like a report card — did you get it right or not? Mike: I'm happy to look back. As a futurist, if I ignored history, I wouldn’t have much credibility.

Mike: Let’s start with interaction. Last year, we talked about “interfaces in new places.” That trend was a bit of a Trojan horse — really a hint at the early stages of spatial computing. Keith: Super straight talk, Mike: VR has felt “two years away” for the last 25 years.

I remember back in the ’90s, we were already talking about VR like it was the next big thing. Mike: And you’re right — that’s okay.

Twenty-five years ago, virtual reality was a novelty you’d experience at the mall. You’d pay 50 bucks (inflation-adjusted) to strap in and play a multiplayer video game. What we’ve seen with interface tech is a progression — from tech to toy to tool. From demo to plaything to utility.

Last year in Tech Trends, we said: major investments are being made — by big tech and small tech alike — that mark the beginning of the end for screens and rectangles. We're moving toward virtual and augmented experiences. Keith: So how’d that trend do? Mike: Incomplete.

I wouldn’t give it an A or a D. We’re seeing early shifts: headsets from big players are becoming less expensive and bulky. Those are evolving into smart glasses, which may eventually become contact lenses.

Since we last talked, AI pins have also emerged — devices that project pixels directly into your field of view. Early days, but promising. Keith: Wow.

So I’m the cynic here — you’re the optimist. I still think we’re not quite there yet. I mean, this was all before Apple Vision Pro came out, right? And now we’re seeing Meta's latest Oculus, and those Ray-Ban smart glasses... Mike: Right.

And I don’t know if the early success of those glasses surprised Meta, or if they planned it — but it definitely seems like they’re doubling down due to actual traction.

Keith: After the Vision Pro, the discussion shifted to: “This is great, but we need a lighter version.” That’s when we started seeing a clearer path to augmented reality. So yeah, it feels like we’re at the beginning — but not quite mainstream yet. Mike: Exactly.

You have to walk before you run. Keith: It reminds me of media hype. Someone wrote recently, “We finally have the killer app for VR.” But I haven’t even clicked the article yet — I’m guessing it’s hyperbole. Mike: Even if it is, we are seeing form factor changes.

Nobody wants to live in the Daft Punk-style helmet forever. But we’re on that path.

Here’s a teaser for the 2025 report: much of the action in spatial computing is shifting away from headsets and hardware. It’s moving toward the data — specifically, the spatial data management that underpins those experiences.

We spoke with companies all over the world, and they said: “3D CAD data has always been confined to architects and designers. Now, it needs to be integrated with enterprise data so AI systems can interact with it.” Keith: That’s fascinating. Give me an example. Mike: Sure.

Benfica — the Portuguese soccer team — is using computer vision to track all the players on the field. They're analyzing positions, tendencies, weaknesses, and strengths. It’s spatial computing... without the headset. Keith: Like Moneyball for soccer. Mike: Exactly. It’s all spatial — just minus the gear on your face.

Keith: So within the enterprise world, you’re seeing more momentum than in the consumer space? Mike: Definitely. We’ve talked about digital twins in manufacturing before. That momentum is still going strong.

We also talked to Paramount Global. They’re doing studio lot tours for new hires using AR and VR. Why? Because the lot is always busy, and much of it is off-limits. This gives them a way to deliver an immersive experience without being physically there.

Oil and gas companies are using drones and spatial data to monitor pipelines. Healthcare systems are building digital twins of patient journeys — from generalist to specialist — at a birds-eye level. It’s taking off.

The challenge is, we sometimes throw the baby out with the bathwater because we don’t want to wear bulky headsets. But spatial computing is more than that. Keith: So jumping back into predictions — does the 2025 report still forecast the death of the screen?

Mike: Not quite death — more like merging. That trend is still around. But yeah, I’m with you... I’m looking at way too many screens too.

But here’s what’s interesting — and we’ll touch on this in our “Hardware Strikes Back” section. We're going to see an enterprise hardware refresh like we haven’t seen in a generation, driven by AI chips and NPUs.

New physical gear is coming — still screen-based for now — but smarter and more integrated. Keith: That makes sense. And as a dad, are you concerned about screen time? Mike: Not really. My 12-year-old still plays sports — basketball, football, goofing around in the yard.

But when he’s online, he’s hanging out in 3D virtual environments more than ever. That’s growing, not shrinking. Keith: I’ve noticed that too. During Thanksgiving, for example, my teenagers actually stayed off their phones. We talked. Played games. It was... normal. Maybe because we had scratch tickets as prizes.

But still — it felt like real family time again.

Mike: That’s another subtrend under spatial computing. During the pandemic, we saw a push for full escape: “Get me out of this family room!” But now, we’re heading into an era of balance. Let’s augment the real world — paint pixels on it — but not escape from it entirely.

And yes, “augmented reality” sounds like jargon. But if you say, “Let’s bring reality online,” it’s easier to get behind that. That’s where spatial computing is heading: reality itself coming online. Keith: All right, let’s move into the next trend.

I want to make sure we hit all of them before we run out of time. The first one is AI Everywhere. You’re saying tasks will become more intuitive and efficient, transforming industries like healthcare, logistics, and education. You even say AI will become as ubiquitous as electricity.

Which leads to my question: will people even care whether a tool is using AI? I mean, I don't think about electricity anymore — I just flip the switch. Will AI become that invisible?

Mike: There’s a lot to unpack there. But yes, that’s the direction we’re heading. I like to think of it as AI underground — or AI undercover. We’re at the point where AI is starting to fade into the background.

I remember when technologists used to obsess over HTTP and the seven-layer TCP/IP stack. But one day... you just didn’t need to know anymore. It became assumed.

We’re starting to see that with AI. Whether it’s the transistor, HTTP, or electricity — it fades from the headline and becomes part of the plumbing. You’ll just ask for a 30-second video — and it’ll use AI.

Or a customer service bot that’s competent and empathetic — and it’ll use AI. So yes, AI will be everywhere. And eventually, you won’t even call it “AI.” Keith: That makes sense. And maybe we’ll only care when it breaks — just like with electricity.

You don’t call an electrician unless something’s wrong. Mike: Right. And I love your example about using AI to create gifts for your theater cast. Right now, you explain how you used AI. But eventually, you’ll just say, “Here’s a cool gift,” and that’ll be it. Keith: Exactly.

I made AI-generated buttons based on photos of their characters. And everyone was impressed — but someday I won’t have to say “AI did this.” It’ll just be normal. Mike: Yep. Reminds me of a college roommate from the ’90s.

He was the first person I knew who would use the Web for daily tasks. “I’ll go online to book the game.” It was noteworthy at the time. Now? Of course you would. Same thing with AI.

You’ll augment yourself with a posse of intelligent agents — and that will be the norm.

Keith: So what’s next for AI? We’re moving from large language models to smaller, more efficient models — and also into agentic AI. That’s been a huge topic on this show lately. Does that mean we’re moving away from the "one-size-fits-all" approach to AI?

Are we going to see more specialized agents for specific tasks? And if so, are users ready to adopt — and even pay for — these more focused tools? Because it feels like, in the past few years, it was all about picking the right big, generalized GPT tool.

But now it’s more like what happened with TV. We had three channels... then cable came along and suddenly we had a hundred options.

Mike: I completely agree. We’re definitely moving from monolithic, “one-chat-box-to-rule-them-all” approaches into a plurality. In our Tech Trends report, we include interviews and case studies with companies making this shift. We talked to Dell — specifically an executive named Vivek Mahindra who oversees a lot of their AI work.

He told us the one-size-fits-all approach isn’t going to be obsolete — it already is.

It’s about right-sizing the tool. You don’t need a jackhammer for everything. Sometimes, you need a ball-peen hammer. Sometimes, a rubber mallet. A small language model isn’t about settling for a "dumber" AI.

It’s about getting a fast, domain-specific model trained on exactly what you need — like call center questions. And frankly, you don’t want customers using that model to debate Socratic philosophy. It’s more useful when it can’t do that.

From an energy usage and clarity-of-intent standpoint, this is a win. There’s an old line in innovation: “Innovation loves constraints.” Ask someone to come up with a big idea and they freeze.

But give them specific boundaries — like "draw a cartoon puppy using six colors and seven names" — and suddenly they’re inspired.

AI agents are similar. When users know what each one is good at — “this one’s the butler,” “that one’s the medical student” — it becomes far easier to interact with them. The constraint creates confidence. Keith: Okay, I get that.

But I haven’t talked to you since this agentic AI trend really took off. So I want you to put your futurist hat on and explain it to me — plainly. Because I’ve been wrestling with what these AI agents are really supposed to do.

Let's take a consumer use case: planning a trip. Right now, I have to research hotels, flights, things to do, which city to visit — and I click and click and click.

The appeal of agentic AI is that I can just say, “Find me the best trip to Florida in January,” and the agent takes care of the rest.

But to do that, I’d have to hand over a lot of my personal data — my airline preferences, my past bookings, my budget. That part of it makes my brain hurt. I’m not sure I’m ready for that level of access. So am I using the wrong use case?

Am I assuming agentic AI is just Jarvis from Iron Man? Do we need to reframe our expectations?

Mike: I love how you think through all this, Keith. And I wouldn’t change a thing — this richness is what makes us human. Here’s what helped me wrap my head around agentic AI: We’re basically talking about computerized concierges — each one with a clear role and specialty.

And crucially, they can delegate to other concierges in a federated way. That means you, as the user, don’t have to worry about how the sausage is made.

Think of it like Downton Abbey. Keith: Oh boy. You’re going to make me feel either really young or really old. Mike: Stay with me.

Lord Grantham didn’t say, “Gas the car, make my lunch, shine my shoes.” He just said, “I’m going to town.” Everyone on staff knew what that meant — the butler had a job, the cook had a job, the chauffeur had a job.

And they coordinated with each other without him needing to micromanage. That’s agentic AI. You say, “I’m going to Cleveland,” and your network of digital agents figures it out — flights, hotels, calendar syncing, whatever’s needed.

Keith: That’s actually brilliant. I haven’t heard agentic AI described as “basically Downton Abbey.” But I love it. Mike: It’s a metaphor I used years ago, before we even had a name for this stuff.

Back then, we were calling it "ambient experience" or "federated experiential agents" — very nerdy, long-winded terms. The Downton Abbey analogy made it click for people. It’s digital staff — each with a role, hierarchy, and chain of command. The user is Lord Grantham.

Keith: I almost prefer the term “AI butler” or “AI concierge.” It’s immediately understandable. And within that world, you’d have reporting lines — some agents report to others. The human stays in control but doesn’t need to be involved in every step. Mike: Exactly.

And from a software architecture standpoint, that’s not even new. It’s like object-oriented programming — functions calling functions. The key difference now is that these agents can act. That’s what “agent” means. You’re granting them autonomy.

Companies like ServiceNow are doing this with their Xanadu platform. Salesforce is doing it with AgentForce. Databricks too. It’s coming — because it’s useful. Keith: You mentioned the big T-word: trust.

Given what we've seen with the ups and downs of generative AI, are we actually ready to trust these agents?

Mike: At Deloitte, we say: Trust trumps tech. None of this matters if you don’t build in governance, risk, ethics, and security from the very beginning — not as a final coat of paint, but as part of detailed design.

And here’s where I’ll loop back to something I mentioned earlier: The utility of AI is often inversely proportional to the stakes. Keith: What do you mean by that? Mike: Well, think of travel planning.

If you book a slightly worse hotel than you wanted, it’s annoying — but it’s not life-threatening. That’s a low-stakes problem. You’re more willing to let the AI take a shot at it. But when it comes to healthcare or banking, that margin of error becomes unacceptable. You need precision.

You need accountability.

That’s why most early agentic AI use cases are in lower-stakes environments — so trust can be earned over time. As it proves itself, it can graduate to higher-stakes tasks. Keith: So you’re saying early trust will be built through safe use cases. Mike: Exactly.

We heard this from folks at ServiceNow, like Chris Bedi, their Chief Customer Officer. He said agents won’t replace humans — but they can work alongside them. Handle the repetitive tasks. Pull background info. Do the digging. Keith: Human in the loop. Mike: Yep.

As long as there’s a human in the loop, bring on the robot reinforcements. Because at the end of the day, a machine can’t be accountable — but a human can. Keith: That’s a great quote. And you know, that reminds me of buying things online.

The first time you did it, there was some hesitation. “Will they really ship the book?” But once you experienced that trust, it became second nature. Mike: Yes!

And that’s how it always goes with tech adoption. There’s a line from Dodgeball: “If you can dodge a wrench, you can dodge a ball.” For online shopping, the value stack evolved. The more you trusted the experience, the more you used it. Same thing with AI agents.

And generational shifts matter. My teenage son recently explained the difference between millennials and Gen Z like this: “Millennials will only buy expensive things on a laptop. Gen Z is fine doing it on their phone.” Keith: Wow. That’s so true.

I’d still sit at a desk for something like buying insurance. Mike: Exactly. But as people get normed to new behaviors, what was once scary becomes second nature.

Keith: All right, I need to move us along. We could talk all day about how we’re personally adopting technology, but I want to make sure we hit the rest of the trends.

Next up is hardware resurgence — this is where you talk about how hardware is “eating the world.” You reference advanced chips, sustainability, edge capabilities, and internal hardware.

What I wanted to ask was: does this also mean we’ll start seeing new hardware designs — beyond the typical laptop, phone, or tablet? Because tablets had their moment and feel less exciting now. And when you mention IoT and robotics, it always feels kind of niche.

Mike: Great questions — and I’m not going to apologize for asking three questions at once. It’s all good. We titled this year’s trend “Hardware Is Eating the World” as a 13-year-later response to Marc Andreessen’s famous line that “software is eating the world.”

Back when I was a venture capitalist, the test was: is it a software company? And is it as-a-service? If yes, it was investable. Why? Because for a generation, the big success stories were built on bits and pixels — not atoms and things.

Software was defensible, scalable, hard to reverse-engineer. It was the business model of the decade. But now, with the rise of AI workloads, we’re seeing a shift. Hardware matters again.

For the first time in a generation, CIOs and CTOs are thinking of hardware not as a commodity, but as a strategic differentiator. It’s showing up in two places: The data center. For the past 10 years, we’ve seen a march toward the cloud. But now?

It’s not a reversal, but it is a reconsideration. For some AI workloads — like an agentic solution in a healthcare call center — having next-gen machines on-prem might be better. Faster. More secure.

So, cloud isn’t going away, but it’s no longer a default “no-brainer.” Endpoint devices — laptops, desktops, phones. With the rise of NPUs and AI chips, there are new capabilities that only work on new machines.

CIOs are now rethinking laptop refresh cycles in ways we haven’t seen in 15 years.

Keith: That’s interesting. But we haven’t yet seen a device on the end-user side that’s really exciting. Are you seeing something we’re not? Mike: Well, I think this is about supply and demand. On the supply side, companies like HP and Dell are investing heavily and getting excited.

I was at a recent conference, and the energy was real — hardware is “back on the menu,” as they say. On the demand side, here’s where the trends connect.

If you believe in: Small language models Domain-specific use cases Governance, security, and risk management ...then there will be more on-device AI.

Let’s say you’re a Department of Defense agency. You need to run AI code on a local, air-gapped device inside a SCIF (Sensitive Compartmented Information Facility). That’s the clearest use case for an AI PC I’ve ever heard.

It starts with niche, vertical-specific needs — and then trickles down into horizontal business use cases. Keith: So, almost the inverse of how consumer AI rolled out. That started general and became specialized. This will start specialized and eventually become general. Mike: Exactly.

Keith: Let’s move to the next trend: IT Amplified. This one focuses on citizen developers, automation, and the shift to lean IT strategies.

I’ve seen companies embrace the citizen developer concept, but my cynical side hears that and thinks: “Oh, so now you want me to build my own tools just to do my job?” That puts up a mental wall.

I’m not quite at the low-code/no-code level yet — it still feels like programming. Mike: Totally fair.

Let’s break this into two parts — FIFO style. First, last year’s trend was about moving from DevOps to DevEx — from developer operations to developer experience. What that really meant was: techies are wired differently, and companies are recognizing that.

They’re shifting incentives and org structures to support how engineers actually work.

For example, instead of promoting a high-performing coder into management (and taking them away from coding), companies are now rewarding them with autonomy, free time, and strategic projects. Not just a briefcase and a new title. At Deloitte, we’ve gone all-in on this over the past 12 months.

We’re building an engineer-first culture — cutting red tape, celebrating deep technical expertise, and aligning rewards accordingly.

Google is doing the same. And now we’re seeing this evolve into IT Amplified. What that means is: instead of saying, “Great, we can use AI to replace Toby,” leading companies are saying, “Great, now Toby can finally work on that strategic project we’ve been putting off for a year.”

AI should augment, not eliminate. It should amplify human capabilities, not replace them. Keith: Okay, that makes sense. I hope companies get that message. But I imagine there’s a wide spectrum — some will embrace it, others may just cut headcount and claim it’s innovation.

The cynic in me wonders: do people even want to work anymore? Not in a generational way, like “kids these days” — more like, if AI really does eliminate the mundane, do people know what they’d actually do with that free time? Mike: That’s such a human question.

There are always extremes. On one end, you’ve got The Jetsons — George Jetson working an hour a day, pushing one button. On the other end, we risk replacing repetitive tasks with... even more repetitive tasks. But what usually happens is the nature of the work elevates.

Most of us work a professional-ish week — balanced by human fatigue, societal norms, and productivity needs. But the type of work evolves.

Our grandparents wouldn’t recognize what we do as work. And that’s not a roast — it’s just how things change. There’s a futurist I admire, Mike Walsh, who says: “Every time we automate, we need to make sure we also elevate.” That’s the responsibility of leaders.

And that’s how AI becomes rocket fuel, not a weight-loss pill. Keith: Wow, that’s profound. I like that a lot.

Keith: All right, onto the next one. This fits under the cyber and trust bucket — specifically post-quantum encryption. Here’s my question: how real is the risk of quantum computers breaking today’s encryption? Is this an urgent red flag? Or just another futuristic pitch?

Also, “quantum computing” needs a new name. You hear it and people immediately check out. I understand AI — but “quantum” still feels like a black box.

Mike: I love that, and I’m so glad you brought props. [Mike steps off-screen and returns with a black-and-white photo.] Mike: Do you recognize this gentleman? Keith: No — but I know I should. Mike: This is Richard Feynman — quantum physicist, Manhattan Project contributor, and brilliant science communicator.

He once said: “Even quantum physicists barely understand quantum mechanics.” So if your eyes glaze over at “quantum computing,” you’re not alone.

Here’s why it matters: There’s a day coming — some call it Y2Q (like Y2K, but quantum). The problem is, we don’t know when that day will arrive.

But once a quantum supercomputer achieves quantum supremacy — meaning it can break today’s encryption like SHA-256 and elliptical curve cryptography — everything changes. Secrets, passwords, infrastructure — vulnerable.

So while the importance is 100%, the urgency is unknown. Could be 2–3 years. Could be more. But smart companies aren’t waiting. NIST has already introduced post-quantum encryption standards, like lattice-based cryptography.

The time to start migrating is now — while you still can, rather than scrambling when it’s too late. Keith: And are companies ready? Mike: Not yet.

The challenge in cybersecurity is that there’s always a present-tense fire. Phishing attacks, ransomware, prompt injection, DDoS... it never ends. So carving out time to plan for a future threat like Y2Q is tough — but necessary. Because when it hits, it will be massive.

Keith: Okay, final trend — let’s get to it. This one is intelligent core modernization, and I think we talked about this in the 2024 edition as well.

So now the trend is: AI will be central to upgrading legacy systems for smarter operations, with an emphasis on human oversight and governance. What’s different this year? Because last year, I remember you used a healthcare analogy — something like preventative health versus emergency care.

Mike: Yep, you’re absolutely right.

And thank you for laughing — because even core modernization folks will admit this trend feels more like vegetables than dessert. The term we hear a lot is technical debt.

And the reason we use the word “debt” is because every dollar spent maintaining old systems is a dollar not spent building the future.

Last year, we said: “An app a day keeps the doctor away.” Meaning, if you invest just a little in proactive modernization, you can delay — or avoid — those expensive overhauls. This year, we’re seeing AI add juice to that idea.

We talked to companies like ServiceNow and a client of ours called Graybar — they’re using AI inside their core systems (even mainframes) to monitor performance, detect when things go from “green” to “yellow” to “red,” and even begin to self-heal before something breaks.

It’s not just saying, “Hey, this needs attention.” It’s saying, “I’m going to start fixing this for you.” And what better place for AI to focus than the back-end muck that nobody wants to touch anyway? That’s where the value is hiding.

Because as we’ve been saying all along today: AI’s biggest promise is unlocking time — time to pursue more valuable, creative, or strategic efforts. Keith: Yeah, that makes total sense. Especially in the places that are invisible to the customer but critical to keeping everything running.

Keith: Mike, it’s always such a pleasure to talk with you. I know we could easily keep going for another hour or two — there’s just so much depth in each of these trends.

For folks who want to dive deeper into all of this — where’s the best place to access the full Tech Trends 2025 report?

Mike: Thanks, Keith — and the feeling’s mutual. We’ll include a link here on the page, but honestly, just type “Deloitte Tech Trends 2025” into your search engine of choice. Make sure you include the year, and you’ll find it. Keith: Great. We’ll definitely have that link.

And Mike, we’ll get you back before the end of next year. I’m sure something will come up where I’ll go, “Oh! I need to talk to Mike about that.” Thanks again for taking the time to be here today. Mike: Always a pleasure, Keith. Thank you.

Keith: That’s all the time we’ve got for this week’s episode. Be sure to like the video, subscribe to the channel, and leave your thoughts in the comments below. Join us every week for new episodes of Today in Tech. I’m Keith Shaw — thanks for watching.