If you think AI is mostly memes and marketing fluff, you’re not alone. Intel’s Robert Hallock was equally skeptical at one point in time. Now, he’s betting that AI, the quiet kind, will be the key to Intel unlocking the kind of power, performance, and battery life truly befitting a next-generation laptop.
In a conversation encompassing everything from ethical AI cocoa farming to the futility of the F1 key, and the rollercoaster ride that is the Gartner Hype Cycle, Hallock shines a light on Intel’s AI ambitions, unpacks the challenges ahead, and offers a counterpoint to the hunt for AI’s killer app… All 450 and counting of them.
An interview with Intel's Robert Hallock
Robert is a senior director of technical marketing at Intel for Core processors and technologies. Prior to joining Intel, Robert spent 12 years in Client and Graphics at AMD, most recently as the director of product and technology marketing for Ryzen processors. Robert has also been a PC hardware reviewer, journalist, and technical writer. He moonlights as a designer of high-performance aftermarket automotive components and is a lifelong PC enthusiast.
What’s your role in AI?
“We’re just trying to make computers faster … and AI is the new face of that.”
Early into our conversation, I asked Robert Hallock if he was what you might call an AI optimist. While he was quick to offer an “I am!” he’s by no means a babe in the woods. His role as senior director of technical marketing for AI at Intel may suggest a dyed-in-the-wool evangelist for all things AI, but Hallock had to be convinced of its capabilities like any other.
“I had the same initial reaction a lot of people had when artificial generalized intelligence came onto the scene (the cloud stuff), I was a little skeptical.”
However, with a career that spans years of technical marketing for processors and GPUs, not to mention stints in journalism as a PC hardware reviewer and technical writer, Hallock has a nose fine-tuned for snake oil — so when he says AI is the key to improving performance and power efficiency in Intel’s new wave of hardware, it’s hard not to buy into that notion.
"I quickly realized there's a lot going on in a different kind of AI — you could call it assistive AI or offline AI. [At Intel] We’re just trying to make computers faster, more power efficient, and AI is the new face of that."
Still, the term AI is broad and carries plenty of baggage and skepticism.
“I think that negativity is not unwarranted, it's also very confusing sometimes because the word AI means a lot of things.
"Are we talking about the AI toaster that has no AI? The cloud kind — which, who knows what happens to your info once you submit it into the chat box? Or are we talking about the offline kind, where you're just running another program on your PC?
"They're not all the same thing, and so I'm an optimist on the ‘assistive local runs on your PC, helps with performance and power.’ I'm big on that kind, and not so big on the other kind.”
Have you seen AI used in a way that you find uncomfortable or worrying?
“Every technology has its bad actors … Don’t be a jerk with AI.”
Since generative AI has broken into the mainstream, it seems like almost anything is possible thanks to these powerful predictive models. Sadly, that potential can be used for good and bad outcomes.
“Obviously, the generative deep fakes are a little concerning. But on the same token, I also think back to the first non-gaming use of graphics cards, which unfortunately, was to crack Wi-Fi passwords. Every technology has its bad actors.
“I think it is incumbent upon companies like Intel to continue promoting ethical and responsible uses of AI — use it to help your email or to write a document, don't be a jerk with AI.”
However, it’s not always the outcome of a model that can be used in bad faith. For AI image generators specifically, a pretty pungent stink is being kicked up online around how this kind of software is trained.
While OpenAI CEO Sam Altman thinks granting creators and artists a share of revenue for creative AI art in their style would be “cool,” Hallock shares how Intel is already taking steps to right certain perceived wrongs against the creative community through AI.
“There's a company we work with and invest in called Bria, and they're a generative AI model where artists get paid for the work that goes into the final image.
"If your picture was 23% of the final image … you get paid 23% of the licensing fees.”
Rael / Laptop Mag: It's sort of the ethical cocoa farming route to pay the artist when it comes to their training data, right?
“Yeah, we've had ethical food, ethical data, ethical everything for a long time, right? That kind of data is not new in the history of PCS. It tangles up with old piracy conversations, for example, but it's just the new face of it.”
What will bring people around to AI's potential?
“A lot of users may not even know that AI is running on their machine, but they're benefiting from it.”
AI might be the shiny new jewel of hardware manufacturers and developers, but many in the general public retain a level of skepticism.
But Hallock believes adoption will happen in waves — starting with the early adopters experimenting with clients like LM Studio, Ollama, and Intel’s AI Playgound, gradually followed by enterprise, content creators, and office workers.
Though PC enthusiasts, of which Hallock is himself a lifelong guild member, might take a little more convincing.
“I think PC enthusiasts will probably not be convinced until there's a major game that benefits from an AI technology of some kind, and I don't mean super sampling … I mean procedural texture generation or voice generation.”
Those undefined by previous categories? They may not even be aware of the transition, but they’ll benefit all the same.
“By 2028, about 80 percent of all computers are going to have dedicated AI accelerators — desktop, notebook, etc.
"A lot of users may not even know that AI Is running on their machine, but they're benefiting from it; they get the performance out of it, they get extra features, they get extra energy, efficiency, and longer battery life.”
Rael / Laptop Mag: Osmosis, essentially? You have the enthusiast, office worker, and content creator who will interact with AI more directly. But, before you know it, it's the new normal?
“That’s my expectation, yeah. This has happened before in the PC industry, too. Many people forget that Graphics have only been inside the CPU for about 20 years. A CPU just used to be a processor.
“Then this little graphics chip comes along, and it was widely ridiculed by users and the media, including myself. I was in the media at the time. I did not see the point.
“Now … everything is GPU-accelerated, but it happened very quietly.
“There was no big banner that said ‘Tada! Congratulations, your integrated GPU is now worth something.’ It just happened gradually over the course of three to five years, and we're all better for it. AI’s going to do the same thing.”
Is there anything in the AI pipeline that might catch people off guard in the short term?
“The potential for memes and tomfoolery is unbounded.”
I asked Robert what upcoming AI developments might sneak up on people. His answer? The next 12 to 18 months are going to be surprisingly lively — with multimodal models and video generation stepping into the spotlight.
“We have the Sora model from OpenAI. We can actually run that on Lunar Lake. Granted, it's a very low-resolution output … but a year ago, creating video was unthinkable — the potential for memes and tomfoolery is unbounded.”
But it’s another possibility that seems to really stoke the embers of excitement in Hallock.
“What if you had a speech-to-text or text input model in your application that would allow people to speak to the tool and use it by request … We've never before had the ability to go, ‘Hey, make me this,’ and the tool will use itself.
“This is pretty awesome for accessibility. Not even for people who have physical limitations … but for anyone who wants to use a more advanced piece of software and get started really quick. I think this is a very cool way to use AI, and it's starting to be explored.”
Rael / Laptop Mag: I don’t know anybody who’s ever hit F1 for help and found what they’ve needed.
“It doesn’t work! It doesn’t work.”
That brought us to a real world example: Intel’s Project SuperBuilder/AI Assistant Builder, which won Laptop Mag’s award for Best AI Debut at MWC 2025 in March.
The platform allows for the creation of chatbots trained on custom data. One OEM had even converted a user manual into a locally running chatbot in the first steps of revolutionizing modern tech support.
“That’s right! We came by it genuinely, actually, because we use tools like that inside Intel to help access information.
“It's shockingly accurate ... I've only been at Intel 18 months, so there's a lot that I don't know about … which would be very hard to digest without this tool. I would just have to know the right person who has these documents. Good luck. There are about 100,000 people here.
“That kind of information access is very, very powerful. AI Assistant Builder allows people to roll their own AI chatbots with access to external documents … PDFs, documents, emails, and websites if you want. And that's part of this platform that we're building for businesses.”
What is distracting us from AI's real potential and how it can impact people's lives?
"There is a bubble coming. That's for sure."
Not all AI applications are created equal, and some, according to Hallock, might be actively pulling focus from where the tech could be making a meaningful impact.
“I think the big one is generating images. I mean, it's obvious it’s fun, but it's either for memes or for trolling, basically … Though it is fun, it's a lot of compute we're spending on memes. There's other stuff we can be doing.”
More seriously, Hallock sees the marketing appropriation of AI as an equally damaging distraction, with the techno-acronym slapped on everything from dog collars to dishwashers..
“There are a million things being called AI. Some are, some are not. It is understandable that many people would just check out.”
The inflation of the term AI in consumer spaces has, in his opinion, launched us to the top of the Gartner Hype Cycle, to the peak of inflated expectations. Beyond it? The ominously named trough of disillusionment.
“You'll probably start to hear chatter soon about this AI bubble. And we're talking about culling the players who aren't really doing AI because eventually people will realize that this is not what it says it is on the can.
“What will be left is the industry players, the software makers, and hardware companies — including Intel, Nvidia, Qualcomm, and AMD.
“This is all legit, serious AI hardware for consumers. I don't know about the other stuff.
“There is a bubble coming. That's for sure.”
Is there a killer app for AI on the horizon?
"Let me ask you this: What’s the killer app for an x86 processor?"
I asked Robert if there’s a surefire way to draw people back from Gartner’s trough of disillusionment.
“That's a really difficult square to move people off of, and the only remedy is a killer app of some kind.
“For every holdout, we have to find that one killer app, and that might take years. But that's okay, that's part of the process.”
And it was around then that I fell into what felt like a perfectly laid trap.
Rael / Laptop Mag: Is that what Intel foresees? That the killer app is coming, but it's not here yet?
“Actually, our real view is that there's no such thing as the killer app. Let me ask you this: What is the killer app for an x86 processor?”
I wrestled with the idea of trying to justify 3D Pinball: Space Cadet as a viable answer, but… Yeah, point taken. Well played, Rob.
“Everybody's going to have a different answer, and so AI, like every other Computing engine before it, is only going to be successful with volume — the number of apps and models that people have access to needs to be hundreds, thousands of choices to hit everyone.”
That kind of scale sounds overwhelming at first, like a bloated tech utopia just waiting to collapse under its own weight. But the way Hallock tells it, it sounds more like the early internet. A time before Instagram, Facebook, and YouTube, when you could stumble across a website built entirely for the art of practicing the underwater jazz flute and feel like you’d found your people.
Is AI on a similar trajectory?
“That's what Intel is working on. We are working overtime to get hundreds of models running. We have 500-plus now working, we're on track for something like 450 AI-driven features this year. It was 12 in August 2023, and now we're at 450.”
For those already sitting in that trough of disillusionment, or yet to see the appeal, Hallock gets it. But he’s betting that function will win the day, especially now it’s baked into the very hardware that powers your Intel machine.
“If nothing else, for these users who are super resistant and reluctant about AI, just understand this is a major part of performance and power in a CPU going forward. And if you care about that, then you have to care about AI too. Even if there isn't the right app for you yet, there will be.”
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
