logo
How to Buy the Best Keycaps for Your Mechanical Keyboard

How to Buy the Best Keycaps for Your Mechanical Keyboard

WIRED24-03-2025

Here's everything you need to know—from keycap sculpts to materials—to make your typing experience more personalized. Photograph: Henri Robbins
If you buy something using links in our stories, we may earn a commission. This helps support our journalism. Learn more. Please also consider subscribing to WIRED
Whether you're buying your first mechanical keyboard or doing research for your endgame build, keycaps are a huge part of not only your keyboard's appearance but also its all-around typing experience. The keycaps you choose can have a massive impact on how you type, what it feels like, and what your keyboard sounds like. That's why it's important to find the right set for your personal needs.
However, there are a lot of different details to consider when looking at keycaps, and it can be overwhelming if you're new to the hobby or out of the loop on certain developments. I've put together this guide to serve as a jumping-off point into the world of keycaps. Here's everything you need to know.
Power up with unlimited access to WIRED . Get best-in-class reporting that's too important to ignore for just $2.50 $1 per month for 1 year. Includes unlimited digital access and exclusive subscriber-only content. Subscribe Today .
The most important part of any keycap set is to make sure the keycaps will fit your keyboard . If not, you'll either be stuck not using the set at all, or mixing your brand-new keycaps with old ones for full coverage. To start, here's what you'll want to look for.
Make sure your keycaps fit your switches! Most modern keyboards use standard MX-style switches with a plus-shaped stem. These have become ubiquitous, and as a result, almost every keycap set offered today will be compatible with them. However, if you want to use keycaps from an older keyboard, there's a good chance they won't fit a newer one. Photograph: Henri Robbins
The three most common 'alternative' switches are Alps switches, which have a small minus-shaped stem, Topre (also referred to as EC) switches, which have a circular stem, and Kailh Choc low-profile switches, which have two prongs on the stem. These keycaps cannot be used on an MX-style keyboard, but any keycap listings will clearly label them.
Next, you'll want to check layouts. While most keyboards have adopted fairly standardized layouts today, a few stand out. The best modern example is the HHKB Studio, which has unique G, H, and B keys, plus a shorter-than-average spacebar, none of which are included in the 'standard' keycap set. Some keycap sets will have specific kits to cover more unusual layouts though.
Make sure your keycap set has full coverage. While most sets today include everything needed for a full-size keyboard, I have seen an influx of keycap sets that sell the number pad separately in exchange for a lower base price, and the same has even been the case for arrows and function row keys, particularly on budget-oriented (or customizable) keycap sets.
After you've figured out what kind of keycaps you need for your keyboard, you should start thinking about what shape you want these keycaps to be. The shape of your keycaps will have the largest impact on how typing feels, and on how comfortable typing is. The most basic aspects of this are the height of the keycap, the shape it is sculpted in, and the presence of different sculpted rows.
But first, it's worth noting that there is no standardized naming convention for keycap profiles—the names themselves are arbitrary and are determined entirely by the keycap manufacturer. Some have logical names (like how OSA keycaps are a fusion of OEM—original equipment manufacturer—and SA—spherical all—sculpts), while others are a bit silly (PBS stands for Penguin Belly Slide).
A keycap's height is the most immediately obvious aspect of its profile. Typically, keycaps are sorted into high-profile, medium-profile, and low-profile categories, which are all determined based on the height they extend past the top of the switch. The most standard height is the Cherry profile, which most people have likely typed on before. These sit a bit above the outer bezel of the keyboard, with the highest point sitting at 9.8 mm and the lowest point at 7.3 mm, and can be easily used without a wrist rest.
On the other side of the spectrum, SA keycaps (which you may have seen on the computers in Severance ) have a high point of almost 17 mm and a low point of 12.6 mm. The difference between the tallest and shortest keycaps is more drastic, and the average height is typically higher, resulting in a deeper typing sound and a greater need for a wrist rest. Photograph: Henri Robbins
Some keycap profiles, like Cherry and SA, will have different keys depending on the row they are in. These rows are typically denoted as R1 (row 1), R2, and R3. R4 (or sometimes R5) sits at the very top of the keyboard, and R1 is at the bottom. These separate rows will typically have different heights and angles, which can make reaching the top rows easier and make typing more comfortable. However, some people prefer a uniform layout where every row of keys has an identical height and shape. There aren't many uniform profiles available, but the most prominent—SA and PBS—are still fairly popular and easy to find in multiple colors.
The sculpt of the keycap also indicates the shape of the keycap's face. The two most popular sculpts are cylindrical and spherical. Cylindrical keycaps will have a U-shaped indentation with peaks on either side and a dip in the center, while spherical keycaps are going to be more bowl-like, with peaks on all four corners and a divot in the center. Popular keyboard maker Keychron typically ships its keycaps with OSA-profile spherical keycaps, while most keyboards from Razer, Corsair, and other big names use Cherry or OEM profile keycaps with a cylindrical sculpt.
Some keycap profiles are perfectly flat, although these are uncommon and typically only used on keyboards with low-profile switches and/or ortho-linear layouts.
A lot of different materials are used to make keycaps. These are some of the most common ones you can find in full keycap sets:
ABS: This is the most common material for a keyboard's original keycaps to be made in. ABS (acrylonitrile butadiene styrene) plastic is everywhere in your life, from LEGO sets to car interiors, and it comes in countless different finishes, textures, and levels of quality. While the thinner ABS plastic used in budget keyboards may not be high quality, some ABS keycaps, like those from GMK and Signature Plastics, are considered to be the best of any material. Regardless of quality, ABS plastic will almost inevitably shine over time the more you type on it—some people avoid it for this, while others see having 'shined' keycaps as a badge of honor. ABS is preferred for its bright and vibrant colors, durability, and smooth feel when typing. ABS also creates a higher-pitched sound when typing compared to other keycap materials.
PBT: A large percentage of aftermarket 'custom' keycaps are made from PBT (polybutylene terephthalate). This material is often a higher-end plastic that focuses on thickness and rigidity while having a more textured top when compared to ABS. Colors on PBT keycaps are often not as vibrant, although the difference between ABS and PBT today is much less significant than it used to be. One of the largest upsides of PBT is that it doesn't easily shine over time, even after decades of use in point-of-sale systems, meaning your keycaps will always look as good as the day you got them. PBT typically creates a deeper typing signature than ABS.
Ceramic: Ceramic keycaps are one of the most recent developments in the keycap world. They are primarily made by Cerakey, a relatively new brand from China. These keycaps provide a deeper typing sound some people describe as marbly. When I used ceramic keycaps previously, I found them to break easily because of their brittle ceramic stems. However, modern ceramic keycaps use a plastic stem to avoid this issue. The largest benefits of a keycap like this are the typing feel and sound: The keycaps themselves are smooth and slick to type on, and they create a deeper and more resonant typing sound compared to plastic keycaps.
Metal: While individual metal keycaps have been around for a long time, full metal keycap sets only became popular in recent years as interest in custom mechanical keyboards increased. A full set of metal keycaps will typically cost upwards of $200, and installing them may require some supporting upgrades to your keyboard such as a heavier switch underneath the spacebar or additional sound dampening to reduce metal pinging when typing.
Awekeys Copper Eagle Full Metal Keycap Set Photograph: Henri Robbins
In writing this article, I tested two sets of metal keycaps: One from NovelKeys, and the other from Awekeys. I had minor issues with fit and finish in both kits, although the NovelKeys set was generally clean and free of manufacturing errors on visible surfaces. On the other hand, the Awekeys set had multiple keys with visible pitting and surface imperfections.
My largest issue with both was one that seemed to be unavoidable: Due to the extreme precision required to manufacture metal keycaps and the relatively lenient manufacturing tolerances of plastic switches, some keycaps in both sets had trouble fitting onto switches, either being too tight or too loose. With the Awekeys set, some were so loose that I could flip the keyboard over or release a key after pressing it down and they would come off. The company recommended using an included tool to bend the keycap stem back into place, which was successful but somewhat nerve-wracking, especially on such an expensive set of keycaps. In my experience, the stems were stiff enough that it was necessary to use a hammer in conjunction with the tool to adjust the stems properly.
Metal keycaps will feel more substantial than plastic keycaps, but will also create a higher-pitched typing sound than most other keycaps, especially in a keyboard with a metal plate. I found one of the main appeals of these keycaps, the cold and metallic feel when typing, quickly dissipates as the keycaps absorb heat from your fingers, and after a few minutes of typing, they are hard to differentiate from smooth plastic. However, their density can make a keypress feel slightly more crisp and immediate compared to a standard keycap, especially on a lighter tactile switch.
One of the most important parts of a set of keycaps is the text on top of them. Even if you already know where your keys are, they bring a sense of character and style to the keyboard, whether you prefer them centered, oriented to one side, or not there at all.
The text on top of keycaps is referred to as 'legends,' and the secondary text on a keycap is the 'sub-legend.' The main legends will denote the primary function of the keys, while sub-legends will indicate a secondary function, a different language, or a callback to another keyboard with interesting sub-legends (like Space Cadet keycaps). There are countless details and specifics on how the legends can be designed and why they may look a certain way, but it ultimately comes down to personal preference and use case. What's more important here is how the text is put onto the keycaps. Photograph: Henri Robbins
Doubleshot: Often considered the gold standard of keycap legends, doubleshot legends will never fade or wear off. This is because the legend is made from plastic instead of being printed onto the keycap. This is done through a two-part injection molding process where the legend is cast in one color of plastic then the rest of the keycap is molded around it, resulting in a legend that goes all the way through the keycap instead of eventually wearing away as the keycap is worn down by typing. The largest benefits of these keycaps are the crispness of the legends, their durability, and the ability to easily combine high-contrast colors, but the cost of making new molds means the options for different fonts and icons are limited compared to printed keycaps, and high-end doubleshot keycaps are often more expensive than other options. While doubleshot legends were previously only available for ABS keycaps, they are now available with PBT keycaps too.
Tripleshot: For keycaps with differently colored sub-legends, a tripleshot keycap is a great option that is incredibly hard to find and more expensive than doubleshot legends. Only a few manufacturers currently produce tripleshot keycaps (such as Domikey and PBTfans), and they are limited almost entirely to Japanese sub-legends. Photograph: Henri Robbins
Dye sublimation: This is the most common printing method for PBT keycaps. Dye sublimation can fully impress the dye into the keycap instead of having it sit on top, much like dyeing a piece of clothing, creating a finish that will rarely wear away. This process can also produce incredibly vibrant and precise legends, which has allowed for fun and unusual keycap sets like RGBEnglish and 21KB's Rainbow Road keycaps. While dye sublimation can last a long time, it can have some downsides. The quality of the legends can vary greatly, with higher-end keycaps looking indistinguishable from doubleshot keycaps, but less expensive options appearing fuzzy or blurry. In addition, dye sublimation—much like any other dyeing process—can usually only apply colors darker than the base material. This is why most dye-sublimated keycaps will either have a lighter base color or will only print black on top of darker keycaps.
Reverse dye sublimation: Reversing this process allows for the entire keycap except for the legends to be dyed, which can create dark keycaps with light legends. However, this process is more difficult than standard sublimation, with mixed results over the years. Some sets of reverse dye-sub keycaps will look fantastic, but other times they will have fading around the edges, inconsistent color coverage, or issues with the sharpness of the legends. In addition, the printing is more susceptible to scratches than raw PBT. However, when done right, reverse dye-sub can create unique color schemes not otherwise possible with dye sublimation.
UV printing: While not as common today as it used to be, UV printing is still popular for creating vibrant, multicolored text on keycaps. UV printing is primarily found on GMK keycaps with unique or multicolored sub-legends, where it is used to add color or design to a key without the need to manufacture additional molds. This printing method has a higher likelihood of fading over time, and can potentially have a slightly raised feel when typing but it has advanced enough in recent years that any well-made set of keycaps with UV printing will still work well for regular use. UV printing is also commonly used for transparent keycaps with solid legends.
Pad printing: This is the method basically all budget-oriented office keyboards today use. In this method, the prints are applied directly onto the keycaps using a silicone pad and sit on top of the keycaps. Inexpensive pad-printed keycaps will typically have subtle outlines around the legends and will wear away quickly compared to any other method. However, this printing method is both high-resolution and inexpensive, hence its popularity.
Generally, there are three primary methods of sale for keycaps: In-stock sales, group buys, and preorders. Each has benefits and downsides, and knowing the difference is important when planning your purchase.
In-stock sales: This is what most keycap sets will be sold as. These keycaps are readily available somewhere, sitting in a warehouse, and will ship to you after your order has been placed. This will be the most reliable and immediate way to get a set of keycaps.
Group buys: With group buys, you give your money to an individual or vendor who will pool all the customers' money together to order keycap sets directly from a manufacturer. This is how keycaps were purchased for years by enthusiasts, and it is typically the only way to get a guaranteed set of limited-run or small-batch keycaps. The group buy method is done to meet the Minimum Order Quantity (MOQ) for a manufacturer to produce a set of keycaps without the vendor needing to purchase all the keycap sets up front and hope they sell. The main concern with group buys is inconsistency: There have been multiple vendors recently who mismanaged money and were unable to fulfill their group buys. Ultimately, this comes down to vetting the vendors you're buying from and ensuring they have a good reputation within the community. Even then, there is still some risk involved.
Extra preorders: When placing an order for a group buy, a vendor will typically order a small number of extra keycaps to cover any issues that arise after the sale, and to have additional sets available for people who weren't able to join the group buy. These are typically called 'extras,' and will be made available for preorder once the manufacturer has confirmed how many sets they have produced. While these are never guaranteed, and will typically be more expensive than the initial group buy price, it can be a good way to get a set you missed out on after the fact.
When it comes to pricing, you can essentially pay as much as you want for a set of keycaps. However, some details can help determine or justify the price of a set. The largest determining factor is the manufacturer of a keycap set. Past that, the price is dependent on how many keycaps are in the set and how complicated the manufacturing of a set is.
Doubleshot ABS keycaps from high-end manufacturers like GMK or Signature Plastics will often run well over $100. This is primarily because these brands are well known for having outstanding color accuracy, durability, and quality control. They are the highest-quality keycaps you can get, but less-expensive keycaps will typically have 90 percent of the quality for 50 percent of the price.
Coming in around or below the $100 mark, PBT keycaps from larger manufacturers like EnjoyPBT will feature a thicker PBT keycap with high-resolution dye-sub printing.
With all of this, it can be overwhelming to figure out what keycaps you might want for your keyboard. There are a few starting places great for finding the perfect set of keycaps for you. Some of my preferred shops are:
Cannonkeys has a massive selection of in-stock PBT keycaps in various colors along with a more limited selection of ABS keycaps. The brand designs many of its sets in-house and offers quite a few sets with both retro and modern styling, taking inspiration from pop culture and vintage keyboards.
Novelkeys is another well-known name in the keyboard world, offering a large selection of ABS and PBT keycap designs. It regularly collaborates with keycap designers, but quite a few of its sets are only available through group buys or limited extras from said group buys.
NovelKeys Aluve Keycaps Photograph: Henri Robbins
21kb is a small manufacturer of beige dye-sub keycaps. While the beige base of these keycaps may not be the most attention-grabbing, the brand has countless options for fun, clever, or otherwise interesting legends that go onto those keycaps, along with the ability to custom-order an individual set to your specifications. These keycaps are some of the highest-quality PBT sets I've purchased, and I use them on my primary keyboard.
Drop is a huge name in the keyboard space, and it has done a lot to advance the world of keycaps. It has large offerings of standard keycaps in everyday color schemes, along with high-profile collaborations (like The Lord of the Rings keycaps) and countless exclusive sets from manufacturers like GMK and Signature Plastics.
Mekibo is a frequent vendor for keycap group buys. The company doesn't stock many keycaps, but I find that every set it offers will be well done and looks great on the right keyboard.
Omnitype offers a large selection of GMK keycaps, along with its in-house over-the-counter line of keycaps. It has been a large part of the keyboard community for years and continually introduces sets that bring something unique to the table.
Divinikey has a lot of keycaps, including PBT keycaps you can't find anywhere else. This is one of the best places to order PBTFans keycaps (Chinese-based KBDFans' line of keycaps), and there's a good number of sets from other international vendors as well.
Outside of these shops, there are a few resources to verify the legitimacy of keyboard vendors and get more recommendations. One of the largest is the MK Vendor Trust and Safety System made by keyboard YouTuber TaehaTypes. This system keeps track of keyboard vendors along multiple metrics to ensure products come from a reliable source.

Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

I replaced Wordle with my own AI-generated games — here's how to make yours
I replaced Wordle with my own AI-generated games — here's how to make yours

Tom's Guide

timea day ago

  • Tom's Guide

I replaced Wordle with my own AI-generated games — here's how to make yours

After months of the same Wordle routine, I decided to create my own word puzzle games using AI and honestly, I'm never going back. Instead of being limited to one puzzle per day, I now have an endless supply of custom games that actually challenge me at my skill level. You don't need any coding experience to make this work. Modern AI chatbots like ChatGPT, Claude, and Gemini can generate fully functional puzzle games in minutes. I tested all three to see which AI creates the most engaging alternatives to Wordle. Turns out, AI is surprisingly good at creating games that keep you entertained during coffee breaks, commutes, or whenever you need a quick mental challenge. Here's how to build your own collection of word games that goes way beyond the daily Wordle limit. Log into ChatGPT and use this simple prompt: "Create a Wordle-style game for us to play. I have 5 guesses to pick the right word." The response is straightforward and gets the job done, delivering a functional word-guessing game that you can play immediately in the chat interface. ChatGPT's approach is intuitive and user-friendly, making it easy to jump straight into playing without any setup. It generates a clean text-based version where you make guesses and receive clear feedback about correct letters and positions. The interface feels natural to use: you simply type your guess and ChatGPT responds with helpful hints about which letters are correct and where they belong. While it lacks the visual pizzazz of colored tiles or interactive elements that make Wordle so satisfying, ChatGPT compensates with clear, well-organized feedback that's easy to follow. This option works perfectly when you need a quick mental challenge during your coffee break and want something that just works without any fuss. Head to Claude and use the exact same prompt: "Create a Wordle-style game for us to play. I have 5 guesses to pick the right word." Claude takes a completely different approach, actually building a functional game interface right within the chat using its artifact system. This was easily my favorite of the three options. Claude creates a fully interactive game with visual feedback, colored tiles, and a proper game board that updates in real-time as you play. Using the mobile app makes this feel like you're actually playing a legitimate Wordle alternative rather than just chatting with an AI. The interface is clean, responsive, and captures that satisfying feeling of seeing your guesses populate the grid with the familiar green, yellow, and gray color coding. It's such a good alternative because you get the complete Wordle experience without being limited to just one puzzle per day. Log into Gemini and try the same prompt. You'll get the most stripped-down version of the three, with minimal visual elements and a very basic text-based interaction. While ChatGPT at least included some emojis to add visual interest, Gemini keeps things extremely plain. However, Gemini does offer one unique advantage — it actively suggests using its Canvas feature to generate actual code for a standalone game. This makes it potentially useful if you want to create something you can share with friends or host elsewhere. The main issue with Gemini's idea of Wordle-style game is it occasionally makes errors that can break the game experience, like incorrectly stating that "fruit" isn't a five-letter word, for example. Once you've gotten past the hallucinations, Gemini played well and did exactly what it was supposed to do, delivering a functional word game that captures the core Wordle mechanics just without the frills. No surprise there. Claude delivered the most polished, interactive experience that actually felt like playing a real game rather than just having a conversation about one. ChatGPT and Gemini both provided functional alternatives, but Claude's ability to create a proper visual interface with real-time updates puts it in a league of its own for this kind of task. Honestly, Claude is criminally underrated as an AI assistant. It consistently delivers more thoughtful, nuanced responses and excels at creative tasks like this game creation challenge. While everyone's talking about ChatGPT, Claude quietly outperforms in areas that actually matter for everyday use and I've been quite vocal about how it's my favorite AI model out there right now. The fact that it can build interactive experiences right within the chat interface is just another example of how it's much more capable than people realize. Now you've learned how to create your own wordle-style game in ChatGPT, Claude and Gemini, why not take a look at our other useful AI guides?Check out 5 mind-blowing ChatGPT image prompts you'll wish you knew sooner and how to choose the right ChatGPT model for any task. Anthropic keeps taking Claude from strength to strength, make sure you explore Claude's latest feature: voice mode. Get instant access to breaking news, the hottest reviews, great deals and helpful tips.

Demis Hassabis On The Future of Work in the Age of AI
Demis Hassabis On The Future of Work in the Age of AI

Yahoo

time2 days ago

  • Yahoo

Demis Hassabis On The Future of Work in the Age of AI

WIRED Editor At Large Steven Levy sits down with Google DeepMind CEO Demis Hassabis for a deep dive discussion on the emergence of AI, the path to Artificial General Intelligence (AGI), and how Google is positioning itself to compete in the future of the workplace. Director: Justin Wolfson Director of Photography: Christopher Eusteche Editor: Cory Stevens Host: Steven Levy Guest: Demis Hassabis Line Producer: Jamie Rasmussen Associate Producer: Brandon White Production Manager: Peter Brunette Production Coordinator: Rhyan Lark Camera Operator: Lauren Pruitt Gaffer: Vincent Cota Sound Mixer: Lily van Leeuwen Production Assistant: Ryan Coppola Post Production Supervisor: Christian Olguin Post Production Coordinator: Stella Shortino Supervising Editor: Erica DeLeo Assistant Editor: Justin Symonds - It's a very intense time in the field. We obviously want all of the brilliant things these AI systems can do, come up with new cures for diseases, new energy sources, incredible things for humanity. That's the promise of AI. But also, there are worries if the first AI systems are built with the wrong value systems or they're built unsafely, that could be also very bad. - Wired sat down with Demis Hassabis, who's the CEO of Google DeepMind, which is the engine of the company's artificial intelligence. He's a Nobel Prize winner and also a knight. We discussed AGI, the future of work, and how Google plans to compete in the age of AI. This is "The Big Interview." [upbeat music] Well, welcome to "The Big Interview," Demis. - Thank you, thanks for having me. - So let's start talking about AGI a little here. Now, you founded DeepMind with the idea that you would solve intelligence and then use intelligence to solve everything else. And I think it was like a 20-year mission. We're like 15 years into it, and you're on track? - I feel like, yeah, we're pretty much dead on track, actually, is what would be our estimate. - That means five years away from what I guess people will call AGI. - Yeah, I think in the next five to 10 years, that would be maybe 50% chance that we'll have what we are defined as AGI, yes. - Well, some of your peers are saying, "Two years, three years," and others say a little more, but that's really close, that's really soon. How do we know that we're that close? - There's a bit of a debate going on in the moment in the field about definitions of AGI, and then obviously, of course, dependent on that. There's different predictions for when it will happen. We've been pretty consistent from the very beginning. And actually, Shane Legg, one of my co-founders and our chief scientist, you know, he helped define the term AGI back in, I think, early 2001 type of timeframe. And we've always thought about it as system that has the ability to exhibit, sort of all the cognitive capabilities we have as humans. And the reason that's important, the reference to the human mind, is the human mind is the only existence proof we have. Maybe in the universe, the general intelligence is possible. So if you want to claim sort of general intelligence, AGI, then you need to show that it generalizes to all these domains. - Is when everything's filled in, all the check marks are filled in, then we have it- - Yes, so I think there are missing capabilities right now. You know, that all of us who have used the latest sort of LLMs and chatbots, will know very well, like on reasoning, on planning, on memory. I don't think today's systems can invent, you know, do true invention, you know, true creativity, hypothesize new scientific theories. They're extremely useful, they're impressive, but they have holes. And actually, one of the main reasons I don't think we are at AGI yet is because of the consistency of responses. You know, in some domains, we have systems that can do International Math Olympiad, math problems to gold medal standard- - Sure. - With our AlphaFold system. But on the other hand, these systems sometimes still trip up on high school maths or even counting the number of letters in a word. - Yeah. - So that to me is not what you would expect. That level of sort of difference in performance across the board is not consistent enough, and therefore shows that these systems are not fully generalizing yet. - But when we get it, is it then like a phase shift that, you know, then all of a sudden things are different, all the check marks are checked? - Yeah. - You know, and we have a thing that can do everything. - Mm-hmm. - Are we then power in a new world? - I think, you know, that again, that is debated, and it's not clear to me whether it's gonna be more of a kind of incremental transition versus a step function. My guess is, it looks like it's gonna be more of an incremental shift. Even if you had a system like that, the physical world, still operates with the physical laws, you know, factories, robots, these other things. So it'll take a while for the effects of that, you know, this sort of digital intelligence, if you like, to really impact, I think, a lot of the real world things. Maybe another decade plus, but there's other theories on that too, where it could come faster. - Yeah, Eric Schmidt, who I think used to work at Google, has said that, "It's almost like a binary thing." He says, "If China, for instance, gets AGI, then we're cooked." Because if someone gets it like 10 minutes, before the next guy, then you can never catch up. You know, because then it'll maintain bigger, bigger leads there. You don't buy that, I guess. - I think it's an unknown. It's one of the many unknowns, which is that, you know, that's sometimes called the hard takeoff scenario, where the idea there is that these AGI systems, they're able to self-improve, maybe code themselves future versus themselves, that maybe they're extremely fast at doing that. So what would be a slight lead, let's say, you know, a few days, could suddenly become a chasm if that was true. But there are many other ways it could go too, where it's more incremental. Some of these self-improvement things are not able to kind of accelerate in that way, then being around the same time, would not make much difference. But it's important, I mean, these issues are the geopolitical issues. I think the systems that are being built, they'll have some imprint of the values and the kind of norms of the designers and the culture that they were embedded in. - [Steven] Mm-hmm. - So, you know, I think it is important, these kinds of international questions. - So when you build AI at Google, you know, you have that in mind. Do you feel competitive imperative to, in case that's true, "Oh my God, we better be first?" - It's a very intense time at the moment in the field as everyone knows. There's so many resources going into it, lots of pressures, lots of things that need to be researched. And there's sort of lots of different types of pressures going on. We obviously want all of the brilliant things that these AI systems can do. You know, I think eventually, we'll be able to advance medicine and science with it, like we've done with AlphaFold, come up with new cures for diseases, new energy sources, incredible things for humanity, that's the promise of AI. But also there are worries both in terms of, you know, if the first AI systems are built with the wrong value systems or they're built unsafely, that could be also very bad. And, you know, there are at least two risks that I worry a lot about. One is, bad actors in whether it's individuals or rogue nations repurposing general purpose AI technology for harmful lens. And then the second one is, obviously, the technical risk of AI itself. As it gets more and more powerful, more and more agentic, can we make sure the guardrails are safe around it? They can't be circumvented. And that interacts with this idea of, you know, what are the first systems that are built by humanity gonna be like? There's commercial imperative- - [Steven] Right. - There's national imperative, and there's a safety aspect to worry about who's in the lead and where those projects are. - A few years ago, the companies were saying, "Please, regulate us. We need regulation." - Mm-hmm, mm-hmm. - And now, in the US at least, the current administration seems less interested in putting regulations on AI than accelerating it so we can beat the Chinese. Are you still asking for regulation? Do you think that that's a miss on our part? - I think, you know, and I've been consistent in this, I think there are these other geopolitical sort of overlays that have to be taken into account, and the world's a very different place to how it was five years ago in many dimensions. But there's also, you know, I think the idea of smart regulation that makes sense around these increasingly powerful systems, I think is gonna be important. I continue to believe that. I think though, and I've been certain on this as well, it sort of needs to be international, which looks hard at the moment in the way the world is working, because these systems, you know, they're gonna affect everyone, and they're digital systems. - Yeah. - So, you know, if you sort of restrict it in one area, that doesn't really help in terms of the overall safety of these systems getting built for the world and as a society. - [Steven] Yeah. - So that's the bigger problem, I think, is some kind of international cooperation or collaboration, I think, is what's required. And then smart regulation, nimble regulation that moves as the knowledge about the research becomes better and better. - Would it ever reach a point for you where you would feel, "Man, we're not putting the guardrails in. You know, we're competing, that we really have to stop, or you can't get involved in that?" - I think a lot of the leaders of the main labs, at least the western labs, you know, there's a small number of them and we do all know each other and talk to each other regularly. And a lot of the lead researchers do. The problem is, is that it's not clear we have the right definitions to agree when that point is. Like, today's systems, although they're impressive as we discussed earlier, they're also very flawed. And I don't think today's systems, are posing any sort of existential risk. - Mm-hmm. - So it's still theoretical, but the problem is that a lot of unknowns, we don't know how fast those will come, and we don't know how risky they will be. But in my view, when there are so many unknowns, then I'm optimistic we'll overcome them. At least technically, I think the geopolitical questions could be actually, end up being trickier, given enough time and enough care and thoughtfulness, you know, sort of using the scientific method as we approach this AGI point. - That makes perfect sense. But on the other hand, if that timeframe is there, we just don't have much time, you know? - No, we don't. We don't have much time. I mean, we're increasingly putting resources into security and things like cyber, and also research into controllability and understanding of these systems, sometimes called mechanistic interpretability. You know, there's a lot of different sub-branches of AI. - Yeah, that's right. I wanna get to interpretability. - Yeah, that are being invested in, and I think even more needs to happen. And then at the same time, we need to also have societal debates more about institutional building. How do we want governance to work? How are we gonna get international agreement, at least on some basic principles, around how these systems are used and deployed and also built? - What about the effect on work on the marketplace? - Yeah. - You know, how much do you feel that AI is going to change people's jobs, you know, the way jobs are distributed in the workforce? - I don't think we've seen, my view is if you talk to economists, they feel like there's not much has changed yet. You know, people are finding these tools useful, certainly in certain domains- - [Steven] Yeah. - Like, things like AlphaFold, many, many scientists are using it to accelerate their work. So it seems to be additive at the moment. We'll see what happens over the next five, 10 years. I think there's gonna be a lot of change with the jobs world, but I think as in the past, what generally tends to happen is new jobs are created that are actually better, that utilize these tools or new technologies, what happened with the internet, what happened with mobile? We'll see if it's different this time. - Yeah. - Obviously everyone always thinks this new one, will be different. And it may be, it will be, but I think for the next few years, it's most likely to be, you know, we'll have these incredible tools that supercharge our productivity, make us really useful for creative tools, and actually almost make us a little bit superhuman in some ways in what we're able to produce individually. So I think there's gonna be a kind of golden era, over the next period of what we're able to do. - Well, if AGI can do everything humans can do, then it would seem that they could do the new jobs too. - That's the next question about like, what AGI brings. But, you know, even if you have those capabilities, there's a lot of things I think we won't want to do with a machine. You know, I sometimes give this example of doctors and nurses. You know, maybe a doctor and what the doctor does and the diagnosis, you know, one could imagine that being helped by AI tool or even having an AI kind of doctor. On the other hand, like nursing, you know, I don't think you'd want a robot to do that. I think there's something about the human empathy aspect of that and the care, and so on, that's particularly humanistic. I think there's lots of examples like that but it's gonna be a different world for sure. - If you would talk to a graduate now, what advice would you give to keep working- - Yeah. - Through the course of a lifetime- - Yeah. - You know, in the age of AGI? - My view is, currently, and of course, this is changing all the time with the technology developing. But right now, you know, if you think of the next five, 10 years as being, the most productive people might be 10X more productive if they are native with these tools. So I think kids today, students today, my encouragement would be immerse yourself in these new systems, understand them. So I think it's still important to study STEM and programming and other things, so that you understand how they're built, maybe you can modify them yourself on top of the models that are available. There's lots of great open source models and so on. And then become, you know, incredible at things like fine-tuning, system prompting, you know, system instructions, all of these additional things that anyone can do. And really know how to get the most out of those tools, and do it for your research work, programming, and things that you are doing on your course. And then come out of that being incredible at utilizing those new tools for whatever it is you're going to do. - Let's look a little beyond the five and 10-year range. Tell me what you envision when you look at our future in 20 years, in 30 years, if this comes about, what's the world like when AGI is everywhere? - Well, if everything goes well, then we should be in an era of what I like to call sort of radical abundance. So, you know, AGI solves some of these key, what I sometimes call root node problems in the world facing society. So a good one, examples would be curing diseases, much healthier, longer lifespans, finding new energy sources, you know, whether that's optimal batteries and better room temperature, superconductors, fusion. And then if that all happens, then we know it should be a kind of era of maximum human flourishing where we travel to the stars and colonize the galaxy. You know, I think the beginning of that will happen in the next 20, 30 years if the next period goes well. - I'm a little skeptical of that. I think we have an unbelievable abundance now, but we don't distribute it, you know, fairly. - Yeah. - I think that we kind of know how to fix climate change, right? We don't need a AGI to tell us how to do it, yet we're not doing it. - I agree with that. I think we being as a species, a society not good at collaborating, and I think climate is a good example. But I think we are still operating, humans are still operating in a zero-sum game mentality. Because actually, the earth is quite finite, relative to the amount of people there are now in our cities. And I mean, this is why our natural habitats, are being destroyed, and it's affecting wildlife and the climate and everything. - [Steven] Yeah. - And it's also partly 'cause people are not willing to accept, we do now to figure out climate. But it would require people to make sacrifices. - Yeah. - And people don't want to. But this radical abundance would be different. We would be in a finally, like, it would feel like a non-zero-sum game. - How will we get [indistinct] to that? Like, you talk about diseases- - Well, I gave you an example. - We have vaccines, and now some people think we shouldn't use it. - Let me give you a very simple example. - Sure. - Water access. This is gonna be a huge issue in the next 10, 20 years. It's already an issue. Countries in different, you know, poorer parts of the world, dryer parts of the world, also obviously compounded by climate change. - [Steven] Yeah. - We have a solution to water access. It's desalination, it's easy. There's plenty of sea water. - Yeah. - Almost all countries have a coastline. But the problem is, it's salty water, but desalination only very rich countries. Some countries do do that, use desalination as a solution to their fresh water problem, but it costs a lot of energy. - Mm-hmm. - But if energy was essentially zero, there was renewable free clean energy, right? Like fusion, suddenly, you solve the water access problem. Water is, who controls a river or what you do with that does not, it becomes much less important than it is today. I think things like water access, you know, if you run forward 20 years, and there isn't a solution like that, could lead to all sorts of conflicts, probably that's the way it's trending- - Mm-hmm, right. - Especially if you include further climate change. - So- - And there's many, many examples like that. You could create rocket fuel easily- - Mm-hmm. - Because you just separate that from seawater, hydrogen and oxygen. It's just energy again. - So you feel that these problems get solved by AGI, by AI, then we're going to, our outlook will change, and we will be- - That's what I hope. Yes, that's what I hope. But that's still a secondary part. So the AGI will give us the radical abundance capability, technically, like the water access. - Yeah. - I then hope, and this is where I think we need some great philosophers or social scientists to be involved. That should hopefully shift our mindset as a society to non-zero-sum. You know, there's still the issue of do you divide even the radical abundance fairly, right? Of course, that's what should happen. But I think there's much more likely, once people start feeling and understanding that there is this almost limitless supply of raw materials and energy and things like that. - Do you think that driving this innovation by profit-making companies is the right way to go? We're most likely to reach that optimistic high point through that? - I think it's the current capitalism or, you know, is the current or the western sort of democratic kind of systems, have so far been proven to be sort of the best drivers of progress. - Mm-hmm. - So I think that's true. My view is that once you get to that sort of stage of radical abundance and post-AGI, I think economics starts changing, even the notion of value and money. And so again, I think we need, I'm not sure why economists are not working harder on this if maybe they don't believe it's that close, right? But if they really did that, like the AGI scientists do, then I think there's a lot of economic new economic theory that's required. - You know, one final thing, I actually agree with you that this is so significant and is gonna have a huge impact. But when I write about it, I always get a lot of response from people who are really angry already about artificial intelligence and what's happening. Have you tasted that? Have you gotten that pushback and anger by a lot of people? It's almost like the industrial revolution people- - Yeah. - Fighting back. - I mean, I think that anytime there's, I haven't personally seen a lot of that, but obviously, I've read and heard a lot about, and it's very understandable. That's all that's happened many times. As you say, industrial revolution, when there's big change, a big revolution. - [Steven] Yeah. - And I think this will be at least as big as the industrial revolution, probably a lot bigger. That's surprising, there's unknowns, it's scary, things will change. But on the other hand, when I talk to people about the passion, the why I'm building AI- - Mm-hmm. - Which is to advance science and medicine- - Right. - And understanding of the world around us. And then I explain to people, you know, and I've demonstrated, it's not just talk. Here's AlphaFold, you know, Nobel Prize winning breakthrough, can help with medicine and drug discovery. Obviously, we're doing this with isomorphic now to extend it into drug discovery, and we can cure terrible diseases that might be afflicting your family. Suddenly, people are like, "Well, of course, we need that." - Right. - It'll be immoral not to have that if that's within our grasp. And the same with climate and energy. - Yeah. - You know, many of the big societal problems, it's not like you know, we know, we've talked about, there's many big challenges facing society today. And I often say I would be very worried about our future if I didn't know something as revolutionary as AI was coming down the line to help with those other challenges. Of course, it's also a challenge itself, right? But at least, it's one of these challenges that can actually help with the others if we get it right. - Well, I hope your optimism holds out and is justified. Thank you so much. - And I'll do my best. Thank you. [upbeat music]

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into the world of global news and events? Download our app today from your preferred app store and start exploring.
app-storeplay-store