The government has axed a scheme for upgrading energy efficiency in public sector buildings. The Public Sector Decarbonisation Scheme (PSDS) delivered more than £2.5bn in its first three phases for measures such as heat pumps, solar panels, insulation and double glazing, with further funding of nearly £1bn recently announced. But the Department for Energy Security and Net Zero (DESNZ) has told Building Design that the scheme has been dropped after the spending review, leaving uncertainty about how upgrades will be funded when the current phase expires in 2028. Source: UK Government/FlickrEd Miliband’s Department for Energy Security and Net Zero is responsible for the scheme The department said it would set out plans for the period after 2028 in due course. In a post on LinkedIn, Dave Welkin, director of sustainability at Gleeds, said he had waited for the release of the spending review with a “sense of trepidation” and was unable to find mention of public sector decarbonisation when Treasury documents were released. “I hoped because it was already committed in the Budget that its omission wasn’t ominous,” he wrote. Yesterday, he was told by Salix Finance, the non-departmental public body that delivers funding for the scheme, that it was no longer being funded. It comes after the withdrawal of funding for the Low Carbon Skills Fund (LCSF) in May. According to the government’s website, PSDS and LCSF were intended to support the reduction of emissions from public sector buildings by 75% by 2037, compared to a 2017 baseline. “Neither LCSF or PSDS were perfect by any means, but they did provide a vital source of funding for local authorities, hospitals, schools and many other public sector organisations to save energy, carbon and money,” Welkin said. “PSDS has helped replace failed heating systems in schools, keeping students warm. It’s replaced roofs on hospitals, helping patients recover from illness. It’s replaced windows in our prisons, improving security and stopping drugs getting behind bars.” However, responding to Welkin’s post, Steve Connolly, chief executive at Arriba Technologies, a low carbon heating and cooling firm, said that the scheme was being “mismanaged” with a small number of professional services firms “scooping up disproportionately large grants for their clients”. The fourth phase of the scheme was confirmed last September, with allocations confirmed only last month. This latest phase, which covers the financial years between 2025/26 and 2027/28, saw the distribution of £940m across the country. A DESNZ spokesperson said: “Our settlement is about investing in Britain’s renewal to create energy security, sprint to clean power by 2030, encourage investment, create jobs and bring down bills for good. “We will deliver £1bn in current allocations of the Public Sector Decarbonisation Scheme until 2028 and, through Great British Energy, have invested in new rooftop solar power and renewable schemes to lower energy bills for schools and hospitals across the UK. “We want to build on this progress by incentivising the public sector to decarbonise, so they can reap the benefits in lower bills and emissions, sharing best practice across government and exploring the use of repayable finance, where appropriate.” A government assessment of phase 3a and 3b projects identified a number of issues with the scheme, including delays and cost inflation, with more than a tenth being abandoned subsequent to grants being offered. Stakeholders interviewed for the report also identified “difficulties in obtaining skilled contractors and equipment”, especially air source heat pumps. The first come first served approach to awarding funding was also said to be “encouraging applicants to opt for more straightforward projects” and “potentially undermining the achievement of PSDS objective by restricting the opportunity for larger [and] more complex measures which may have delivered greater carbon reduction benefits”. But the consensus among stakeholders and industry representatives interviewed for the report was that the scheme was “currently key to sustaining the existing UK heat pump market” and that it was “seen as vital in enabling many public sector organisations to invest in heat decarbonisation”.

Outdated coding practices and memory-unsafe languages like C are putting software, including cryptographic libraries, at risk. Fortunately, memory-safe languages like Rust, along with formal verification tools, are now mature enough to be used at scale, helping prevent issues like crashes, data corruption, flawed implementation, and side-channel attacks. To address these vulnerabilities and improve memory safety, we’re rewriting SymCrypt (opens in new tab)—Microsoft’s open-source cryptographic library—in Rust. We’re also incorporating formal verification methods. SymCrypt is used in Windows, Azure Linux, Xbox, and other platforms. Currently, SymCrypt is primarily written in cross-platform C, with limited use of hardware-specific optimizations through intrinsics (compiler-provided low-level functions) and assembly language (direct processor instructions). It provides a wide range of algorithms, including AES-GCM, SHA, ECDSA, and the more recent post-quantum algorithms ML-KEM and ML-DSA.  Formal verification will confirm that implementations behave as intended and don’t deviate from algorithm specifications, critical for preventing attacks. We’ll also analyze compiled code to detect side-channel leaks caused by timing or hardware-level behavior. Proving Rust program properties with Aeneas Program verification is the process of proving that a piece of code will always satisfy a given property, no matter the input. Rust’s type system profoundly improves the prospects for program verification by providing strong ownership guarantees, by construction, using a discipline known as “aliasing xor mutability”. For example, reasoning about C code often requires proving that two non-const pointers are live and non-overlapping, a property that can depend on external client code. In contrast, Rust’s type system guarantees this property for any two mutably borrowed references. As a result, new tools have emerged specifically for verifying Rust code. We chose Aeneas (opens in new tab) because it helps provide a clean separation between code and proofs. Developed by Microsoft Azure Research in partnership with Inria, the French National Institute for Research in Digital Science and Technology, Aeneas connects to proof assistants like Lean (opens in new tab), allowing us to draw on a large body of mathematical proofs—especially valuable given the mathematical nature of cryptographic algorithms—and benefit from Lean’s active user community. Compiling Rust to C supports backward compatibility   We recognize that switching to Rust isn’t feasible for all use cases, so we’ll continue to support, extend, and certify C-based APIs as long as users need them. Users won’t see any changes, as Rust runs underneath the existing C APIs. Some users compile our C code directly and may rely on specific toolchains or compiler features that complicate the adoption of Rust code. To address this, we will use Eurydice (opens in new tab), a Rust-to-C compiler developed by Microsoft Azure Research, to replace handwritten C code with C generated from formally verified Rust. Eurydice (opens in new tab) compiles directly from Rust’s MIR intermediate language, and the resulting C code will be checked into the SymCrypt repository alongside the original Rust source code. As more users adopt Rust, we’ll continue supporting this compilation path for those who build SymCrypt from source code but aren’t ready to use the Rust compiler. In the long term, we hope to transition users to either use precompiled SymCrypt binaries (via C or Rust APIs), or compile from source code in Rust, at which point the Rust-to-C compilation path will no longer be needed. Microsoft research podcast Ideas: AI and democracy with Madeleine Daepp and Robert Osazuwa Ness As the “biggest election year in history” comes to an end, researchers Madeleine Daepp and Robert Osazuwa Ness and Democracy Forward GM Ginny Badanes discuss AI’s impact on democracy, including the tech’s use in Taiwan and India. Listen now Opens in a new tab Timing analysis with Revizor  Even software that has been verified for functional correctness can remain vulnerable to low-level security threats, such as side channels caused by timing leaks or speculative execution. These threats operate at the hardware level and can leak private information, such as memory load addresses, branch targets, or division operands, even when the source code is provably correct.  To address this, we’re extending Revizor (opens in new tab), a tool developed by Microsoft Azure Research, to more effectively analyze SymCrypt binaries. Revizor models microarchitectural leakage and uses fuzzing techniques to systematically uncover instructions that may expose private information through known hardware-level effects.   Earlier cryptographic libraries relied on constant-time programming to avoid operations on secret data. However, recent research has shown that this alone is insufficient with today’s CPUs, where every new optimization may open a new side channel.  By analyzing binary code for specific compilers and platforms, our extended Revizor tool enables deeper scrutiny of vulnerabilities that aren’t visible in the source code. Verified Rust implementations begin with ML-KEM This long-term effort is in alignment with the Microsoft Secure Future Initiative and brings together experts across Microsoft, building on decades of Microsoft Research investment in program verification and security tooling. A preliminary version of ML-KEM in Rust is now available on the preview feature/verifiedcrypto (opens in new tab) branch of the SymCrypt repository. We encourage users to try the Rust build and share feedback (opens in new tab). Looking ahead, we plan to support direct use of the same cryptographic library in Rust without requiring C bindings.  Over the coming months, we plan to rewrite, verify, and ship several algorithms in Rust as part of SymCrypt. As our investment in Rust deepens, we expect to gain new insights into how to best leverage the language for high-assurance cryptographic implementations with low-level optimizations.  As performance is key to scalability and sustainability, we’re holding new implementations to a high bar using our benchmarking tools to match or exceed existing systems. Looking forward  This is a pivotal moment for high-assurance software. Microsoft’s investment in Rust and formal verification presents a rare opportunity to advance one of our key libraries. We’re excited to scale this work and ultimately deliver an industrial-grade, Rust-based, FIPS-certified cryptographic library. Opens in a new tab

News Climate This paint ‘sweats’ to keep your house cool A new cement-based paint cools buildings by combining sunlight reflection and water evaporation A new cooling paint reflects sunlight and mimics sweat to reduce air-conditioning use. Marie LaFauci/Getty Images By Larissa G. Capella June 13, 2025 at 11:00 am A cool house without air conditioning may soon be possible. Scientists in Singapore have developed a new type of paint that reflects sunlight and cools surfaces by slowly evaporating water. Unlike other commercially available cooling paints, which are designed to repel water to protect the underlying material, the new one even works in hot, humid places, offering a low-energy way to stay cool, researchers report June 5 in Science. “The key is passive cooling,” which requires no energy input, says material scientist Li Hong In other words, it works without using electricity or mechanical systems. Right now, radiative cooling is the most common type of passive cooling used in materials, including certain paints. It works by reflecting sunlight and radiating heat from a surface such as walls or roofs, into the sky. But in humid places like Singapore, water vapor in the air traps heat near the surface, which prevents it from escaping into the atmosphere and keeps the surfaces warm. In response, Hong and two other material scientists from Nanyang Technological University developed a cement-based paint that combines three cooling strategies: radiative cooling, evaporative cooling, which our skin uses, and solar reflection. In the study, the scientists painted three small houses: one with regular white paint, one with commercial cooling paint that uses only radiative cooling and one with their new formula. After two years of sun and rain in Singapore, the first two paints had turned yellow. But “our paint was still white,” says coauthor Jipeng Fei. Unlike other colors, white helps materials maintain their high reflectivity and cooling performance. Sign up for our newsletter We summarize the week's scientific breakthroughs every Thursday.

Transcript [MUSIC]      [BOOK PASSAGE]   PETER LEE: “In ‘The Little Black Bag,’ a classic science fiction story, a high-tech doctor’s kit of the future is accidentally transported back to the 1950s, into the shaky hands of a washed-up, alcoholic doctor. The ultimate medical tool, it redeems the doctor wielding it, allowing him to practice gratifyingly heroic medicine. … The tale ends badly for the doctor and his treacherous assistant, but it offered a picture of how advanced technology could transform medicine—powerful when it was written nearly 75 years ago and still so today. What would be the Al equivalent of that little black bag? At this moment when new capabilities are emerging, how do we imagine them into medicine?”   [END OF BOOK PASSAGE]     [THEME MUSIC]     This is The AI Revolution in Medicine, Revisited. I’m your host, Peter Lee.    Shortly after OpenAI’s GPT-4 was publicly released, Carey Goldberg, Dr. Zak Kohane, and I published The AI Revolution in Medicine to help educate the world of healthcare and medical research about the transformative impact this new generative AI technology could have. But because we wrote the book when GPT-4 was still a secret, we had to speculate. Now, two years later, what did we get right, and what did we get wrong?     In this series, we’ll talk to clinicians, patients, hospital administrators, and others to understand the reality of AI in the field and where we go from here.   [THEME MUSIC FADES] The book passage I read at the top is from “Chapter 10: The Big Black Bag.”  In imagining AI in medicine, Carey, Zak, and I included in our book two fictional accounts. In the first, a medical resident consults GPT-4 on her personal phone as the patient in front of her crashes. Within seconds, it offers an alternate response based on recent literature. In the second account, a 90-year-old woman with several chronic conditions is living independently and receiving near-constant medical support from an AI aide.    In our conversations with the guests we’ve spoken to so far, we’ve caught a glimpse of these predicted futures, seeing how clinicians and patients are actually using AI today and how developers are leveraging the technology in the healthcare products and services they’re creating. In fact, that first fictional account isn’t so fictional after all, as most of the doctors in the real world actually appear to be using AI at least occasionally—and sometimes much more than occasionally—to help in their daily clinical work. And as for the second fictional account, which is more of a science fiction account, it seems we are indeed on the verge of a new way of delivering and receiving healthcare, though the future is still very much open.  As we continue to examine the current state of AI in healthcare and its potential to transform the field, I’m pleased to welcome Bill Gates and Sébastien Bubeck.   Bill may be best known as the co-founder of Microsoft, having created the company with his childhood friend Paul Allen in 1975. He’s now the founder of Breakthrough Energy, which aims to advance clean energy innovation, and TerraPower, a company developing groundbreaking nuclear energy and science technologies. He also chairs the world’s largest philanthropic organization, the Gates Foundation, and focuses on solving a variety of health challenges around the globe and here at home.  Sébastien is a research lead at OpenAI. He was previously a distinguished scientist, vice president of AI, and a colleague of mine here at Microsoft, where his work included spearheading the development of the family of small language models known as Phi. While at Microsoft, he also coauthored the discussion-provoking 2023 paper “Sparks of Artificial General Intelligence,” which presented the results of early experiments with GPT-4 conducted by a small team from Microsoft Research.    [TRANSITION MUSIC]   Here’s my conversation with Bill Gates and Sébastien Bubeck.  LEE: Bill, welcome.  BILL GATES: Thank you.  LEE: Seb …  SÉBASTIEN BUBECK: Yeah. Hi, hi, Peter. Nice to be here.  LEE: You know, one of the things that I’ve been doing just to get the conversation warmed up is to talk about origin stories, and what I mean about origin stories is, you know, what was the first contact that you had with large language models or the concept of generative AI that convinced you or made you think that something really important was happening?  And so, Bill, I think I’ve heard the story about, you know, the time when the OpenAI folks—Sam Altman, Greg Brockman, and others—showed you something, but could we hear from you what those early encounters were like and what was going through your mind?   GATES: Well, I’d been visiting OpenAI soon after it was created to see things like GPT-2 and to see the little arm they had that was trying to match human manipulation and, you know, looking at their games like Dota that they were trying to get as good as human play. And honestly, I didn’t think the language model stuff they were doing, even when they got to GPT-3, would show the ability to learn, you know, in the same sense that a human reads a biology book and is able to take that knowledge and access it not only to pass a test but also to create new medicines.  And so my challenge to them was that if their LLM could get a five on the advanced placement biology test, then I would say, OK, it took biologic knowledge and encoded it in an accessible way and that I didn’t expect them to do that very quickly but it would be profound.   And it was only about six months after I challenged them to do that, that an early version of GPT-4 they brought up to a dinner at my house, and in fact, it answered most of the questions that night very well. The one it got totally wrong, we were … because it was so good, we kept thinking, Oh, we must be wrong. It turned out it was a math weakness [LAUGHTER] that, you know, we later understood that that was an area of, weirdly, of incredible weakness of those early models. But, you know, that was when I realized, OK, the age of cheap intelligence was at its beginning.  LEE: Yeah. So I guess it seems like you had something similar to me in that my first encounters, I actually harbored some skepticism. Is it fair to say you were skeptical before that?  GATES: Well, the idea that we’ve figured out how to encode and access knowledge in this very deep sense without even understanding the nature of the encoding, …  LEE: Right.   GATES: … that is a bit weird.   LEE: Yeah.  GATES: We have an algorithm that creates the computation, but even say, OK, where is the president’s birthday stored in there? Where is this fact stored in there? The fact that even now when we’re playing around, getting a little bit more sense of it, it’s opaque to us what the semantic encoding is, it’s, kind of, amazing to me. I thought the invention of knowledge storage would be an explicit way of encoding knowledge, not an implicit statistical training.  LEE: Yeah, yeah. All right. So, Seb, you know, on this same topic, you know, I got—as we say at Microsoft—I got pulled into the tent. [LAUGHS]  BUBECK: Yes.   LEE: Because this was a very secret project. And then, um, I had the opportunity to select a small number of researchers in MSR [Microsoft Research] to join and start investigating this thing seriously. And the first person I pulled in was you.  BUBECK: Yeah.  LEE: And so what were your first encounters? Because I actually don’t remember what happened then.  BUBECK: Oh, I remember it very well. [LAUGHS] My first encounter with GPT-4 was in a meeting with the two of you, actually. But my kind of first contact, the first moment where I realized that something was happening with generative AI, was before that. And I agree with Bill that I also wasn’t too impressed by GPT-3.  I though that it was kind of, you know, very naturally mimicking the web, sort of parroting what was written there in a nice way. Still in a way which seemed very impressive. But it wasn’t really intelligent in any way. But shortly after GPT-3, there was a model before GPT-4 that really shocked me, and this was the first image generation model, DALL-E 1.  So that was in 2021. And I will forever remember the press release of OpenAI where they had this prompt of an avocado chair and then you had this image of the avocado chair. [LAUGHTER] And what really shocked me is that clearly the model kind of “understood” what is a chair, what is an avocado, and was able to merge those concepts.  So this was really, to me, the first moment where I saw some understanding in those models.   LEE: So this was, just to get the timing right, that was before I pulled you into the tent.  BUBECK: That was before. That was like a year before.  LEE: Right.   BUBECK: And now I will tell you how, you know, we went from that moment to the meeting with the two of you and GPT-4.  So once I saw this kind of understanding, I thought, OK, fine. It understands concept, but it’s still not able to reason. It cannot—as, you know, Bill was saying—it cannot learn from your document. It cannot reason.   So I set out to try to prove that. You know, this is what I was in the business of at the time, trying to prove things in mathematics. So I was trying to prove that basically autoregressive transformers could never reason. So I was trying to prove this. And after a year of work, I had something reasonable to show. And so I had the meeting with the two of you, and I had this example where I wanted to say, there is no way that an LLM is going to be able to do x.  And then as soon as I … I don’t know if you remember, Bill. But as soon as I said that, you said, oh, but wait a second. I had, you know, the OpenAI crew at my house recently, and they showed me a new model. Why don’t we ask this new model this question?   LEE: Yeah. BUBECK: And we did, and it solved it on the spot. And that really, honestly, just changed my life. Like, you know, I had been working for a year trying to say that this was impossible. And just right there, it was shown to be possible.   LEE: [LAUGHS] One of the very first things I got interested in—because I was really thinking a lot about healthcare—was healthcare and medicine.  And I don’t know if the two of you remember, but I ended up doing a lot of tests. I ran through, you know, step one and step two of the US Medical Licensing Exam. Did a whole bunch of other things. I wrote this big report. It was, you know, I can’t remember … a couple hundred pages.   And I needed to share this with someone. I didn’t … there weren’t too many people I could share it with. So I sent, I think, a copy to you, Bill. Sent a copy to you, Seb.   I hardly slept for about a week putting that report together. And, yeah, and I kept working on it. But I was far from alone. I think everyone who was in the tent, so to speak, in those early days was going through something pretty similar. All right. So I think … of course, a lot of what I put in the report also ended up being examples that made it into the book.  But the main purpose of this conversation isn’t to reminisce about [LAUGHS] or indulge in those reminiscences but to talk about what’s happening in healthcare and medicine. And, you know, as I said, we wrote this book. We did it very, very quickly. Seb, you helped. Bill, you know, you provided a review and some endorsements.  But, you know, honestly, we didn’t know what we were talking about because no one had access to this thing. And so we just made a bunch of guesses. So really, the whole thing I wanted to probe with the two of you is, now with two years of experience out in the world, what, you know, what do we think is happening today?  You know, is AI actually having an impact, positive or negative, on healthcare and medicine? And what do we now think is going to happen in the next two years, five years, or 10 years? And so I realize it’s a little bit too abstract to just ask it that way. So let me just try to narrow the discussion and guide us a little bit.   Um, the kind of administrative and clerical work, paperwork, around healthcare—and we made a lot of guesses about that—that appears to be going well, but, you know, Bill, I know we’ve discussed that sometimes that you think there ought to be a lot more going on. Do you have a viewpoint on how AI is actually finding its way into reducing paperwork?  GATES: Well, I’m stunned … I don’t think there should be a patient-doctor meeting where the AI is not sitting in and both transcribing, offering to help with the paperwork, and even making suggestions, although the doctor will be the one, you know, who makes the final decision about the diagnosis and whatever prescription gets done.   It’s so helpful. You know, when that patient goes home and their, you know, son who wants to understand what happened has some questions, that AI should be available to continue that conversation. And the way you can improve that experience and streamline things and, you know, involve the people who advise you. I don’t understand why that’s not more adopted, because there you still have the human in the loop making that final decision.  But even for, like, follow-up calls to make sure the patient did things, to understand if they have concerns and knowing when to escalate back to the doctor, the benefit is incredible. And, you know, that thing is ready for prime time. That paradigm is ready for prime time, in my view.  LEE: Yeah, there are some good products, but it seems like the number one use right now—and we kind of got this from some of the previous guests in previous episodes—is the use of AI just to respond to emails from patients. [LAUGHTER] Does that make sense to you?  BUBECK: Yeah. So maybe I want to second what Bill was saying but maybe take a step back first. You know, two years ago, like, the concept of clinical scribes, which is one of the things that we’re talking about right now, it would have sounded, in fact, it sounded two years ago, borderline dangerous. Because everybody was worried about hallucinations. What happened if you have this AI listening in and then it transcribes, you know, something wrong?  Now, two years later, I think it’s mostly working. And in fact, it is not yet, you know, fully adopted. You’re right. But it is in production. It is used, you know, in many, many places. So this rate of progress is astounding because it wasn’t obvious that we would be able to overcome those obstacles of hallucination. It’s not to say that hallucinations are fully solved. In the case of the closed system, they are.   Now, I think more generally what’s going on in the background is that there is something that we, that certainly I, underestimated, which is this management overhead. So I think the reason why this is not adopted everywhere is really a training and teaching aspect. People need to be taught, like, those systems, how to interact with them.  And one example that I really like, a study that recently appeared where they tried to use ChatGPT for diagnosis and they were comparing doctors without and with ChatGPT (opens in new tab). And the amazing thing … so this was a set of cases where the accuracy of the doctors alone was around 75%. ChatGPT alone was 90%. So that’s already kind of mind blowing. But then the kicker is that doctors with ChatGPT was 80%.   Intelligence alone is not enough. It’s also how it’s presented, how you interact with it. And ChatGPT, it’s an amazing tool. Obviously, I absolutely love it. But it’s not … you don’t want a doctor to have to type in, you know, prompts and use it that way.  It should be, as Bill was saying, kind of running continuously in the background, sending you notifications. And you have to be really careful of the rate at which those notifications are being sent. Because if they are too frequent, then the doctor will learn to ignore them. So you have to … all of those things matter, in fact, at least as much as the level of intelligence of the machine.  LEE: One of the things I think about, Bill, in that scenario that you described, doctors do some thinking about the patient when they write the note. So, you know, I’m always a little uncertain whether it’s actually … you know, you wouldn’t necessarily want to fully automate this, I don’t think. Or at least there needs to be some prompt to the doctor to make sure that the doctor puts some thought into what happened in the encounter with the patient. Does that make sense to you at all?  GATES: At this stage, you know, I’d still put the onus on the doctor to write the conclusions and the summary and not delegate that.  The tradeoffs you make a little bit are somewhat dependent on the situation you’re in. If you’re in Africa, So, yes, the doctor’s still going to have to do a lot of work, but just the quality of letting the patient and the people around them interact and ask questions and have things explained, that alone is such a quality improvement. It’s mind blowing.   LEE: So since you mentioned, you know, Africa—and, of course, this touches on the mission and some of the priorities of the Gates Foundation and this idea of democratization of access to expert medical care—what’s the most interesting stuff going on right now? Are there people and organizations or technologies that are impressing you or that you’re tracking?  GATES: Yeah. So the Gates Foundation has given out a lot of grants to people in Africa doing education, agriculture but more healthcare examples than anything. And the way these things start off, they often start out either being patient-centric in a narrow situation, like, OK, I’m a pregnant woman; talk to me. Or, I have infectious disease symptoms; talk to me. Or they’re connected to a health worker where they’re helping that worker get their job done. And we have lots of pilots out, you know, in both of those cases.   The dream would be eventually to have the thing the patient consults be so broad that it’s like having a doctor available who understands the local things.   LEE: Right.   GATES: We’re not there yet. But over the next two or three years, you know, particularly given the worsening financial constraints against African health systems, where the withdrawal of money has been dramatic, you know, figuring out how to take this—what I sometimes call “free intelligence”—and build a quality health system around that, we will have to be more radical in low-income countries than any rich country is ever going to be.   LEE: Also, there’s maybe a different regulatory environment, so some of those things maybe are easier? Because right now, I think the world hasn’t figured out how to and whether to regulate, let’s say, an AI that might give a medical diagnosis or write a prescription for a medication.  BUBECK: Yeah. I think one issue with this, and it’s also slowing down the deployment of AI in healthcare more generally, is a lack of proper benchmark. Because, you know, you were mentioning the USMLE [United States Medical Licensing Examination], for example. That’s a great test to test human beings and their knowledge of healthcare and medicine. But it’s not a great test to give to an AI.  It’s not asking the right questions. So finding what are the right questions to test whether an AI system is ready to give diagnosis in a constrained setting, that’s a very, very important direction, which to my surprise, is not yet accelerating at the rate that I was hoping for.  LEE: OK, so that gives me an excuse to get more now into the core AI tech because something I’ve discussed with both of you is this issue of what are the right tests. And you both know the very first test I give to any new spin of an LLM is I present a patient, the results—a mythical patient—the results of my physical exam, my mythical physical exam. Maybe some results of some initial labs. And then I present or propose a differential diagnosis. And if you’re not in medicine, a differential diagnosis you can just think of as a prioritized list of the possible diagnoses that fit with all that data. And in that proposed differential, I always intentionally make two mistakes.  I make a textbook technical error in one of the possible elements of the differential diagnosis, and I have an error of omission. And, you know, I just want to know, does the LLM understand what I’m talking about? And all the good ones out there do now. But then I want to know, can it spot the errors? And then most importantly, is it willing to tell me I’m wrong, that I’ve made a mistake?   That last piece seems really hard for AI today. And so let me ask you first, Seb, because at the time of this taping, of course, there was a new spin of GPT-4o last week that became overly sycophantic. In other words, it was actually prone in that test of mine not only to not tell me I’m wrong, but it actually praised me for the creativity of my differential. [LAUGHTER] What’s up with that?  BUBECK: Yeah, I guess it’s a testament to the fact that training those models is still more of an art than a science. So it’s a difficult job. Just to be clear with the audience, we have rolled back that [LAUGHS] version of GPT-4o, so now we don’t have the sycophant version out there.  Yeah, no, it’s a really difficult question. It has to do … as you said, it’s very technical. It has to do with the post-training and how, like, where do you nudge the model? So, you know, there is this very classical by now technique called RLHF [reinforcement learning from human feedback], where you push the model in the direction of a certain reward model. So the reward model is just telling the model, you know, what behavior is good, what behavior is bad.  But this reward model is itself an LLM, and, you know, Bill was saying at the very beginning of the conversation that we don’t really understand how those LLMs deal with concepts like, you know, where is the capital of France located? Things like that. It is the same thing for this reward model. We don’t know why it says that it prefers one output to another, and whether this is correlated with some sycophancy is, you know, something that we discovered basically just now. That if you push too hard in optimization on this reward model, you will get a sycophant model.  So it’s kind of … what I’m trying to say is we became too good at what we were doing, and we ended up, in fact, in a trap of the reward model.  LEE: I mean, you do want … it’s a difficult balance because you do want models to follow your desires and …  BUBECK: It’s a very difficult, very difficult balance.  LEE: So this brings up then the following question for me, which is the extent to which we think we’ll need to have specially trained models for things. So let me start with you, Bill. Do you have a point of view on whether we will need to, you know, quote-unquote take AI models to med school? Have them specially trained? Like, if you were going to deploy something to give medical care in underserved parts of the world, do we need to do something special to create those models?  GATES: We certainly need to teach them the African languages and the unique dialects so that the multimedia interactions are very high quality. We certainly need to teach them the disease prevalence and unique disease patterns like, you know, neglected tropical diseases and malaria. So we need to gather a set of facts that somebody trying to go for a US customer base, you know, wouldn’t necessarily have that in there.  Those two things are actually very straightforward because the additional training time is small. I’d say for the next few years, we’ll also need to do reinforcement learning about the context of being a doctor and how important certain behaviors are. Humans learn over the course of their life to some degree that, I’m in a different context and the way I behave in terms of being willing to criticize or be nice, you know, how important is it? Who’s here? What’s my relationship to them?   Right now, these machines don’t have that broad social experience. And so if you know it’s going to be used for health things, a lot of reinforcement learning of the very best humans in that context would still be valuable. Eventually, the models will, having read all the literature of the world about good doctors, bad doctors, it’ll understand as soon as you say, “I want you to be a doctor diagnosing somebody.” All of the implicit reinforcement that fits that situation, you know, will be there. LEE: Yeah. GATES: And so I hope three years from now, we don’t have to do that reinforcement learning. But today, for any medical context, you would want a lot of data to reinforce tone, willingness to say things when, you know, there might be something significant at stake.  LEE: Yeah. So, you know, something Bill said, kind of, reminds me of another thing that I think we missed, which is, the context also … and the specialization also pertains to different, I guess, what we still call “modes,” although I don’t know if the idea of multimodal is the same as it was two years ago. But, you know, what do you make of all of the hubbub around—in fact, within Microsoft Research, this is a big deal, but I think we’re far from alone—you know, medical images and vision, video, proteins and molecules, cell, you know, cellular data and so on.  BUBECK: Yeah. OK. So there is a lot to say to everything … to the last, you know, couple of minutes. Maybe on the specialization aspect, you know, I think there is, hiding behind this, a really fundamental scientific question of whether eventually we have a singular AGI [artificial general intelligence] that kind of knows everything and you can just put, you know, explain your own context and it will just get it and understand everything.  That’s one vision. I have to say, I don’t particularly believe in this vision. In fact, we humans are not like that at all. I think, hopefully, we are general intelligences, yet we have to specialize a lot. And, you know, I did myself a lot of RL, reinforcement learning, on mathematics. Like, that’s what I did, you know, spent a lot of time doing that. And I didn’t improve on other aspects. You know, in fact, I probably degraded in other aspects. [LAUGHTER] So it’s … I think it’s an important example to have in mind.  LEE: I think I might disagree with you on that, though, because, like, doesn’t a model have to see both good science and bad science in order to be able to gain the ability to discern between the two?  BUBECK: Yeah, no, that absolutely. I think there is value in seeing the generality, in having a very broad base. But then you, kind of, specialize on verticals. And this is where also, you know, open-weights model, which we haven’t talked about yet, are really important because they allow you to provide this broad base to everyone. And then you can specialize on top of it.  LEE: So we have about three hours of stuff to talk about, but our time is actually running low. BUBECK: Yes, yes, yes.   LEE: So I think I want … there’s a more provocative question. It’s almost a silly question, but I need to ask it of the two of you, which is, is there a future, you know, where AI replaces doctors or replaces, you know, medical specialties that we have today? So what does the world look like, say, five years from now?  GATES: Well, it’s important to distinguish healthcare discovery activity from healthcare delivery activity. We focused mostly on delivery. I think it’s very much within the realm of possibility that the AI is not only accelerating healthcare discovery but substituting for a lot of the roles of, you know, I’m an organic chemist, or I run various types of assays. I can see those, which are, you know, testable-output-type jobs but with still very high value, I can see, you know, some replacement in those areas before the doctor.   The doctor, still understanding the human condition and long-term dialogues, you know, they’ve had a lifetime of reinforcement of that, particularly when you get into areas like mental health. So I wouldn’t say in five years, either people will choose to adopt it, but it will be profound that there’ll be this nearly free intelligence that can do follow-up, that can help you, you know, make sure you went through different possibilities.  And so I’d say, yes, we’ll have doctors, but I’d say healthcare will be massively transformed in its quality and in efficiency by AI in that time period.  LEE: Is there a comparison, useful comparison, say, between doctors and, say, programmers, computer programmers, or doctors and, I don’t know, lawyers?  GATES: Programming is another one that has, kind of, a mathematical correctness to it, you know, and so the objective function that you’re trying to reinforce to, as soon as you can understand the state machines, you can have something that’s “checkable”; that’s correct. So I think programming, you know, which is weird to say, that the machine will beat us at most programming tasks before we let it take over roles that have deep empathy, you know, physical presence and social understanding in them.  LEE: Yeah. By the way, you know, I fully expect in five years that AI will produce mathematical proofs that are checkable for validity, easily checkable, because they’ll be written in a proof-checking language like Lean or something but will be so complex that no human mathematician can understand them. I expect that to happen.   I can imagine in some fields, like cellular biology, we could have the same situation in the future because the molecular pathways, the chemistry, biochemistry of human cells or living cells is as complex as any mathematics, and so it seems possible that we may be in a state where in wet lab, we see, Oh yeah, this actually works, but no one can understand why.  BUBECK: Yeah, absolutely. I mean, I think I really agree with Bill’s distinction of the discovery and the delivery, and indeed, the discovery’s when you can check things, and at the end, there is an artifact that you can verify. You know, you can run the protocol in the wet lab and see [if you have] produced what you wanted. So I absolutely agree with that.   And in fact, you know, we don’t have to talk five years from now. I don’t know if you know, but just recently, there was a paper that was published on a scientific discovery using o3- mini (opens in new tab). So this is really amazing. And, you know, just very quickly, just so people know, it was about this statistical physics model, the frustrated Potts model, which has to do with coloring, and basically, the case of three colors, like, more than two colors was open for a long time, and o3 was able to reduce the case of three colors to two colors.   LEE: Yeah.  BUBECK: Which is just, like, astounding. And this is not … this is now. This is happening right now. So this is something that I personally didn’t expect it would happen so quickly, and it’s due to those reasoning models.   Now, on the delivery side, I would add something more to it for the reason why doctors and, in fact, lawyers and coders will remain for a long time, and it’s because we still don’t understand how those models generalize. Like, at the end of the day, we are not able to tell you when they are confronted with a really new, novel situation, whether they will work or not.  Nobody is able to give you that guarantee. And I think until we understand this generalization better, we’re not going to be willing to just let the system in the wild without human supervision.  LEE: But don’t human doctors, human specialists … so, for example, a cardiologist sees a patient in a certain way that a nephrologist …  BUBECK: Yeah. LEE: … or an endocrinologist might not. BUBECK: That’s right. But another cardiologist will understand and, kind of, expect a certain level of generalization from their peer. And this, we just don’t have it with AI models. Now, of course, you’re exactly right. That generalization is also hard for humans. Like, if you have a human trained for one task and you put them into another task, then you don’t … you often don’t know. LEE: OK. You know, the podcast is focused on what’s happened over the last two years. But now, I’d like one provocative prediction about what you think the world of AI and medicine is going to be at some point in the future. You pick your timeframe. I don’t care if it’s two years or 20 years from now, but, you know, what do you think will be different about AI in medicine in that future than today?  BUBECK: Yeah, I think the deployment is going to accelerate soon. Like, we’re really not missing very much. There is this enormous capability overhang. Like, even if progress completely stopped, with current systems, we can do a lot more than what we’re doing right now. So I think this will … this has to be realized, you know, sooner rather than later.  And I think it’s probably dependent on these benchmarks and proper evaluation and tying this with regulation. So these are things that take time in human society and for good reason. But now we already are at two years; you know, give it another two years and it should be really …   LEE: Will AI prescribe your medicines? Write your prescriptions?  BUBECK: I think yes. I think yes.  LEE: OK. Bill?  GATES: Well, I think the next two years, we’ll have massive pilots, and so the amount of use of the AI, still in a copilot-type mode, you know, we should get millions of patient visits, you know, both in general medicine and in the mental health side, as well. And I think that’s going to build up both the data and the confidence to give the AI some additional autonomy. You know, are you going to let it talk to you at night when you’re panicked about your mental health with some ability to escalate? And, you know, I’ve gone so far as to tell politicians with national health systems that if they deploy AI appropriately, that the quality of care, the overload of the doctors, the improvement in the economics will be enough that their voters will be stunned because they just don’t expect this, and, you know, they could be reelected [LAUGHTER] just on this one thing of fixing what is a very overloaded and economically challenged health system in these rich countries.  You know, my personal role is going to be to make sure that in the poorer countries, there isn’t some lag; in fact, in many cases, that we’ll be more aggressive because, you know, we’re comparing to having no access to doctors at all. And, you know, so I think whether it’s India or Africa, there’ll be lessons that are globally valuable because we need medical intelligence. And, you know, thank god AI is going to provide a lot of that.  LEE: Well, on that optimistic note, I think that’s a good way to end. Bill, Seb, really appreciate all of this.   I think the most fundamental prediction we made in the book is that AI would actually find its way into the practice of medicine, and I think that that at least has come true, maybe in different ways than we expected, but it’s come true, and I think it’ll only accelerate from here. So thanks again, both of you.  [TRANSITION MUSIC]  GATES: Yeah. Thanks, you guys.  BUBECK: Thank you, Peter. Thanks, Bill.  LEE: I just always feel such a sense of privilege to have a chance to interact and actually work with people like Bill and Sébastien.    With Bill, I’m always amazed at how practically minded he is. He’s really thinking about the nuts and bolts of what AI might be able to do for people, and his thoughts about underserved parts of the world, the idea that we might actually be able to empower people with access to expert medical knowledge, I think is both inspiring and amazing.   And then, Seb, Sébastien Bubeck, he’s just absolutely a brilliant mind. He has a really firm grip on the deep mathematics of artificial intelligence and brings that to bear in his research and development work. And where that mathematics takes him isn’t just into the nuts and bolts of algorithms but into philosophical questions about the nature of intelligence.   One of the things that Sébastien brought up was the state of evaluation of AI systems. And indeed, he was fairly critical in our conversation. But of course, the world of AI research and development is just moving so fast, and indeed, since we recorded our conversation, OpenAI, in fact, released a new evaluation metric that is directly relevant to medical applications, and that is something called HealthBench. And Microsoft Research also released a new evaluation approach or process called ADeLe.   HealthBench and ADeLe are examples of new approaches to evaluating AI models that are less about testing their knowledge and ability to pass multiple-choice exams and instead are evaluation approaches designed to assess how well AI models are able to complete tasks that actually arise every day in typical healthcare or biomedical research settings. These are examples of really important good work that speak to how well AI models work in the real world of healthcare and biomedical research and how well they can collaborate with human beings in those settings.  You know, I asked Bill and Seb to make some predictions about the future. You know, my own answer, I expect that we’re going to be able to use AI to change how we diagnose patients, change how we decide treatment options.   If you’re a doctor or a nurse and you encounter a patient, you’ll ask questions, do a physical exam, you know, call out for labs just like you do today, but then you’ll be able to engage with AI based on all of that data and just ask, you know, based on all the other people who have gone through the same experience, who have similar data, how were they diagnosed? How were they treated? What were their outcomes? And what does that mean for the patient I have right now? Some people call it the “patients like me” paradigm. And I think that’s going to become real because of AI within our lifetimes. That idea of really grounding the delivery in healthcare and medical practice through data and intelligence, I actually now don’t see any barriers to that future becoming real.  [THEME MUSIC]  I’d like to extend another big thank you to Bill and Sébastien for their time. And to our listeners, as always, it’s a pleasure to have you along for the ride. I hope you’ll join us for our remaining conversations, as well as a second coauthor roundtable with Carey and Zak.   Until next time.   [MUSIC FADES]

Solar air heating is among the most cost-effective applications of solar thermal energy. These systems are used for space heating and preheating fresh air for ventilation, typically using glazed or unglazed perforated solar collectors. The collectors draw in outside air, heat it using solar energy, and then distribute it through ductwork to meet building heating and fresh air needs. In 2024, Canada led again the world for the at least seventh year in a row in solar air heating adoption. The four key suppliers – Trigo Energies, Conserval Engineering, Matrix Energy, and Aéronergie – reported a combined 26,203 m2 (282,046 ft2) of collector area sold last year. Several of these providers are optimistic about the growing demand. These findings come from the newly released Canadian Solar Thermal Market Survey 2024, commissioned by Natural Resources Canada. Canada is the global leader in solar air heating. The market is driven by a strong network of experienced system suppliers, optimized technologies, and a few small favorable funding programs – especially in the province of Quebec. Architects and developers are increasingly turning to these cost-effective, façade-integrated systems as a practical solution for reducing onsite natural gas consumption. Despite its cold climate, Canada benefits from strong solar potential with solar irradiance in many areas rivaling or even exceeding that of parts of Europe. This makes solar air heating not only viable, but especially valuable in buildings with high fresh air requirements including schools, hospitals, and offices. The projects highlighted in this article showcase the versatility and relevance of solar air heating across a range of building types, from new constructions to retrofits. Figure 1: Preheating air for industrial buildings: 2,750 m2 (29,600 ft2) of Calento SL solar air collectors cover all south-west and south-east facing facades of the FAB3R factory in Trois-Rivières, Quebec. The hourly unitary flow rate is set at 41 m3/m2 or 2.23 cfm/ft2 of collector area, at the lower range because only a limited number of intake fans was close enough to the solar façade to avoid long ventilation ductwork. Photo: Trigo Energies Quebec’s solar air heating boom: the Trigo Energies story Trigo Energies makes almost 90 per cent of its sales in Quebec. “We profit from great subsidies, as solar air systems are supported by several organizations in our province – the electricity utility Hydro Quebec, the gas utility Energir and the Ministry of Natural Resources,” explained Christian Vachon, Vice President Technologies and R&D at Trigo Energies. Trigo Energies currently has nine employees directly involved in planning, engineering and installing solar air heating systems and teams up with several partner contractors to install mostly retrofit projects. “A high degree of engineering is required to fit a solar heating system into an existing factory,” emphasized Vachon. “Knowledge about HVAC engineering is as important as experience with solar thermal and architecture.” One recent Trigo installation is at the FAB3R factory in Trois-Rivières. FAB3R specializes in manufacturing, repairing, and refurbishing large industrial equipment. Its air heating and ventilation system needed urgent renovation because of leakages and discomfort for the workers. “Due to many positive references he had from industries in the area, the owner of FAB3R contacted us,” explained Vachon. “The existence of subsidies helped the client to go for a retrofitting project including solar façade at once instead of fixing the problems one bit at a time.” Approximately 50 per cent of the investment costs for both the solar air heating and the renovation of the indoor ventilation system were covered by grants and subsidies. FAB3R profited from an Energir grant targeted at solar preheating, plus an investment subsidy from the Government of Quebec’s EcoPerformance Programme.   Blue or black, but always efficient: the advanced absorber coating In October 2024, the majority of the new 2,750 m² (29,600 ft2) solar façade at FAB3R began operation (see figure 1). According to Vachon, the system is expected to cover approximately 13 per cent of the factory’s annual heating demand, which is otherwise met by natural gas. Trigo Energies equipped the façade with its high-performance Calento SL collectors, featuring a notable innovation: a selective, low-emissivity coating that withstands outdoor conditions. Introduced by Trigo in 2019 and manufactured by Almeco Group from Italy, this advanced coating is engineered to maximize solar absorption while minimizing heat loss via infrared emission, enhancing the overall efficiency of the system. The high efficiency coating is now standard in Trigo’s air heating systems. According to the manufacturer, the improved collector design shows a 25 to 35 per cent increase in yield over the former generation of solar air collectors with black paint. Testing conducted at Queen’s University confirms this performance advantage. Researchers measured the performance of transpired solar air collectors both with and without a selective coating, mounted side-by-side on a south-facing vertical wall. The results showed that the collectors with the selective coating produced 1.3 to 1.5 times more energy than those without it. In 2024, the monitoring results were jointly published by Queen’s University and Canmat Energy in a paper titled Performance Comparison of a Transpired Air Solar Collector with Low-E Surface Coating. Selective coating, also used on other solar thermal technologies including glazed flat plate or vacuum tube collectors, has a distinctive blue color. Trigo customers can, however, choose between blue and black finishes. “By going from the normal blue selective coating to black selective coating, which Almeco is specially producing for Trigo, we lose about 1 per cent in solar efficiency,” explained Vachon. Figure 2: Building-integrated solar air heating façade with MatrixAir collectors at the firehall building in Mont Saint Hilaire, south of Montreal. The 190 m2 (2,045 ft2) south-facing wall preheats the fresh air, reducing natural gas consumption by 18 per cent compared to the conventional make-up system. Architect: Leclerc Architecture. Photo: Matrix Energy Matrix Energy: collaborating with architects and engineers in new builds The key target customer group of Matrix Energy are public buildings – mainly new construction. “Since the pandemic, schools are more conscious about fresh air, and solar preheating of the incoming fresh air has a positive impact over the entire school year,” noted Brian Wilkinson, President of Matrix Energy. Matrix Energy supplies systems across Canada, working with local partners to source and process the metal sheets used in their MatrixAir collectors. These metal sheets are perforated and then formed into architectural cladding profiles. The company exclusively offers unglazed, single-stage collectors, citing fire safety concerns associated with polymeric covers. “We have strong relationships with many architects and engineers who appreciate the simplicity and cost-effectiveness of transpired solar air heating systems,” said President Brian Wilkinson, describing the company’s sales approach. “Matrix handles system design and supplies the necessary materials, while installation is carried out by specialized cladding and HVAC contractors overseen by on-site architects and engineers,” Wilkinson added. Finding the right flow: the importance of unitary airflow rates One of the key design factors in solar air heating systems is the amount of air that passes through each square meter of the perforated metal absorber,  known as the unitary airflow rate. The principle is straightforward: higher airflow rates deliver more total heat to the building, while lower flow rates result in higher outlet air temperatures. Striking the right balance between air volume and temperature gain is essential for efficient system performance. For unglazed collectors mounted on building façades, typical hourly flow rates should range between 120 and 170 (m3/h/m2), or 6.6 to 9.4 cfm/ft2. However, Wilkinson suggests that an hourly airflow rate of around 130 m³/h/m² (7.2 cfm/ft2) offers the best cost-benefit balance for building owners. If the airflow is lower, the system will deliver higher air temperatures, but it would then need a much larger collector area to achieve the same air volume and optimum performance, he explained. It’s also crucial for the flow rate to overcome external wind pressure. As wind passes over the absorber, air flow through the collector’s perforations is reduced, resulting in heat losses to the environment. This effect becomes even more pronounced in taller buildings, where wind exposure is greater. To ensure the system performs well even in these conditions, higher hourly airflow rates typically between 150 and 170 m³/m² (8.3 to 9.4 cfm/ft2)  are necessary. Figure 3: One of three apartment blocks of the Maple House in Toronto’s Canary District. Around 160 m2 (1,722 ft2) of SolarWall collectors clad the two-storey mechanical penthouse on the roof. The rental flats have been occupied since the beginning of 2024. Collaborators: architects-Alliance, Claude Cormier et Associés, Thornton Tomasetti, RWDI, Cole Engineering, DesignAgency, MVShore, BA Group, EllisDon. Photo: Conserval Engineering Solar air heating systems support LEED-certified building designs Solar air collectors are also well-suited for use in multi-unit residential buildings. A prime example is the Canary District in Toronto (see Figure 3), where single-stage SolarWall collectors from Conserval Engineering have been installed on several MURBs to clad the mechanical penthouses. “These penthouses are an ideal location for our air heating collectors, as they contain the make-up air units that supply corridor ventilation throughout the building,” explained Victoria Hollick, Vice President of Conserval Engineering. “The walls are typically finished with metal façades, which can be seamlessly replaced with a SolarWall system – maintaining the architectural language without disruption.” To date, nine solar air heating systems have been commissioned in the Canary District, covering a total collector area of over 1,000 m² (10,764 ft2). “Our customers have many motivations to integrate SolarWall technology into their new construction or retrofit projects, either carbon reduction, ESG, or green building certification targets,” explained Hollick. The use of solar air collectors in the Canary District was proposed by architects from the Danish firm Cobe. The black-colored SolarWall system preheats incoming air before it is distributed to the building’s corridors and common areas, reducing reliance on natural gas heating and supporting the pursuit of LEED Gold certification. Hollick estimates the amount of gas saved between 10 to 20 per cent of the total heating load for the corridor ventilation of the multi-unit residential buildings. Additional energy-saving strategies include a 50/50 window-to-wall ratio with high-performance glazing, green roofs, high-efficiency mechanical systems, LED lighting, and Energy Star-certified appliances. The ideal orientation for a SolarWall system is due south. However, the systems can be built at any orientation up to 90° east and west, explained Hollick. A SolarWall at 90° would have approximately 60 per cent of the energy production of the same area facing south.Canada’s expertise in solar air heating continues to set a global benchmark, driven by supporting R&D, by innovative technologies, strategic partnerships, and a growing portfolio of high-impact projects. With strong policy support and proven performance, solar air heating is poised to play a key role in the country’s energy-efficient building future. Figure 4: Claude-Bechard Building in Quebec is a showcase project for sustainable architecture with a 72 m2 (775 ft2) Lubi solar air heating wall from Aéronergie. It serves as a regional administrative center. Architectural firm: Goulet et Lebel Architectes. Photo: Art Massif Bärbel Epp is the general manager of the German Agency solrico, whose focus is on solar market research and international communication. The post Op-ed: Canada’s leadership in solar air heating—Innovation and flagship projects appeared first on Canadian Architect.

The desert that surrounds Taliesin West, Frank Lloyd Wright’s winter home and studio, is no stranger to camping. Which is perhaps why it is the perfect place to unveil the Airstream Frank Lloyd Wright Usonian Limited Edition Travel Trailer, a new collaboration between the architect’s eponymous foundation and the American travel trailer brand.When Wright arrived in the Sonoran Desert in December of 1937, he made two purchases. First, 600 acres of land, on which Taliesin West would eventually sit. Then, shortly after, a handful of tents for his apprentices to sleep in while they helped build the new property. Even once construction finished, it became a tradition that his disciples would build temporary shelters among the cacti, bushes, and sandy soil. “This was a camp, and Wright was moved by the way canvas from the tents diffused light. That’s what inspired the canvas roofs on Taliesin West today,” Sally Russel, the director of licensing at the Frank Lloyd Wright Foundation, said at a press briefing at Taliesin West about the trailer.You might also like: What Was It Like Living at Frank Lloyd Wright’s Taliesin West?The trailer door features a pattern called the Gordon Leaf motif, which was created by Taliesin apprentice Eugene Masselink. Photo: Andrew PielageCoincidentally, Airstream’s founder, Wally Byam, began designing trailers for people who didn’t like sleeping on the ground in tents—a sect his first wife belonged to. Nearly 100 years later, the Usonian trailer lets owners enjoy the desert (or any part of the world) Wright-style, while still taking advantage of modern comforts like a bed, shower, and kitchen. “I’ve been dropping the idea of a Frank Lloyd Wright trailer into the thought mill at Airstream for about 20 years,” Bob Wheeler, the president and CEO of Airstream, said at the briefing.The kitchen includes under cabinet lighting and warm, wood-toned cabinets. Photo: Andrew PielageInside the Frank Lloyd Wright AirstreamAt just over 28 feet long, the trailer is among the larger of Airstream’s offerings, which range from 16 to 33 feet. From the outside, the company’s instantly recognizable aluminum shell offers little evidence of the idiosyncrasy that’s on full display inside. But from the moment the door opens—which is printed with a leaf motif designed by a Taliesin apprentice—Wright’s influence is all encompassing.You might also like: 7 Stylish Mobile Homes Owned by Celebrities

From garage apartments and carriage houses to backyard guest suites and stand-alone home offices, ADUs are one of the most ingenious ways to maximize your property's available space—and your home's property value. Also known as an accessory dwelling unit, these outbuildings aren't a new concept, but they are becoming more prevalent, thanks to an ultra-competitive housing market. As multigenerational living has become increasingly popular among Americans and home prices continue to rise, more homeowners are looking to expand their homes' footprint with additional living spaces, whether for personal or investment purposes. Accessory dwelling units are essentially your everyday tiny homes. However, unlike tiny home floor plans, ADUs can be attached to a single-family house, depending on the type.As ADUs have become more popular, it only makes sense that interior designers are taking on more projects involving the structures. Whether you're interested in putting in an ADU or want to renovate the current one you have, keep reading to hear from two experts on what's trending in interiors for these unique living units.Related StoriesWhat is an ADU?Joey Puterbaugh and Kristian Alveo for KebonyDesigner Kim Lewis looked to the outdoors when designing this ADU, where the living room opens up to a large patio. An accessory dwelling unit, or ADU, is an independent, secondary housing unit that resides on a single-family lot. This is not your backyard shed or detached garage—but it could be. There are a few types of ADUs, but they all mainly fall into these three categories: detached, internal, and attached, according to the American Planning Association (APA). A backyard shed or detached garage that's been converted into an ADU falls into the detached category. However, a detached ADU can be a new construction, too. It just needs to be able to stand alone.An internal ADU is a converted portion of an existing home. Think of a basement apartment or a different floor of a home that's been converted into an additional independent living space.Additions to new or existing homes fall under the category of attached ADUs. These share at least one wall with the single-family home, and are also sometimes called "bump-out" ADUs. What Are ADUs Used For?When someone talks about an in-law suite or an attached Airbnb, they're likely talking about an ADU. “Home office and mother-in-law suites are still the main requests and purposes,” Kim Lewis of Kim Lewis Designs says. She's worked on a handful of ADU projects in the past, so the designer is familiar with these additions. A lot of times, homeowners who add on an ADU are doing so to expand their square footage, both for themselves or family members. Lewis explains that multigenerational living is on the rise in the United States, which is why ADUs are becoming more popular. Kori Sassower, principal agent at The Kori Sassower Team, agrees, adding, “Whether it is an aging parent or a 20-something-year-old child looking to save some money, the flexibility of an ADU has become paramount. We have a housing shortage, taxes are increasing, and ADUs are a very good, safe choice as a home improvement.”However, some homeowners are using ADUs as investment opportunities. “Depending on the area, some clients see this as an investment opportunity; utilizing the asset as a short-term rental where property restrictions allow for it,” Lewis explains. As the market for ADUs increases, Lewis predicts that there will be more legislation introduced that allows these home additions in certain jurisdictions. We have a housing shortage, taxes are increasing, ADUs are a very good, safe choice as a home improvement. ADU Trends Experts Are SeeingWhen it comes to the interiors of ADUs, they tend to follow the trends we're seeing in single-family homes. Joy-filled design, unique features, bold paint colors—they're not basic white boxes. Below, find the six biggest trends experts are seeing in ADUs. Architectural Leaps“ADUs are taking more interesting shapes in architecture,” Lewis says. As 3D printed homes become more popular, Lewis predicts that we'll see more curves in architecture with ADUs, like barrel roofs, radius rooflines, and rounded corners on decking. In a market heavily saturated with boxy, square, tiny homes, she's expecting more progressive architecture with the use of large glass walls to bring the outdoors in. Bold AppliancesChase Daniel"I love how we embraced a darker, moody palette, even though the rooms were small," Lewis says of this ADU.Colorful appliances are trending in every type of home right now, and they're not exclusive to large kitchens. Lewis says that small spaces can pack a big punch when it comes to color, and she's seeing more of that in the ADUs (and traditional interiors) she's designing. “I love a dark burgundy cabinet right now,” she adds.Vast Outdoor SpacesJoey Puterbaugh and Kristian Alveo for KebonyThe outdoor space in this ADU by Lewis would make anyone jealous. Large, open outdoor spaces are very common with ADUs, since the living quarters are smaller than those of an average single-family home. “Large decking can be used to create a more useful footprint around the perimeter of the ADU, tying the accessory space to the main home,” Lewis says. “A desire for nature, sustainable living, and more farming at home might point us to more gardening trends with ADUs.” Wellness RetreatsSassower is seeing more ADUs pop up with a focus on wellness rather than actual living. Personal spas, specifically, have been very popular as the weather gets warmer. Sports CenteredSimilar to basement or garage home gyms, ADUs can be used for more athletic purposes, too, says Sassower. She's seen everything from golf simulators and mini basketball courts to interiors that nod to a ski lodge.Personality-Based DesignThe living room of an ADU renovation that Lewis completed in Fredericksburg, TX, features a dark, moody palette. Deep olive green walls and a matching couch make the space feel sophisticated yet warm. Chase DanielThe powder room in the same Texas ADU by Lewis is a true jewel box space, with a standout fluted stone sink.Chase DanielLike regular interior trends, ADU trends have been leaning towards design with more character. “I'm looking forward to using more limewash and Venetian plasters, and I'm always here for a funky tile to add character,” Lewis says.Follow House Beautiful on Instagram and TikTok.

With reclaimed land previously allocated to cars, the Grijalva River boardwalk offers generous public spaces and reconnects the Mexican city of Villahermosa to its river In Villahermosa, nature reigns supreme. Surrounded by rivers, lagoons, wild vegetation and the scorching heat of a humid tropical climate, the city’s identity is shaped by intense and unpredictable natural forces. The capital of the Mexican state of Tabasco was founded in 1564 on the banks of the Grijalva River, a vital trade route that has significantly shaped the city’s development. For locals, the river has long been both blessing and threat; major floods have been recorded since the 17th century. A devastating flood in 2007 submerged what officials estimated to be 80 per cent of the city, damaging or destroying more than 120,000 homes. In the aftermath of the inundation, high concrete retaining walls were built along both banks of the Grijalva River to prevent further flooding. While this was an understandable measure at first glance, it consequently caused residents to lose both their visual and physical connection with the river. As a result, people moved, particularly from the western bank where the historical centre is located, to new areas further away from the Grijalva River. The riverfront was left to deteriorate into a troubled zone. On the eastern bank, the neighbourhood of Gaviotas was already considered unsafe before the flood, yet it maintained more of its residential character. In 2022, 15 years after the dramatic flood, then‑president Andrés Manuel López Obrador, more commonly known as AMLO, announced the construction of a new 6km‑long riverfront promenade in Villahermosa, the capital of his home state. The idea was to enable the population to once again take pride in and live with their river, looking to Paris and Rome as examples. The monumental task, with its large urban scale and the population’s psychological trauma, was entrusted to the Ministry of Agricultural, Territorial and Urban Development (SEDATU) as part of their Programa de Mejoramiento Urbano (Urban Improvement Programme, or PMU). This programme aimed to use architecture as an ‘instrument of social transformation’. High expectations were placed on these projects; architects were asked to create ‘places of national pride’ while improving everyday living conditions. The architectural trio of Alejandro Castro Jiménez Labora, Mauricio Rocha Iturbide, and Óscar Rodríguez Castañeda, along with their teams, were commissioned to design a linear park along both banks of the Grijalva. Each architect contributed their strength: Castro brought his expertise in poetic urban furniture; Rocha his sensitive and atmospheric architectural approach; and Rodríguez his thoughtful urban and traffic planning skills. The SEDATU team provided technical and participatory expertise, enabling contextual sensitivity by sharing essential information about the site’s topography, soil conditions and water flows. From the city’s existing observatory, the Torre del Caballero landmark, visitors enjoy an excellent view over the redesigned riverbanks. The historical centre and the Gaviotas neighbourhood now form a single ensemble, while the intervention carefully responds to the different conditions found along the length of the river. The project’s main objective is to reclaim some of the land previously allocated to cars and create a promenade for pedestrians and slower vehicles, punctuated with public spaces and facilities. On both sides of the river, cars are now limited to just one or two grey asphalt lanes. Running alongside are generous cycle paths and pedestrian walkways made of earth‑coloured concrete. Speed bumps in the same material and colour connect the pavements on either side of the road while helping to limit traffic speed to 30km/h, further enhancing pedestrian safety. Several design elements are found along almost the entire promenade. A ribbon of light‑grey benches delineates the edge of the elevated riverfront; stone walls, steps and ramps are used to negotiate the slight changes in level; planters and lush vegetation soften the transition to the walkways, creating a welcome buffer from street traffic. The most visually striking components are the tall, red‑pigmented concrete light poles on the elevated path, adorned with elegant L‑shaped steel light fixtures, which establish a strong and cohesive visual rhythm. Only upon closer inspection you notice the 2007 retaining walls peeking through the dense tropical vegetation. Removing these unattractive concrete barriers was never an option; they stand as a symbol of successful flood protection for the local population. The architectural team ingeniously built the elevated promenade atop the existing wall – an effective concealment from the street side while simultaneously inviting residents to reconnect with the Grijalva.  At the foot of the observatory, directly below the retaining wall, the earth‑toned concrete platforms of the Carlos A Madrazo Becerra Park stretch towards the river. Visitors can access the park via a ramp from the promenade on the western bank or by ferry from the opposite side. In the park, concrete furnishings invite visitors to linger among tropical vegetation set against tall natural stone walls. Importantly, it is a space that is durable and requires minimal maintenance – a survival formula for public parks in the Mexican context. Small traces on the concrete benches reveal that the park weathered its baptism of fire last year: the design accommodates the river’s natural dynamics, adapting to fluctuating water levels without compromising public safety. Beyond providing much‑needed shade, the extensive planting of native, low‑maintenance plants on both riverbanks has improved soil stability. Above the park, on a broad extension of the elevated pathway, stand three long, elegant buildings with large cantilevered roofs supported by hefty beams resting on distinctive double columns. The tall glass walls that enclose the interiors are set back, creating a visual flow between interior and exterior spaces. While the beams evoke timber construction, they – like the columns – are made of the same pigmented concrete used for the promenade paving. Despite their refined composition, these structures have remained largely unused since their completion over a year ago, neither serving their intended function as restaurants nor hosting alternative uses. Even the beautifully designed park sees only limited public engagement. The ambitious goal of SEDATU with the PMU projects to ‘counteract violence and strengthen the social fabric’ appears, for now, to have fallen short in this area. According to national statistics, Villahermosa ranks first in perceived insecurity among Mexican cities. This sense of insecurity is tangible on the promenade by the city centre, where buildings that look abandoned contribute to an atmosphere of neglect. The situation is markedly different on the opposite riverbank, in the Gaviotas neighbourhood. Construction of the 3.5km promenade on this side began in 2021 with three open pavilions housing several small kiosks, which quickly evolved into popular taco stands. The Plaza Solidaridad, revitalised by the architectural trio, draws people from the surrounding vibrant neighbourhood. Further south, the final section that was built is a large sports area and children’s playground, which were embraced by the local community even before their official inauguration in February 2024. Especially after sunset, when the air cools, the well‑lit Gaviotas riverfront comes to life. During daylight hours, however, air‑conditioned shopping centres remain the preferred gathering places for the residents of Villahermosa. Rocha describes the city’s new promenade as a ‘jazz composition’, a striking metaphor that speaks of rhythmic complexity and the freedom to improvise. With just a few designed elements and carefully selected colours, the architects have harmoniously layered the river’s urban spaces. The project is earning international recognition but, in Mexico, it faced sharp criticism and was overshadowed by accusations of nepotism. Castro is a friend of AMLO’s son, and the fact that the intervention took place in the home state of the then‑president, coupled with its substantial budget by local standards, drew considerable attention. According to residents, this undermined public acceptance. When asked about the negative press, Rocha speaks of the need to develop a ‘crisis muscle’; he says architects working on public projects in Mexico must ‘let go of perfectionism’ as much lies beyond their control.  During AMLO’s six‑year term, which ended in 2024, SEDATU implemented 1,300 PMU projects in 193 highly marginalised municipalities across the country. While many of these interventions undoubtedly improved people’s quality of life, the Villahermosa riverside project also reveals architecture’s limitations, exposing some of the programme’s weaknesses: architectural interventions often act as sticking plasters on an extensively damaged urban fabric. They are handed over from a national ministry with comprehensive expertise and funding to local governments lacking the means to sustain them. Although SEDATU conducted participatory consultations during the project’s implementation, this engagement was absent once the project was completed. Public acceptance and appropriation can take time; what this project does is send an invitation out. 2025-06-05 Reuben J Brown Share AR June 2025RoadsBuy Now

Giant Squid’s upcoming title, Sword of the Sea, is a well-crafted union of all the titles the studio and Game Director Matt Nava have worked on. The aquatic wonder of Abzû, The Pathless’s sense of adventure, and Journey’s beautiful desert world. Sword of the Sea weaves those influences together so skillfully that it feels like this atmospheric surfing adventure was the goal all along.  I got a chance to go hands-on with Sword of the Sea and talk with Giant Squid about the new game, coming day one to PlayStation Plus Game Catalog on August 19. Surfing across the sands and sea The game begins as the player awakens as The Wraith, a vessel waiting to be filled, who is given the task of bringing the oceans back to the world. Appearing to be the last being left alive to get the job done, you receive a mystical surfboard and hit the dunes.  The starting area is a vast playground where you can get accustomed to the board, perform tricks, and learn how to interact with the world. As you zoom across the desert, you come across different nodes known as Ocean Seeds that allow you to cleanse the land and restore parts of the ocean and marine life.  Beyond pleasing aesthetics, water and sea life directly impact gameplay. The Wraith is significantly faster on water, leading to greater speed for bigger jumps and more tricks. Some sea life create new paths, like buoyant jellyfish that make great jump pads and long strands of climbable seaweed. Watching the landscape terraform was a highlight, as was searching for all the land’s secrets.  “Sword of the Sea is really all about the spiritual magic of surfing,” Nava explains. “ It’s inspired by my own experiences snowboarding, skateboarding, and surfing throughout my life. When you think of these extreme sports, the first thing that comes to mind is probably the high speed and the danger. But there’s more to it than just an adrenaline rush. When you are surfing on waves in the ocean, you are literally on the boundary between our world and the world beneath the waves, a world beyond our understanding.”  Golden triangles known as Tetra populate the world, acting as currency you can trade to mysterious vendors to learn more tricks, acquire skills, and open new paths. Want to see it all? You’ll have to detour off the main path to see and afford it all. Different vendors, triangles, and Trick Attack arenas are hidden throughout the world.  Show off your tricks and feel the terrain Pulling off stylish moves in Sword of the Sea is straightforward. When surfing, once you jump you can press X again to perform a variety of double jump tricks. Once you acquire advanced tricks from the vendor, you hold L1 and press Triangle, Square, Circle, or X to do special grabs, flips, and spins. They score more points, and combining grabs and flips in different ways create new tricks. While the timing is a little trickier, you can give yourself a much-needed jump boost and score high-point combos when performed just right.  Sword of the Sea constantly keeps score of the tricks you pull and how advanced they are. Hidden Trick Attack arenas let you put those skills to good use. What does racking up a high Trick Score lead to? Yet another mystery we’ll have to wait to uncover. Though the game takes place in the desert, there are plenty of different objects to grind on besides sea and sand. I came across ruins, ceramic tile roofs, and helpful aquatic life. The team at Giant Squid takes full advantage of the DualSense controller haptics to ensure every surface feels and sounds distinct. As I surfed along rooftops, I could hear the melodic tones of tiles underneath and the tingle in my fingers. The sand offers a coarse sensation, and you feel and hear the rushing water as you speed through on the makeshift highways parts of the ocean create. “The haptic feedback of the DualSense controller reacts to each surface differently so you can feel it when you cross from one to the next,” says Nava.  “Combined with special sound effects that play from the controller, it gives a very realistic sense of touching the surfaces you see in the game. We are very excited by the extra layer of detail the haptics adds to the experience of surfing in Sword of the Sea.” You’re free to explore in any given direction, and hard-to-reach places usually reward you with something appropriate for the time and skill it takes to reach them. However, as you explore, you will discover murals and other environmental elements that keep the story ever-present. A tale of loss and destruction, and trying to find out how you fit into it all is an intriguing thread to follow. Especially since at the end of my session, it was clear that a mysterious character was following me. It appeared to be another wraith, but whoever the pursuer was didn’t give vibes of good intentions.  It’s easy to compare Sword of the Sea to its predecessors, but the game truly feels like an evolution of everything the Giant Squid team has worked on to this point. The game’s sights, sounds, and feel connected me with the land, and I can’t wait to hop on the board again and see what’s really out there.  “Our games have a really recognizable and unique art style, and tell atmospheric stories with their bold color and music,” says Nava. “We start with a feeling, an experience, and a message that we want the player to feel. Really, all our games are all about movement, and how it can connect you with nature and the world in different ways. Sword of the Sea builds on the ideas in all our previous games to create something that is both very Giant Squid and an exciting new adventure.” Looking to reel in more Sword of the Sea details? Check out the latest trailer and more details from State of Play. More from June’s State of Play State of Play June 2025: all announcements, trailers