Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More January 2025 shook the AI landscape. The seemingly unstoppable OpenAI and the powerful American tech giants were shocked by what we can certainly call an underdog in the area of large language models (LLMs). DeepSeek, a Chinese firm not on anyone’s radar, suddenly challenged OpenAI. It is not that DeepSeek-R1 was better than the top models from American giants; it was slightly behind in terms of the benchmarks, but it suddenly made everyone think about the efficiency in terms of hardware and energy usage. Given the unavailability of the best high-end hardware, it seems that DeepSeek was motivated to innovate in the area of efficiency, which was a lesser concern for larger players. OpenAI has claimed they have evidence suggesting DeepSeek may have used their model for training, but we have no concrete proof to support this. So, whether it is true or it’s OpenAI simply trying to appease their investors is a topic of debate. However, DeepSeek has published their work, and people have verified that the results are reproducible at least on a much smaller scale. But how could DeepSeek attain such cost-savings while American companies could not? The short answer is simple: They had more motivation. The long answer requires a little bit more of a technical explanation. DeepSeek used KV-cache optimization One important cost-saving for GPU memory was optimization of the Key-Value cache used in every attention layer in an LLM. LLMs are made up of transformer blocks, each of which comprises an attention layer followed by a regular vanilla feed-forward network. The feed-forward network conceptually models arbitrary relationships, but in practice, it is difficult for it to always determine patterns in the data. The attention layer solves this problem for language modeling. The model processes texts using tokens, but for simplicity, we will refer to them as words. In an LLM, each word gets assigned a vector in a high dimension (say, a thousand dimensions). Conceptually, each dimension represents a concept, like being hot or cold, being green, being soft, being a noun. A word’s vector representation is its meaning and values according to each dimension. However, our language allows other words to modify the meaning of each word. For example, an apple has a meaning. But we can have a green apple as a modified version. A more extreme example of modification would be that an apple in an iPhone context differs from an apple in a meadow context. How do we let our system modify the vector meaning of a word based on another word? This is where attention comes in. The attention model assigns two other vectors to each word: a key and a query. The query represents the qualities of a word’s meaning that can be modified, and the key represents the type of modifications it can provide to other words. For example, the word ‘green’ can provide information about color and green-ness. So, the key of the word ‘green’ will have a high value on the ‘green-ness’ dimension. On the other hand, the word ‘apple’ can be green or not, so the query vector of ‘apple’ would also have a high value for the green-ness dimension. If we take the dot product of the key of ‘green’ with the query of ‘apple,’ the product should be relatively large compared to the product of the key of ‘table’ and the query of ‘apple.’ The attention layer then adds a small fraction of the value of the word ‘green’ to the value of the word ‘apple’. This way, the value of the word ‘apple’ is modified to be a little greener. When the LLM generates text, it does so one word after another. When it generates a word, all the previously generated words become part of its context. However, the keys and values of those words are already computed. When another word is added to the context, its value needs to be updated based on its query and the keys and values of all the previous words. That’s why all those values are stored in the GPU memory. This is the KV cache. DeepSeek determined that the key and the value of a word are related. So, the meaning of the word green and its ability to affect greenness are obviously very closely related. So, it is possible to compress both as a single (and maybe smaller) vector and decompress while processing very easily. DeepSeek has found that it does affect their performance on benchmarks, but it saves a lot of GPU memory. DeepSeek applied MoE The nature of a neural network is that the entire network needs to be evaluated (or computed) for every query. However, not all of this is useful computation. Knowledge of the world sits in the weights or parameters of a network. Knowledge about the Eiffel Tower is not used to answer questions about the history of South American tribes. Knowing that an apple is a fruit is not useful while answering questions about the general theory of relativity. However, when the network is computed, all parts of the network are processed regardless. This incurs huge computation costs during text generation that should ideally be avoided. This is where the idea of the mixture-of-experts (MoE) comes in. In an MoE model, the neural network is divided into multiple smaller networks called experts. Note that the ‘expert’ in the subject matter is not explicitly defined; the network figures it out during training. However, the networks assign some relevance score to each query and only activate the parts with higher matching scores. This provides huge cost savings in computation. Note that some questions need expertise in multiple areas to be answered properly, and the performance of such queries will be degraded. However, because the areas are figured out from the data, the number of such questions is minimised. The importance of reinforcement learning An LLM is taught to think through a chain-of-thought model, with the model fine-tuned to imitate thinking before delivering the answer. The model is asked to verbalize its thought (generate the thought before generating the answer). The model is then evaluated both on the thought and the answer, and trained with reinforcement learning (rewarded for a correct match and penalized for an incorrect match with the training data). This requires expensive training data with the thought token. DeepSeek only asked the system to generate the thoughts between the tags <think> and </think> and to generate the answers between the tags <answer> and </answer>. The model is rewarded or penalized purely based on the form (the use of the tags) and the match of the answers. This required much less expensive training data. During the early phase of RL, the model tried generated very little thought, which resulted in incorrect answers. Eventually, the model learned to generate both long and coherent thoughts, which is what DeepSeek calls the ‘a-ha’ moment. After this point, the quality of the answers improved quite a lot. DeepSeek employs several additional optimization tricks. However, they are highly technical, so I will not delve into them here. In any technology research, we first need to see what is possible before improving efficiency. This is a natural progression. DeepSeek’s contribution to the LLM landscape is phenomenal. The academic contribution cannot be ignored, whether or not they are trained using OpenAI output. It can also transform the way startups operate. But there is no reason for OpenAI or the other American giants to despair. This is how research works — one group benefits from the research of the other groups. DeepSeek certainly benefited from the earlier research performed by Google, OpenAI and numerous other researchers. However, the idea that OpenAI will dominate the LLM world indefinitely is now very unlikely. No amount of regulatory lobbying or finger-pointing will preserve their monopoly. The technology is already in the hands of many and out in the open, making its progress unstoppable. Although this may be a little bit of a headache for the investors of OpenAI, it’s ultimately a win for the rest of us. While the future belongs to many, we will always be thankful to early contributors like Google and OpenAI. Debasish Ray Chawdhuri is senior principal engineer at Talentica Software. Daily insights on business use cases with VB Daily If you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory shifts to practical deployments, so you can share insights for maximum ROI. Read our Privacy Policy Thanks for subscribing. Check out more VB newsletters here. An error occured.

Do you want to rewatch old CCTV footage or cutscenes in Blue Prince? There's actually a method that unlocks this particular feature using terminals. Of course, you also need to gather a few clues along the way. Oh, and before we continue, please be reminded that this article contains minor spoilers.How to replay cutscenes in Blue Prince - CCTV footage password guideTo replay cutscenes and watch CCTV footage, you must input a specific password based on clues found in the estate. Similar to our other guides, we look at the hints first just in case you want to discover these on your own. Then, we talk about the actual solution.Where to find the clues for the CCTV recordings?If you're eager to replay cutscenes in Blue Prince, you might want to check a couple of areas for hints:Continue Reading at GameSpot

In Blue Prince, you can unlock both temporary upgrades that fade at the end of the day and permanent upgrades that continue day after day. Some permanent changes are things like adding new floorplans, upgrading rooms, and more.

In A24’s new fantasy adventure movie The Legend of Ochi, the titular apelike creatures communicate via strange chirps and whistles. Their unique vocals become a pivotal plot point, especially after lonely 12-year-old Yuri (Helena Zengel) runs off to return a baby ochi to its colony deep in the forest. For the most part, people like Yuri’s father (Willem Dafoe) view the ochi as aggressive, but Yuri learns that they’re not as dangerous as her village believes.  [Ed. note: This post contains slight spoilers for The Legend of Ochi.] Yuri eventually learns how to communicate with the creature, and later, learns from her estranged mother (Emily Watson) that the ochi sometimes engage in a beautiful, harmonious song all together. Writer-director Isaiah Saxon had a very clear idea of what he wanted the ochi to sound like — a mix of musical birdsong and dolphin noises, but realistically made by a primate. One serendipitous encounter on YouTube really made everything click together.  “I was looking for human sounds, and I looked up the term ‘throat-whistling’ on YouTube,” Saxon told Polygon ahead of the movie’s release. “I found this guy named Paulythebirdman Manalatos. He just had one video on his account, and it was him in a hoodie in his basement. And he’s like, ‘Hey guys, I figured out this thing I can do in the back of my throat. I sound like a bird.’ And he goes, ‘Aah!’ The thing that comes out of his mouth is the actual audio that we used in the movie. I’ve ripped it from YouTube. The first time the ochi speak is that recording from that YouTube video.” Saxon says he reached out to Manalatos, who was immediately drawn to the movie’s script.  “He said, ‘Oh my God, this is my life story. My mom was out of the picture, and I turned to black metal and throat-whistling to express myself,’” Saxon said. “We got him into the booth, recorded him for a couple of days, just going through the whole script, and just like any other actor, hitting every little emotional beat through throat-whistling. It was incredible. In the edit, I mixed in a little bit of mockingbird and raven and whale for the larger adult ochi, but it’s almost all him.” Manalatos taught Saxon how to throat-whistle, though he says he can’t do it to the extent that Manalatos can. And even though Yuri eventually is able to replicate the noises, Zengel didn’t actually make them — she just mimed the process. The technique, it turns out, is very intensive on the vocal cords. “I taught Helena, and she could kind of do it too, but the amount of time she needed to do it, she could have permanently ruined her voice,” Saxon explained. “I was like, That’s medically unsafe.” The Legend of Ochi is out in limited release now, expanding to wide release on April 25.

My brain on AI.Gemini AI for image generationThere’s an itch I can’t quite scratch, a low-humming anxiety beneath the surface of my daily work. It started subtly. I’d use an AI tool to quickly refactor some code, summarize a dense document, or brainstorm ideas. The efficiency gain was undeniable, intoxicating even. But lately, I’ve noticed a shift. I reach for these tools not just for speed, but because the cognitive path feels… easier. Too easy.It reminds me of muscle atrophy. Stop lifting weights, and your muscles weaken. Stop walking, and your endurance fades. I’m starting to fear a similar phenomenon happening in my mind. Am I exercising my critical thinking, my problem-solving abilities, my creativity muscles? Or am I letting them soften, relying on the AI to do the heavy lifting?I find myself generating answers more than truly thinking through problems. There’s a difference. Generating feels like pulling a pre-packaged meal from the freezer — convenient, fast, often looks good. Thinking feels like selecting fresh ingredients, chopping, seasoning, tasting, adjusting — a process that’s messier, slower, but ultimately builds skill and deeper understanding. I worry that by optimizing for the generated meal, I’m losing the culinary art of thought itself.What fades when I outsource my thinking?This touches something deeper than just forgetting facts or syntax that an AI can instantly recall; it feels like it affects the core processes of how I engage with complexity.First, there’s critical judgment. When an AI serves up a polished answer or a block of code, the friction I’d normally encounter — sifting through sources, weighing evidence, spotting flaws — often evaporates. I might skim, nod, and integrate. I realize now that friction was essential; it was the workout. It’s where I learned to sense bullshit, to question assumptions, to evaluate the strength of an argument or the elegance of a solution. Each time I bypass that effort, that judgment muscle gets a little less exercise. My fear is becoming a passive recipient, less able to distinguish the truly insightful from the merely plausible, or even the subtly wrong. In a world overflowing with information and generated content, exercising sound judgment feels like the bedrock skill, the one thing that truly matters. And it seems to be built through hard-won experience, not summoned on demand.Then there’s problem framing. AI tools are getting remarkably good at solving problems once they’re clearly defined. Yet, I believe true value often lies in the murky, upfront work of framing the problem correctly. Understanding the real needs, the hidden constraints, the human context, the ‘why’ behind the ‘what’. This requires nuance, empathy, and a grasp of the bigger picture — qualities that feel inherently human. If I increasingly lean on AI from the start, I risk short-circuiting this crucial diagnostic phase. I might get faster answers, but are they answers to the right questions?Synthesis and original insight also feel threatened. Effective synthesis involves more than just stitching summaries together. It’s about spotting novel connections between disparate ideas, creating something new from existing parts. AI can recombine elements from its training data in incredible ways. But genuine breakthroughs often seem to arise from lived experience, from cross-pollinating ideas from completely different fields, from intuitive leaps that defy predictable correlations. When I outsource the act of assembling ideas, I reduce the chances for those serendipitous mental collisions that spark true originality. I practice efficient recombination, perhaps at the expense of deep invention.My ability to maintain deep focus is another casualty. The temptation to ‘just ask the AI’ when I hit a tricky section of code or a difficult conceptual hurdle is immense. It offers an immediate escape hatch from the discomfort of sustained mental effort. This bypasses the very state of intense concentration — what some call ‘Deep Work’ — that’s necessary for producing high-value, hard-to-replicate results. I trade the satisfaction of wrestling through a complex problem for the quick dopamine hit of an instant solution, potentially rewiring my brain for shallower engagement over time.Finally, there’s the erosion of tacit knowledge. Some understanding isn’t written down; it’s absorbed through the pores by doing. It’s the intuitive feel for debugging a complex system, the gut sense that guides an experienced architect, the subtle understanding of team dynamics. Relying on AI to fix things or provide the path forward might solve the immediate issue, but it prevents me from accumulating that rich, hard-earned, embodied understanding that only comes through struggle and direct experience.My future thinking: partnering, not replacingI don’t intend this as a forecast of doom or a call to unplug entirely. I don’t believe human thinking is becoming obsolete. Instead, I see its highest value points shifting. I envision the future involving me working with the machine, rather than competing against it. The atrophy I fear is a real risk, but the outcome isn’t set in stone.I picture cognition like a multi-layered process. AI is rapidly automating the lower layers: pulling information, spotting basic patterns, handling routine tasks, summarizing, translating, generating standard content. So, where do I, where do we, continue to provide unique and growing value?It seems to lie in the higher-order functions:Judgment and Wisdom: As AI floods the world with content, the human ability to discern quality, truth, relevance, ethical implications, and long-term consequences becomes exponentially more valuable. This requires context, life experience, and a framework of values that models simply don’t possess. I become the curator, the editor, the conscience.Strategic Questioning and Direction Setting: AI needs goals; it needs purpose. My ability to ask insightful, penetrating questions — the questions that define the real problem, that set a meaningful direction — becomes a critical meta-skill. I shift from being an answer-finder to being a question-architect, guiding the powerful tools towards worthwhile ends.Cross-Domain Synthesis and True Creativity: While AI synthesizes within its data, I can connect ideas across completely different fields, drawing on unique experiences and intuition. This is where unexpected innovations often arise. It’s about leveraging my specific knowledge in novel combinations.Empathy and Human Connection: Understanding users, colleagues, the nuances of human interaction, the emotional landscape — these remain profoundly human strengths. Building trust, fostering collaboration, considering the human impact of technology requires sensibilities AI can only simulate.Systems Thinking and Integration: Seeing the entire ecosystem — how technical components interact with market dynamics, user behavior, and social trends — is crucial. AI might optimize a piece, but I need to ensure the whole system works, that it’s resilient, ethical, and serves its intended purpose.Knowing facts AI can retrieve instantly feels less important for my future value than the ability to apply that knowledge effectively, guided by sound judgment and clear purpose. It requires using the immense leverage AI provides without letting it dictate my intellectual agenda.Cultivating my cognitive garden: My antidotes to atrophyAvoiding this mental softening isn’t passive. It requires conscious effort, deliberate practice — tending to my mind like a garden that needs active cultivation. Here’s how I’m trying to approach it:Wielding the tool deliberately:Purposeful Use: I try to be explicit about why I’m using AI for a given task. Is it genuinely freeing me up for higher-order thinking, or am I just avoiding effort? I aim to use it for specific, well-defined tasks where it provides clear leverage, rather than as a default for all thinking.Prompting as Discipline: I’ve found that crafting a truly effective prompt forces me to clarify my own thinking first. What result do I actually need? What are the constraints? What does success look like? This process itself is a valuable cognitive exercise.Critical Review: I treat AI output as a first draft, a suggestion, a sparring partner — never the final word. I actively review, question, and refine it. This keeps my judgment muscle engaged.Embracing cognitive resistance:Manual First: For tasks involving skills I want to preserve, I often try to tackle them manually first. Write the initial outline, sketch the core logic, wrestle with the argument. Then I might bring in AI to help refine, check, or explore alternatives. I need to consciously choose the cognitive stairs sometimes, even when the elevator is available.Seeking Challenge: I try to read books and articles that stretch my understanding, engage in discussions that challenge my perspectives, and occasionally tackle problems just outside my comfort zone. This builds mental resilience.Protecting Deep Focus: I actively schedule and defend blocks of uninterrupted time for focused work. Turning off distractions and allowing myself to sink into complex problems feels essential for producing work of real depth and for maintaining my ability to concentrate.Deepening my specific knowledge:Going Deep: I focus on cultivating deep expertise in my core domains — the intricacies of Android development, system design, maybe even exploring intersections with other interests. Broad AI knowledge is becoming commoditized; deep, specific knowledge built through focused effort and experience feels far more durable and valuable.Connecting the Dots: I actively look for ways to connect ideas from different areas of my knowledge and experience. Reading widely and talking to people with different backgrounds helps spark these cross-domain insights.Elevating questions over answers:Problem Definition First: I’m trying to invest more time and mental energy upfront in deeply understanding and framing the problems I’m working on. Asking “why” multiple times, mapping the context, clarifying the true goal before seeking solutions. This feels like higher-leverage work.Interrogating the AI: When an AI provides an answer, I ask myself: What assumptions is it making? What context is it missing? How might its training data bias the result? This helps me use the tool more critically.Practicing meta-cognition:Thinking About Thinking: I try to regularly step back and reflect on how I’m using these tools and how they’re affecting my thought processes. Is this specific use case empowering me or making me intellectually lazy? Journaling or simply pausing to consider this helps.Explaining to Solidify: Trying to explain a concept clearly to someone else — without leaning on the AI — is a powerful way to test and deepen my own understanding. It forces articulation and reveals gaps.Valuing the analog and disconnected:Offline Thinking: I find immense value in stepping away from the screen. Going for walks, thinking with pen and paper, allowing my mind to wander without digital input often leads to clearer thoughts and unexpected ideas. Boredom can be surprisingly productive.Focused Reading: Engaging with physical books or long-form articles forces a type of sustained attention that feels increasingly rare and valuable.Playing the long game: My human algorithmI see this fear of skill atrophy as a protective instinct, rather than simple pessimism. It’s my mind’s way of saying, “Hey, don’t get too comfortable, don’t outsource the functions that make you fundamentally you.” AI offers staggering capabilities, unprecedented leverage. But leverage needs a firm hand and a clear mind guiding it.My goal is to integrate AI wisely and intentionally, rather than simply resisting it. It means playing the long game — focusing on the sustainable cultivation of the skills that matter most, going beyond just immediate productivity gains: judgment, creativity, critical inquiry, wisdom. These feel like core components of my own human algorithm, the one that allows me to navigate complexity, solve meaningful problems, and build things of lasting value — whether that’s robust software, strong relationships, or a thoughtful approach to the future I’m helping shape.The muscle of my mind doesn’t have to waste away. But keeping it strong in the face of these powerful new conveniences requires conscious choice and deliberate effort, every single day. The challenge lies in using the tools effectively without letting them dictate my thinking, ensuring that the ghost in the machine remains firmly human.p.s: I used AI to restructure sentences and grammar checks within this article. Only after I wrote my thoughts first.References & further: 1. Taste is your unfair advantage2. The 3-Level Prompting System That Makes AI Insanely Useful3. Mind Mapping your way to eternal sunshine….Notebook LM way!Wrestling with skill atrophy in the age of generated thought was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.

You might not be familiar with the phrase "peanut butter platform heels" but it apparently originates from a scientific experiment, where peanut butter was transformed into a diamond-like structure, under very high pressure—hence the "heels" reference.Except this never happened. The phrase is complete nonsense, but was given a definition and backstory by Google AI Overviews when asked by writer Meaghan Wilson-Anastasios, as per this Threads post (which contains some other amusing examples). View on Threads The internet picked this up and ran with it. Apparently, "you can't lick a badger twice" means you can't trick someone twice (Bluesky), "a loose dog won't surf" means something is unlikely to happen (Wired), and "the bicycle eats first" is a way of saying that you should prioritize your nutrition when training for a cycle ride (Futurism).Google, however, is not amused. I was keen to put together my own collection of nonsense phrases and apparent meanings, but it seems the trick is no longer possible: Google will now refuse to show an AI Overview or tell you you're mistaken if you try and get an explanation of a nonsensical phrase.If you go to an actual AI chatbot, it's a little different. I ran some quick tests with Gemini, Claude, and ChatGPT, and the bots attempt to explain these phrases logically, while also flagging that they appear to be nonsensical, and don't seem to be in common use. That's a much more nuanced approach, with context that has been lacking from AI Overviews. Someone on Threads noticed you can type any random sentence into Google, then add “meaning” afterwards, and you’ll get an AI explanation of a famous idiom or phrase you just made up. Here is mine[image or embed]— Greg Jenner (@gregjenner.bsky.social) 23 April 2025 at 11:15 Now, AI Overviews are still labeled as "experimental," but most people won't take much notice of that. They'll assume the information they see is accurate and reliable, built on information scraped from web articles.And while Google's engineers may have wised up to this particular type of mistake, much like the glue on pizza one last year, it probably won't be long before another similar issue crops up. It speaks to some basic problems with getting all of our information from AI, rather than references written by actual humans.What's going on?Fundamentally, these AI Overviews are built to provide answers and synthesize information even if there's no exact match for your query—which is where this phrase-definition problem starts. The AI feature is also perhaps not the best judge of what is and isn't reliable information on the internet.Looking to fix a laptop problem? Previously you'd get a list of blue links from Reddit and various support forums (and maybe Lifehacker), but with AI Overviews, Google sucks up everything it can find on those links and tries to patch together a smart answer—even if no one has had the specific problem you're asking about. Sometimes that can be helpful, and sometimes you might end up making your problems worse. Credit: Lifehacker Anecdotally, I've also noticed AI bots have a tendency to want to agree with prompts, and affirm what a prompt says, even if it's inaccurate. These models are eager to please, and essentially want to be helpful even if they can't be. Depending on how you word your query, you can get AI to agree with something that isn't right.I didn't manage to get any nonsensical idioms defined by Google AI Overviews, but I did ask the AI why R.E.M.'s second album was recorded in London: That was down to the choice of producer Joe Boyd, the AI Overview told me. But in fact, R.E.M.'s second album wasn't recorded in London, it was recorded in North Carolina—it's the third LP that was recorded in London, and produced by Joe Boyd.The actual Gemini app gives the right response: that the second album wasn't recorded in London. But the way AI Overviews attempt to combine multiple online sources into a coherent whole seems to be rather suspect in terms of its accuracy, especially if your search query makes some confident claims of its own. With the right encouragement, Google will get its music chronology wrong. Credit: Lifehacker "When people do nonsensical or 'false premise' searches, our systems will try to find the most relevant results based on the limited web content available," Google told Android Authority in an official statement. "This is true of Search overall, and in some cases, AI Overviews will also trigger in an effort to provide helpful context."We seem to be barreling towards having search engines that always respond with AI rather than information compiled by actual people, but of course AI has never fixed a faucet, tested an iPhone camera, or listened to R.E.M.—it's just synthesizing vast amounts of data from people who have, and trying to compose answers by figuring out which word is most likely to go in front of the previous one.

New devices are still hitting our desks at Engadget at a rapid pace. Over the last two weeks, we've offered up in-depth analysis of cameras, earbuds, GPUs and a portable display. Plus, there are follow-ups on two of this spring's biggest TV shows and a little something for the gamers. Read on to catch up on everything you might've missed in the last fortnight.  Panasonic S1R II If you're looking for a camera that excels at both photos and video that's more affordable than what Sony, Nikon and Canon offer, contributing reporter Steve Dent recommends the S1R II. "The S1R II is Panasonic’s best hybrid mirrorless camera to date, offering a great balance of photography and video powers," he said. "It’s also the cheapest new camera in the high-resolution hybrid full-frame category, undercutting rivals like Canon’s R5 II and the Nikon Z8." NVIDIA RTX 5060 Ti (16GB) Devindra is back with another GPU review, and this time he put the NVIDIA RTX 5060 Ti through its paces. Price hikes are the biggest concern here amidst the current retail market (even before potential tariffs kick in). "On paper, NVIDIA has done a lot right with the 16GB GeForce RTX 5060 Ti," he explained. "It’ll be more than enough for demanding games in 1080p and 1440p, even if you let loose a bit with ray tracing. But it’s also relying on DLSS 4 upscaling for much of that performance, which may make some wary about the 5060 Ti’s actual power." Espresso 15 Pro Espresso Displays is an Engadget favorite as far as portable monitors are concerned, but senior reviews reporter Sam Rutherford argues the company needed to bridge the gap between its more affordable options and its priciest. The Espresso 15 Pro isn't cheap, but it does offer almost everything you'd want. "It features well above average brightness, a sleek but sturdy design and super simple setup," he said. "It also comes with a few special features like Glide and added touch support for Macs that help you get more out of the devices you already own. And thanks to a wealth of accessories, it can adapt to almost any use case." Audio-Technica ATH-CKS50TW2 The idea of wireless earbuds with 25 hours of battery life seems impossible, but Audio-Technica made it happen. The company's ATH-CKS50TW2 lasts twice as long as more premium competition with active noise cancellation (ANC) on, but it blows them away with that mode disabled. A-T's trademark warm, inviting sound profile is on display here too. "More specifically, the stock audio isn’t overly tuned, so bass remains pleasantly thumpy when needed and dialed down when it’s not," I wrote. The Last of Us, Andor and Clair Obscur Expedition 33 Nathan has been keeping up with season two of The Last of Us on an episode-by-episode basis and Devindra penned a full review of the new season of Andor. UK bureau chief Mat Smith spent some time playing Clair Obscur: Expedition 33, noting that the game "does a great job setting up its world in a way that allows everyone to get on board."This article originally appeared on Engadget at https://www.engadget.com/engadget-review-recap-panasonic-s1r-ii-nvidia-rtx-5060-ti-and-more-130005749.html?src=rss

We chatted with Alexandru Costin about the new innovations coming to Firefly

Every now and then, you run into a tool that truly wows you. It’s rare—especially nowadays, when everyone and their cousin is coming out with overhyped AI-centric codswallop that’s almost always more impressive on paper than in practice. And that, if you ask me, makes it all the more satisfying when you track down a tool that really, truly impresses. My friend, today is one of those days. Prepare to have your mind blown. Be the first to find all sorts of little-known tech treasures with my free Cool Tools newsletter from The Intelligence. One useful new discovery in your inbox every Wednesday! Your instant audio enhancer Our tool for today comes from a company you’ve almost certainly heard of. But I’d be willing to wager you didn’t know it offered this off-the-beaten-path treasure. ➜ The gem of which we speak is a simple little web app called, fittingly, ​Enhance Speech​—by Adobe. Enhance Speech lets you upload any audio recording of someone speaking. It’s technically designed for podcasters, but it could be useful for practically anything—a spoken memo, a recorded conversation, even a recorded phone call​. The site takes any recording you feed it and instantly improves its audio quality—removing background noise and enhancing the sound of the speaking so it’s crisp, clear, and easy as can be to hear, no matter how sloppily or in what kind of environment it was recorded. Adobe’s Enhance Speech site couldn’t be much simpler or easier to use. ⌚ It’ll take you about two minutes to perform an enhancement. ✅ And you don’t need to create an account or anything: Just open up the Adobe Enhance Speech site​ in any browser, on any device in front of you. Click or tap the Choose Files button—or drag and drop an audio file directly from your device onto the page, if you’re using a computer. Enhance Speech works with most common audio file formats, including WAV, MP3, AAC, FLAC, and M4A. In a matter of moments, the site will serve up an enhanced version of your recording that you can either play on the page or download. You can play your enhanced audio right on the site or download it with one quick click. You can also just check out the built-in sample on the site’s main page to see how impressive of a difference its enhancements make. It really is something. Enhance Speech is ​entirely web-based​—no downloads or installations required. It’s free to use with recordings up to 30 minutes in length and 500MB in size, with a one-hour-per-day upload limit. You can lift those limits and unlock a variety of advanced features with a premium plan, but that absolutely isn’t necessary for the service’s core features—and the limits are plenty generous for most casual use. Enhance Speech follows Adobe’s standard privacy policy​, which ensures no personal data is ever shared or used in any eyebrow-raising way. Ready for more productivity-boosting goodness? Check out my free Cool Tools newsletter for an instant introduction to an audio app that’ll tune up your days in delightful ways—and a new off-the-beaten-path gem in your inbox every Wednesday!

Music has always been about more than just sound, with album artwork serving as a crucial visual companion to the auditory experience. From the psychedelic covers of the 1960s to the iconic imagery of classic rock albums, these visual elements have helped forge deeper connections between listeners and the music they love. However, as digital streaming has become the dominant way people consume music, the tactile and visual aspects of music packaging have often been reduced to tiny thumbnails on smartphone screens, diminishing what was once a rich multisensory experience. The resurgence of vinyl records has breathed new life into album packaging, with artists and designers seizing the opportunity to create physical experiences that digital platforms simply cannot replicate. OK GO, a band already renowned for their visually innovative music videos, has embraced this renaissance with their latest album “And the Adjacent Possible,” featuring a revolutionary vinyl jacket design that transforms the simple act of opening an album into a moment of wonder and discovery that perfectly complements their boundary-pushing musical approach. Designer: Yuri Suzuki, Claudio Ripol When vinyl enthusiasts first open this remarkable gatefold cover, they’re greeted with an unexpected surprise as a three-dimensional pop-up structure springs to life before their eyes. The intricate design features a striking red geodesic dome perched atop a smaller green dome of identical geometric construction, creating an architectural marvel that seems to defy the limitations of paper and cardboard. This structural achievement immediately distinguishes the album from conventional flat artwork, demanding attention and interaction from the listener before the needle even touches the record. The true magic of this design reveals itself through the clever incorporation of a mirror-like surface beneath the geometric structures. This reflective base creates a perfect symmetrical illusion, transforming the partial domes into what appear to be complete spheres floating mysteriously within the confines of the album cover. The optical effect is both startling and delightful, evoking the childlike wonder of discovering a pop-up book for the first time while simultaneously presenting a sophisticated artistic statement that rewards close examination and contemplation. Far from being a mere decorative flourish, the intricate design elements directly connect to the thematic essence of the album itself. Each carefully engineered fold and geometric pattern serves as a physical metaphor for the album’s exploration of interconnected possibilities, symmetry, and movement. The brilliance of this design lies in its ability to balance simplicity with sophistication. While the pop-up mechanism draws on familiar childhood book techniques, its execution demonstrates remarkable precision and artistic restraint. The limited color palette and geometric forms create a visually striking presentation without overwhelming the senses, allowing the structural elements to take center stage in a way that complements rather than competes with the music contained within the vinyl sleeve. For collectors and vinyl enthusiasts, this innovative packaging transforms the album into something approaching a kinetic sculpture, blurring the boundaries between music, visual art, and interactive design. While the quality of OK GO’s music remains unchanged by its container, the overall experience of engaging with the album becomes significantly richer through this thoughtful design approach.The post Pop-up Book-inspired Vinyl Jacket adds a whole new dimension to OK GO’s new album first appeared on Yanko Design.