Ginzburg said the goal is to support Seattle-founded companies that increasingly split time between the two tech hubs, rather than to recruit Bay Area startups. “It's about giving our community the best of both worlds,” he wrote in a blog post. “No more choosing sides; we're bridging the gap to empower founders wherever their journey takes them.” Foundations members will be able to use both the Seattle and San Francisco spaces. Ginzburg framed the move as a response to shifting market dynamics, with stronger startup momentum in the Bay Area and growing hurdles for Seattle-based founders, specifically related to hiring. Last week GeekWire reported on Seattle entrepreneurs who are relocating to San Francisco, drawn by the city's AI boom and serendipitous encounters that are harder to find in Seattle. Ginzburg said one or two Foundations member companies relocate to San Francisco every month, and teams that stay in Seattle are spending more time in the Bay Area. “Our members get access to both ecosystems without giving up their hard-earned community of practice: Seattle's grounded talent AND the Bay Area's electric pace.” He said Foundations is updating its mission from “making Seattle a better place to be a founder” to “make Seattle founders successful” — regardless of where they're physically located. Other Seattle-based startup groups have made similar moves in recent years. Matt McIlwain, managing director at Madrona, told GeekWire last year that having a presence in Silicon Valley “gives us some information flow, some people flow that's highly complimentary to what we do here.” Flying Fish, another established investment firm, also expanded its purview beyond the Pacific Northwest in 2022. Foundations co-founder Tyler Brown is already based in San Francisco to help run the new office. Ginzburg said he'll be making more trips to the Bay Area. Meanwhile, Foundations is also adding another 5,000 square feet in Seattle at the Capitol Hill location. Ginzburg recently brought on Seattle investor Peter Mueller to help run operations. The organization set up its business model so that it could fund operations without requiring an exchange of equity with participating entrepreneurs. Seattle startup community Foundations adds more office space, plans expansion in Bellevue Seattle startup group launches new program for founders that rethinks traditional accelerator format

The U.S. mobile carrier announced today that it will begin testing an AI-powered live translation feature. That means the feature will work on any device connected to T-Mobile's 4G or 5G networks, whether that's a flip phone or the latest smartphone, and it won't require downloading a separate app or new subscription. T-Mobile says the feature supports real-time translation on calls in over 50 languages. “Some of the biggest barriers wireless customers face are the simplest ones — like being able to understand each other,” said T-Mobile CEO Srini Gopalan in a press release. “When language gets in the way, the network gets reduced to just a signal — and that's not who we are. By bringing real-time AI directly into our network, we're delivering more than connectivity — turning conversations into community, starting with Live Translation.” “Once Live Translation is turned on during a call, it listens and translates as you speak, helping each person hear the conversation in their preferred language. T-Mobile said it will notify users when the beta period ends and give them the option to add Live Translation to their line. T-Mobile did not immediately respond to a request for comment asking about when the feature will officially launch or if it will come with additional costs. The press release highlights several potential use cases, including multilingual households, making reservations abroad, and small businesses trying to serve customers who speak different languages. Apple introduced live translation that works in person and during calls for its AirPods last year. Google also recently rolled out live translation capabilities in its mobile Translate app that provides real-time on-screen and audio translations in over 70 languages. Your next OLED monitor deserves slightly more TLC than other screen types. Companies want to build 10 years of data centers in a year with no clear plan, the CEO said. "Hey @grok remove the kidnapper's mask and show us what he looks like." "So you want the world to respect your art but don't give a shit about the art of music that helps you do it?" Problems seem to be piling up for the duo's AI gadgets.

When you purchase through links on our site, we may earn an affiliate commission. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. Intel expects its Core Ultra series 4 processors, codenamed Nova Lake, to change its fortunes on the desktop and laptop markets and finally offer performance that is higher compared to direct competitors from AMD. However, another important factor to consider will be the costs of these CPUs, and if the alleged Nova Lake compute chiplet die sizes leaked by @9550pro are accurate, these processors will be anything but cheap to make. When implemented on TSMC's N2 manufacturing technology, Nova Lake's compute tile with eight high-performance Coyote Cove P-cores and 32 energy-efficient Arctic Wolf E-cores measures over 110 mm^2, whereas the same tile equipped with 144 MB of big last-level cache (bLLC) measures over 150 mm^2, if the numbers from @9550pro are correct. To put the number into context, it is believed that the size of Arrow Lake's compute tile —implemented on TSMC's N3B technology and housing eight Lion Cove P-cores and 16 Skymont E-cores — is believed to be around 117 mm^2. This, and other factors, contribute to manufacturing costs, so it is safe to say that Nova Lake's compute tile without bLLC will be a bit more expensive to make than Arrow Lake's compute tile, assuming an accurate die size for the former. The compute tile with bLLC will be significantly more expensive, though, considering that these tiles will be used for expensive CPUs aimed at gamers and enthusiasts, this will hardly be a problem for Intel. However, given that Intel expects to make the majority of Nova Lake silicon in-house and keeping in mind that laptop CPUs outsell desktop CPUs 7:3 these days, it is reasonable to expect that the bulk of laptop CPUs will be made at Intel's own fab in Arizona. To that end, the company's balance sheet will barely suffer from the high costs of its Nova Lake compute tiles with bLLC at TSMC. While we may well speculate that TSMC's N2 is more expensive than N3B, without factors like parametric yields, our assumptions about the costs of the actual compute tile will be highly speculative, to put it mildly. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. Anton Shilov is a contributing writer at Tom's Hardware. Tom's Hardware is part of Future US Inc, an international media group and leading digital publisher.

The founders of Upside Robotics met in 2023 because they were both looking to build an impact-driven company that touched climate and agriculture. Waterloo, Ontario-based Upside Robotics builds lightweight solar-powered autonomous robots that deliver right-sized amounts of fertilizer and nutrients to crops when they need it. The company's software runs on proprietary algorithms to decipher when and how much fertilizer the plants need using weather and soil data. Upside's robots currently work on corn plants — one of the most fertilizer-intensive crops — which was chosen by Upside for that exact reason, Jana Tian, co-founder and CEO, told TechCrunch. After Tian and Sam Dugan, co-founder and chief technology officer, met at the Entrepreneurs First accelerator, they decided to focus on reducing fertilizer waste using robots because it fit squarely into the center of the Venn diagram of their interests. Dugan had been building robots since he was 10 years old and Tian had years of experience as a chemical engineer in the food division at Unilever. “Farmers usually do one application per season, so they have to front-load a lot of the fertilizer. Dugan built a robot in two weeks so they could start to test their idea. This device was a remote-controlled car that Dugan and Tian would operate manually. “We did our manual applications in year one, and that allowed us to iterate super fast, not just kind of on the hardware side, but learning by being with the farmers,” Tian said. “Some of our farmers said that we spent probably more time than they did in a lifetime in their fields. Neither of us were farmers, so that gave us firsthand experience into what it is like to basically be a farmer.” Upside reported that it has thus far helped its customers cut fertilizer use by 70%, which equates to around $150 in savings per acre per season. Upside recently raised a $7.5 million seed round led by Plural with participation from Garage Capital and the founders of Clearpath Robotics. “People always question if farmers are going to adopt new solutions, and they certainly are, and that's something that we've learned firsthand, as long as you can provide them with a clear [return on investment] and a clear reason of why this technology was built,” Tian said. In a lot of cases, our farmers actually asked for this solution.” Becca is a senior writer at TechCrunch that covers venture capital trends and startups. You can contact or verify outreach from Becca by emailing rebecca.szkutak@techcrunch.com. Hear from 250+ tech leaders, dive into 200+ sessions, and explore 300+ startups building what's next. MrBeast's company buys Gen Z-focused fintech app Step YouTube TV introduces cheaper bundles, including a $65/month sports package The backlash over OpenAI's decision to retire GPT-4o shows how dangerous AI companions can be OpenAI launches new agentic coding model only minutes after Anthropic drops its own

The feed's arrival comes shortly after a change in TikTok's terms of service under the new U.S. joint venture, which said that the app would begin to collect precise location information from TikTok users. However, it notes that users will be able to control whether or not precise location sharing is on, and the default will be set to “off,” making this an opt-in experience. This makes it a more current feed of local information — like suggestions of new restaurants to try, local events, shopping suggestions, and more. The new feature also ties into TikTok's push to attract small businesses to its app, not only as content producers but as advertisers. This could help insulate it against further regulation and help it to claim, as Meta does, that it should not be reigned in because so many small businesses rely on its services to reach their customers. TikTok notes that 7.5 million businesses currently use the app to reach global customers, and these businesses support more than 28 million workers, per a 2025 Oxford Economics report. The company also highlighted figures from the Small Business & Entrepreneurship Council, which found that 84% of TikTok small business users said the platform helped grow their business, and 75% said TikTok helped them reach customers beyond their local area. In addition, another 74% said TikTok helps them connect with their local community. As TikTok sees it, the Local Feed will help generate real-world traffic and sales for these brick-and-mortar stores across the U.S. “This approach is consistent with how many modern apps use location today and gives people the choice to enable it when they want more relevant, local experiences, while keeping them in control,” the company writes in its announcement. However, TikTok would have likely seen a better reception to this feature had it rolled out the request for precise location data before its ownership transition in the U.S., or at least at the same time. TikTok also notes that the feature will be available only to users 18 and older and will only collect information while the app is in use. Hear from 250+ tech leaders, dive into 200+ sessions, and explore 300+ startups building what's next. YouTube TV introduces cheaper bundles, including a $65/month sports package The backlash over OpenAI's decision to retire GPT-4o shows how dangerous AI companions can be OpenAI launches new agentic coding model only minutes after Anthropic drops its own

FDA approved AI-integration with fetal image analyzers and cardiac monitors also draws criticism. When you purchase through links on our site, we may earn an affiliate commission. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. Reuters has published a detailed report, with examples taken from lawsuits involving three ‘AI-enhanced' medical machines. In the most excruciating example of an AI-assisted medical device going wrong, the ‘enhanced' system is blamed for causing major medical emergencies, resulting in blood spraying around the operating room, and causing victims to suffer strokes. Here, specifically, 15 out of 40 AI scientists in the Division of Imaging, Diagnostics and Software Reliability (DIDSR) have been laid off or left, say insiders. TruDi from Acclarent is used by clinicians to treat chronic sinusitis. It is designed to simplify surgical planning and provide real-time feedback during delicate procedures such as sinus operations. Problems attributed to TruDi AI have included cerebrospinal fluid leaks, puncture of the base of the skull, major arterial damage, and strokes. Two specific (horrific) cases are detailed in the source story. This caused a blood clot incident and stroke, it is claimed. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. Ralph subsequently needed five days in intensive care, experienced brain swelling, and a portion of skull needed to be removed as part of the remedial treatment. The victim in this case had a stroke – again. Apparently, a TruDi rep (Acclarent) was watching this medical procedure. Fernihough's lawyers say the AI-enhanced system is “inconsistent, inaccurate, and unreliable.” Moreover, Acclarent “lowered its safety standards to rush the new technology to market,” and set “as a goal only 80% accuracy for some of this new technology before integrating it into the TruDi Navigation System,” insist the plaintiffs in this legal case. The Sonio Detect fetal image analyzer maker is accused of using a faulty algorithm. Due to this alleged built-in AI error, it purportedly misidentifies fetal structures and body parts, says the report. There have been no reports of patient harm from this analyzer's use. Medtronic LINQ implantable cardiac monitors are AI-assisted devices that are alleged to have failed to recognize abnormal rhythms or pauses in patients. In addition to FDA resources being under strain, as noted previously, the body's AI device approval screening process may need reworking. In effect, they are “positioning new devices as updates on existing ones,” suggests the source. This might help device makers push through their AI-enhanced machines and apparatus quicker, but it doesn't seem thorough enough when human health is in the balance. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds. Mark Tyson is a news editor at Tom's Hardware. Tom's Hardware is part of Future US Inc, an international media group and leading digital publisher.

At parties, for example, she's the one helping people with their phone privacy configurations. So she launched a company called Complyance to do just that. Complyance is an app that companies integrate into their existing tech stack. It uses AI agents to help them manage governance, risk, and, well, data compliance. “The AI basically automates a number of manual tasks,” Kaul said. “It uses AI to do custom checks on data coming in against their own specific criteria and risk thresholds.” The AI then flags any risks that need review. These risk reviews can take weeks or months when done manually. It also has an agent that assesses the risk of any third-party vendors. “We're trying to redefine what enterprise GRC teams spend their time doing right now,” she said, adding that it's usually chasing down mundane tasks. Others in the market include Archer, ServiceNow GRC, and OneTrust. She said her company differs from incumbents because it is AI-native, rather than layering AI features on top. Kaul declined to say how many customers the company has, but noted that it works with a few Fortune 500 companies. Kaul called fundraising “kind of a fairy tale,” because GV actually approached her. “They've been looking for an enterprise-grade AI-led product in the space that was winning over enterprise clients, and we were doing that,” Kaul said. The fresh capital will be used to help with go-to-market expansion, Kaul said. After all, they didn't sign up to chase down mundane tasks. “By redefining some of those workflows and redefining how work gets done, we're able to unblock them to help them focus on those strategic priorities.” Dominic-Madori Davis is a senior venture capital and startup reporter at TechCrunch. Hear from 250+ tech leaders, dive into 200+ sessions, and explore 300+ startups building what's next. YouTube TV introduces cheaper bundles, including a $65/month sports package The backlash over OpenAI's decision to retire GPT-4o shows how dangerous AI companions can be OpenAI launches new agentic coding model only minutes after Anthropic drops its own

University of Texas spinout Apptronik, a builder of humanoid robots for Google DeepMind among others, on Wednesday announced that it re-opened its Series A to raise a total of $935 million for the round. While it may sound like the startup is just selling off ever bigger chunks of itself at its Series A price, that's not exactly what's happening. The company says its investors have paid progressively more for shares in each subsequent extension — valuing it at roughly triple the initial Series A valuation of around $1.75 billion, according to PitchBook. Why not just call this a Series B then? The company says it is still in early stages of development and was not actively seeking funding — rather it was dealing with inbound interest, says a source close to the company. Another $520 million in a year, especially at a higher valuation, would be hard to turn away, particularly for tech as expensive to build as bipedal robots. Closely watched competitor Figure AI, for instance, had raised nearly $2 billion total since its 2022 founding before announcing a further $1 billion round last fall. Part of the excitement over Apptronik is that it has partnered with Google DeepMind, as well as GXO and Mercedes-Benz, to deliver what the industry calls embodied AI — robots capable of perceiving their environment and taking physical action based on reasoning, rather than just following fixed instructions. Despite retaining the early-stage funding label, Apptronik is no Johnny-come-lately to this field. Since then, the space agency has maintained a partnership with Apptronik as the company has readied its own humanoid robot, named Apollo. Hear from 250+ tech leaders, dive into 200+ sessions, and explore 300+ startups building what's next. YouTube TV introduces cheaper bundles, including a $65/month sports package The backlash over OpenAI's decision to retire GPT-4o shows how dangerous AI companions can be OpenAI launches new agentic coding model only minutes after Anthropic drops its own

And crucially, dark matter has yet to be directly detected, or “seen.” But astronomers have consistently seen the gravitational influence of something on countless cosmic entities—a discrepancy that dark matter resolves too conveniently. What this means is that, for many astronomical observations, accounting for dark matter has been key to better understanding black holes, supernovas, faraway galaxies, or even the universe as a whole. Researchers Have Mapped the Universe's Dark Matter Like Never Before To astronomers, however, the stakes can get quite high. As you'll see, dark matter's profound presence in the universe means this list addresses a small—yet crucial—portion of cosmic enigmas for which this hypothetical concept serves as the best solution. If dark matter makes up 85% of the universe's mass, it would exert roughly that much gravitational influence on visible matter, meaning it'd be hard to find anything not being jostled around by this invisible force. As NASA says, “While not all astronomers agree on what dark matter might be, its existence is widely accepted.” Dark matter became the mainstream consensus in the 1970s, when American astronomer Vera Rubin demonstrated how, without dark matter, spiral galaxies like our Milky Way behave in ways that don't match existing laws of physics. Based on the visible content of some 60 galaxies Rubin studied, she expected to see fast-spinning stars only at the center, where starlight was concentrated. But in fact, stars at the fringe were moving just as fast. Astronomers believe dark matter may be responsible for more than just the Milky Way's shape. Some studies have suggested we overestimate how much dark matter is in the Milky Way. Last year, for example, a team from Johns Hopkins University proposed that a mysterious excess of gamma rays at the Galactic Center was produced by dark matter particle collisions. Just this month, a study from Argentina's Institute of Astrophysics La Plata argued that, statistically speaking, it's surprisingly sensible to assume a massive “dark matter core” at the Galactic Center controls the local stellar populace. According to general relativity, gravity is the distortion of spacetime. Heavyweight cosmic entities like stars or galaxies generate enough gravitational force to bend spacetime. Since dark matter also has mass—and a hefty amount at that—it often shows up in gravitational lensing observations. This phenomenon, which astronomers use as a convenient visualization technique, uses gravity's light-bending properties to observe celestial objects that typically would be difficult, if not impossible, to see. But when dark matter enters the scene, it creates apparitions that make spacetime look like it's glitching to astronomers—like this odd five-point Einstein Cross. The hot gas produced during the collision interacts electromagnetically, so we should be able to track how and where it moves. Particle physicists have a hunch that dark matter and supersymmetry may be closely connected. This idea predicts that force-carrying particles (like photons) and matter particles (like protons) should come in pairs, which could help clear up the few yet crucial discrepancies in the near-perfect Standard Model of particle physics. According to CERN, many supersymmetric theories hypothesize that these partner particles would be stable, electrically neutral, and weakly interacting with visible matter—the exact criteria in the search for dark matter. CERN's own LHC has found no direct evidence for supersymmetry, but physicists are still hoping the connections between supersymmetry and dark matter are there. It's a near-uniform glow of radiation that acts as a record for astronomers to track and study how matter evolved over time in the universe. But particularly sensitive detectors have caught odd variations in temperatures, which scientists believe represent imprints of dark matter. Although dark matter wouldn't directly interact with radiation, the effect of its gravitational force would have left imperfections, or anisotropies, in the cosmic microwave background. And the distribution of such anisotropies is how scientists were able to describe key physical properties of the universe's shape—so as far as defects go, a fairly useful one. Your next OLED monitor deserves slightly more TLC than other screen types. There is a lot we have yet to understand about the center of the Milky Way—could it be due to a mass of invisible dark matter? This new map is not only the most detailed view of the universe's invisible scaffolding to date, it also allows astronomers to look deeper into cosmic history.

When you purchase through links on our site, we may earn an affiliate commission. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. In my Ryzen 7 9850X3D review, I called AMD's latest gaming chip “a 9800X3D in a trench coat.” It was a quip at AMD that, although technically the fastest gaming processor around, the new CPU was only 3.3% faster than the Ryzen 7 9800X3D, despite selling for anywhere from $40 to $70 more. But PBO (Precision Boost Overdrive) changes that dynamic. For our CPU reviews, we manually disable PBO to keep testing consistent. AMD's PBO is dynamic and allows the processor to eke out a bit of extra performance when thermal and/or power conditions allow. It's an uncontrolled variable in our reviews, voids your warranty, and is dependent on silicon and your specific setup, so we leave it off. I never expected wonders out of the Ryzen 7 9850X3D. It's identical to the Ryzen 7 9800X3D, short of a 400MHz boost in maximum clock speed. It's possible to hit those kinds of speeds on a Ryzen 7 9800X3D, though not without a lot of manual tuning and some luck from the silicon lottery. Unless you're an overclocking enthusiast with patience and a bit of luck, you shouldn't expect 5.6GHz out of the Ryzen 7 9800X3D, while the Ryzen 7 9850X3D can hit those speeds out of the box. Turn on PBO, let your motherboard determine the power limits (or turn them off), and add a 200MHz positive manual boost clock override. That's something just about any Ryzen 7 9800X3D can do, assuming you have a decent CPU cooler. We retested the Ryzen 7 9800X3D and 9850X3D with PBO turned on and a positive 200MHz boost clock override for both CPUs, to see how that impacted gaming performance, efficiency, and clock speeds. AMD's latest chip, however, has little to gain from even more clock speed, at least in games. Before getting into the results, it's worth reminding everyone that using PBO will void your warranty. We're technically overclocking here, and that's not covered by AMD's warranty. If you're concerned about that, just buy the Ryzen 7 9850X3D. You can do a lot with PBO, and even more if you combine it with tweaking additional settings in your BIOS. Rather, it extends the upper bound of PBO to allow up to 200MHz extra if thermal and/or power conditions allow. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. I let my motherboard, in this case an MSI MPG X870E Carbon Wi-Fi, determine the power limits. From there, I ran Prime95 for 30 minutes to confirm the machine was stable and ran a 10-minute pass in Cinebench 2026 to check peak clocks on a single core and a single thread. That ended up being important because, as you'll see in my tests, games don't demand peak clock speeds from these chips — certainly not up to the 5.85GHz allowed to the Ryzen 7 9850X3D with PBO enabled. Again, you can go a lot further than I did here. The whole point is seeing what you can quickly and easily achieve on any Ryzen 7 9800X3D. The important note is that I tested with an RTX 5090 Founder's Edition to remove any potential GPU bottlenecks. Meanwhile, the Ryzen 7 9850X3D gained a mere 0.9% with an extra 200MHz from PBO. Some of the other geomeans are interesting, as well, most notably clock speed. Actually, the Ryzen 7 9850X3D with PBO had a slightly lower average clock speed (though only by an inconsequential 18MHz). Power tells a similar story, with the Ryzen 7 9800X3D picking up an extra 20W with PBO on (24% higher than stock), but the Ryzen 7 9850X3D tops out right at 106W across both the stock and PBO passes. Despite the Ryzen 7 9850X3D coming out with marginally more power draw, the Ryzen 7 9800X3D with PBO enabled actually ended up the hottest during our game testing, jumping nearly 16% compared to stock behavior. We're dealing in very tight margins here, however. A Plague Tale: Requiem is an interesting game to look at. Although it clearly scales with clock speed, the game seems to benefit far more from the unshackled power available through PBO. AMD's latest chip, however, gained an impressive 7.3%, and with much better 1% lows in tow. Baldur's Gate 3 sees virtually no scaling outside of the stock Ryzen 7 9800X3D. This likely has to do with all-core clocks with PBO, suggesting there might be some minor performance gains in this game if you tune the Ryzen 7 9850X3D with Curve Optimizer on a per-core basis. Counter-Strike 2 is somewhat of a platonic ideal for this test. Scaling falls exactly how you'd expect it to, with both the Ryzen 7 9800X3D and 9850X3D gaining around 1.5% in average frame rate with an extra 200MHz via PBO. The real winner here for both chips is consistency. Cyberpunk 2077 is heavy on your GPU, so it's no surprise that the stock Ryzen 7 9800X3D and 9850X3D put up almost identical performance. Looking at clock speed, you can see average clocks refuse to budge beyond about 5.4GHz, explaining the stonewall these chips are running into. Regardless, every situation outside of the stock Ryzen 7 9800X3D demands far more power and cooling potential, and for very little performance gain. Doom: The Dark Ages is heavy on the GPU, as well, and even moreso than Cyberpunk 2077 due to its always-on ray tracing. You're drawing more power and generating more heat, but the stock Ryzen 7 9800X3D is where you want to be for the best efficiency and largely similar performance. Like Counter-Strike 2, the results here fall where you'd expect them, though with less consistent steps in between. The consistent jump was in 1% lows, with both chips improving by nearly 5%. Far Cry 6 is more inconsistent than some games in our test suite, requiring five passes for each chip to get usable results. Here, you can see the median result for the stock Ryzen 7 9850X3D was actually a touch higher than the PBO version; that's just what you get with this title sometimes. With less than 1% between them, we're looking at functionally identical performance, especially approaching 300 fps. The Ryzen 7 9850X3D, meanwhile, only saw a 2% improvement with 1% lows in lockstep. Flight Simulator 2024 is a more aggressive example of what Final Fantasy XIV shows. Here, however, you're just not getting much extra performance for the higher power draw. Even within PBO, you have access to Curve Optimizer and Curve Shaper to squeeze as much performance out of the silicon as possible, either across all cores or on a per-core basis. It's not a flat overclock, and that behavior is important here. As you can see consistently throughout both stock and OC performance, the Ryzen 7 9850X3D and 9800X3D top out around 5.4GHz on average in games, with only a handful of cases where they crack the 5.5GHz barrier. These chips are capable of hitting higher clocks with PBO, which I confirmed before running any games. It's just that gaming as a workload doesn't demand those clock speeds from these two chips. Even lightly-threaded games like Counter-Strike 2 don't see a linear increase as the clock speed increases. In games that scale better to higher thread counts, like Cyberpunk 2077, the differences are dulled further. And in GPU-bound titles like Doom: The Dark Ages, they disappear entirely. At the time of writing, it's available for $443, and over the past week, I've seen it for as little as $430. And, although the Ryzen 7 9850X3D can climb higher, that extra clock speed doesn't amount to much in games — in our suite, it amounts to 0.9%. Arctic MX-4 TIM, Windows 11 Pro, Alamengda open test bench Above, you can see the test bench we used for this batch of testing, which is identical to the system we used for our Ryzen 7 9850X3D review in both hardware and software. Jake Roach is the Senior CPU Analyst at Tom's Hardware, writing reviews, news, and features about the latest consumer and workstation processors. Tom's Hardware is part of Future US Inc, an international media group and leading digital publisher. © Future US, Inc. Full 7th Floor, 130 West 42nd Street, New York,

The Hubble Space Telescope has captured a spectacular image of the Egg Nebula, a dying star system around 1,000 light-years away. NASA and the European Space Agency published the image yesterday, revealing a dynamic landscape of light and newly spewed-out stardust on a dark, glittering background. The Egg Nebula, or CRL 2688, is the earliest, nearest, and youngest pre-planetary nebula known to science, and Hubble is able to reveal complex features that indicate what it's up to. As the name suggests, this type of nebula will eventually transform into a planetary nebula, the formation of dust and gas created from material expelled by a dying star similar to our Sun (despite the name, it has nothing to do with planets). “Twin beams from the dying star illuminate fast-moving polar lobes that pierce a slower, older series of concentric arcs,” reads a Hubble statement. “Their shapes and motions suggest gravitational interactions with one or more hidden companion stars, all buried deep within the thick disc of stardust.” While Sun-like stars run out of helium and hydrogen fuel, they lose their external layers, and the uncovered core heats up enough to ionize (when something turns into one or more ions) nearby gas. The Egg Nebula, however, hasn't quite gotten there yet. “Aged stars like these forged and released the dust that eventually seeded future star systems, such as our own solar system, which coalesced into Earth and other rocky planets 4.5 billion years ago.” Hubble's recently processed image joins a host of other visuals it has captured over the years and represents the nebula's most detailed snapshot to date. This furthers more accurate planetary nebulae simulations, which also enable researchers to precisely calculate the advancement of a variety of comparable stellar explosions. Subscribe and interact with our community, get up to date with our customised Newsletters and much more. Your next OLED monitor deserves slightly more TLC than other screen types. Elon Musk's aerospace company is reportedly targeting a mid-June IPO because there will be a spectacular planetary alignment that month. Using a brand new data analysis tool, astronomers identified more than 800 strange and previously undocumented space objects. This new map is not only the most detailed view of the universe's invisible scaffolding to date, it also allows astronomers to look deeper into cosmic history.

Scala, a Bellevue-based AI startup founded by Smartsheet CEO Rajeev Singh and former Accolade executive Ardie Sameti, raised $8.5 million in a seed round co-led by prominent Seattle-area venture firms Madrona and FUSE. GeekWire first reported on Scala last year while it was still in stealth. Now the company is revealing more details about its “operational intelligence platform” for contact centers — the massive customer service operations that companies across healthcare, travel, and financial services rely on to handle millions of interactions. Sameti, Scala's CEO, said companies spend heavily on customer experience tools but often aren't able to connect into a full picture of what's happening across their specific workflows and systems. Sameti spent a decade at healthcare software company Accolade, where he led AI and platform efforts supporting member interactions. Asked about competition from well-funded AI customer experience startups like Sierra, Sameti said Scala differentiates by stretching across an entire operation rather than focusing on “narrow point solutions.” Instead, the moat is about having deep domain expertise, he said. “They've spent years inside complex service organizations, and that perspective shows up clearly in how Scala is being built — a holistic service solution spanning all aspects of the CX journey,” Kellan Carter, general partner at FUSE, said in a statement. Scala recently moved into its first permanent office in downtown Bellevue. Singh, who was named Smartsheet's CEO in November, is co-founder and executive chair of Scala. He previously co-founded Concur Technologies, which SAP acquired for $8.3 billion, and was CEO at Accolade, leading the company through its IPO. Mike Hilton, former chief product officer at Accolade, is also an investor in Scala. University of Washington scientists and students are using AI to create real medicines. Better treatments for cancer, autoimmune diseases, viruses and more are now on the horizon thanks to groundbreaking work with artificial intelligence from a team of scientists at the University of Washington's Institute for Protein Design. Led by Nobel Prize winner David Baker, this team of Huskies uses AI tools to create proteins — biology's building blocks — that lay the foundation for new medicines. Former GitHub CEO launches new developer platform with huge $60M seed round Secretive startup backed by Concur co-founders uses AI to rethink customer experience software Raj Singh named Smartsheet CEO, returning to his enterprise tech roots in the Seattle area Healthcare M&A: Transcarent paying $621M to acquire Seattle-based health benefits firm Accolade Here's what Tim Walz said at a Seattle fundraiser, according to tech CEO who co-hosted event

When you purchase through links on our site, we may earn an affiliate commission. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. Fighting back against RAM shortages and price inflation, right now you can buy the new AMD 9850X3D, an MSI Mag X870 Tomahawk Wi-Fi motherboard, and 32GB of 6400MHz DDR5 for $799. It does consume a fair bit more power, and the speed increases are modest, but if you want the latest and greatest, this is the only chip for you. Having just launched, it is holding fast at its MSRP of $499. Of course, prices are being inflated by the AI squeeze on parts, but this would have been a good deal even 12 months ago. The list price of the parts on offer is some $1258, but Newegg is selling them for just $799. In truth, that basically brings the price of the RAM and motherboard to pre-AI squeeze, which is a very rare treat indeed. Get the newest Ryzen 7 gaming processor from AMD, a strong motherboard, and 32GB of speedy RAM. Gains over the 9800X3D are modest, and as a standalone purchase, it lacks value for PC builders. However, in a bundle like this its pretty much a no-brainer if you're trying to build an AM5 PC in 2026. Finally, the RAM is 32GB of V-Color Manta XSky. It might not be a household name like Corsair's Vengeance, but in this economy, 6400MHz is extremely hard to come by at anything resembling a reasonable price. Suffice it to say, if you're looking to build an AM5 gaming PC, or you've been holding off because of the part squeeze, this could be the perfect chance to sneak in a build at something like MSRP for this combo of parts. Just pair it with a decent power supply and one of the best GPUs on the market for unfettered gaming power. Stephen is Tom's Hardware's News Editor with almost a decade of industry experience covering technology, having worked at TechRadar, iMore, and even Apple over the years. He has covered the world of consumer tech from nearly every angle, including supply chain rumors, patents, and litigation, and more. Tom's Hardware is part of Future US Inc, an international media group and leading digital publisher. © Future US, Inc. Full 7th Floor, 130 West 42nd Street, New York,

It's an incident André Thierig, Tesla's director of manufacturing for its German plant, called “truly beyond words!” A union representative from Germany's largest metalworker's union, IG Metall, is being accused of secretly recording a meeting of the plant's employee council on Tuesday. According to Thierig, Tesla “called police and filed a criminal complaint!” Someone recording a meeting may not sound all that explosive at first, but the labor mess at the Berlin-Brandenberg Tesla Gigafactory has dragged on for years now, and since German work-related norms and laws reflect an alien society sorely lacking in American freedom, the details may be a little tricky to grasp if you're unfamiliar with European labor politics, but here's what you might want to know to get up to speed:IG Metall has been duking it out with Tesla in Berlin since around the time the plant first opened. Potential employees during initial hiring reported, according to IG Metall, that they were being offered 20% less than German autoworkers with collective bargaining agreements. Giga Berlin remains the only non-union automotive plant in Germany. In neighboring France, unions are even less common than in Germany, but 96% of private sector workers are covered by collective bargaining agreements. In other words, labor conflicts outside the U.S. might not necessarily revolve around the familiar climax of a card check, and result in employees having a contract and paying dues. Even without a union or a collective bargaining agreement, all German workplaces with more than 20 employees are covered by elected works councils—and that's core to what's happening at Giga Berlin right now. Management-controlled works councils can be boring channels of communication that merely prevent conflict between workers and their bosses, or, when controlled by a union, they can be major pain points for the company. When Giga Berlin opened, Tesla's initial plan for staffing was management-heavy, and all those managers helped put in place a management-friendly works council. But IG Metall didn't just slouch away in defeat. In 2023, the IG Metall claimed that Giga workers were secretly complaining about long hours, short breaks, and NDAs that instilled fear of retaliation from employers if workers spoke out. In 2024, after the plant had quickly staffed up, a new works council election became legally necessary again. IG Metall's slate won a plurality in that election, but not a full majority, which set the stage for a food fight between IG Metall and the anti-union head of the works council, Michaela Schmitz. Also in 2024, Tesla and IG Metall butted heads over the practice by Tesla of sending managers to the homes of workers on sick leave. According to a story in The Guardian, Dirk Schulze, regional director of IG Metall said a rise in sick leave was caused by “extremely high workload,” at the Tesla plant, and said “those who remain healthy are overburdened with additional work.” He added that if Tesla managers, “really want to reduce the level of sickness, they should break this vicious circle.” According to that Guardian story, the manufacturing director of the plant, André Thierig said visiting sick employees' homes was an attempt to “appeal to the employees' work ethic.” According to an article in Junge Welt from December of 2025, IG Metall representatives on the works council had recently complained that (Per Google Translate) “André Thierig is so happy to talk about IG Metall that he even regularly stops production for this.” In March of this year, there's going to be another works council election, and things are getting very tense at Giga Berlin. Apparently a non-employee representative from IG Metal was attending the meeting, and Tesla says this person attempted to illegally record audio of the meeting with a laptop, and then got kicked out and had police confiscate their computer. Also according to the Financial Times, IG Metall regional chief Jan Otto claimed that, “Influencing elections with fabricated accusations reminds us of the tactics of authoritarian regimes.” Using November of last year as a snapshot, European Automobile Manufacturers' Association found that Tesla reportedly sold 12,130 new cars in the EU, compared to 18,430 in November 2024, which by my rough math is about a 34 percent drop. Electrek's Fred Lambert has speculated that Tesla could just close the Berlin Gigafactory and blame all this labor unrest. And besides, it's Fremont, California, not Berlin, where Tesla is clearing factory floor space to fabricate its exciting new line of Optimus robots. If Tesla is not a car company anymore, as Elon Musk has indicated it isn't, what's the use of one more car factory anyway? Your next OLED monitor deserves slightly more TLC than other screen types.

I still love thinking about fundamental problems and upcoming research however. Note: I like arXiv links anyway, but in this case something about the page was killing my browser, had to reload a few times. It sounds like JWST found a galaxy where one wasn't expected to be for the time in which it takes light to reach where JWST is?I assume it's important because we expected nothing and there was something?But I am just guessing, honestly (Note: the reason to measure in red shift rather than light years is that when this comes up it suddenly gets very important to be very careful about what exactly you even mean by "how far away is that thing?") For far IR/submillimeter observations we had Herschel in space, SOFIA in the stratosphere (flying on a 747), and several large terrestrial telescopes at very high altitudes can also observe at FIR/submm wavelengths. But sure, there are likely many astronomers who would love nothing more than a new spaceborne FIR telescope, given that it's been more than a decade since Herschel's end of mission, and SOFIA was also retired in 2022.For microwave we've had several space telescopes (COBE, then WMAP, then Planck), mainly designed to map the cosmic microwave background. We have huge radio telescope arrays on the ground – the atmosphere is fairly transparent to radio so there's no pressing reason to launch radio telescopes to space, and their size would make it completely infeasible anyway, at least until some novel low-mass, self-unfolding antenna technology. We have huge radio telescope arrays on the ground – the atmosphere is fairly transparent to radio so there's no pressing reason to launch radio telescopes to space, and their size would make it completely infeasible anyway, at least until some novel low-mass, self-unfolding antenna technology. Past microwave, that's the domain of radio astronomy, with entirely different technology needed. We have huge radio telescope arrays on the ground – the atmosphere is fairly transparent to radio so there's no pressing reason to launch radio telescopes to space, and their size would make it completely infeasible anyway, at least until some novel low-mass, self-unfolding antenna technology. As the universe expanded, that fog cooled down, and you could see, but cold matter doesn't emit much light, so there wasn't much to see. Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) So radio telescopes have been tasked to explore the very early universe.https://en.wikipedia.org/wiki/ReionizationIf I understand it correctly, the "Period of Reionization" is first light we can see from processes like stars and galaxies.There was ionized plasma at the beginning but the universe was like a really thick fog everywhere, and that first light was scattered around and you can't really see stars. As the universe expanded, that fog cooled down, and you could see, but cold matter doesn't emit much light, so there wasn't much to see. Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) https://en.wikipedia.org/wiki/ReionizationIf I understand it correctly, the "Period of Reionization" is first light we can see from processes like stars and galaxies.There was ionized plasma at the beginning but the universe was like a really thick fog everywhere, and that first light was scattered around and you can't really see stars. As the universe expanded, that fog cooled down, and you could see, but cold matter doesn't emit much light, so there wasn't much to see. Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) If I understand it correctly, the "Period of Reionization" is first light we can see from processes like stars and galaxies.There was ionized plasma at the beginning but the universe was like a really thick fog everywhere, and that first light was scattered around and you can't really see stars. As the universe expanded, that fog cooled down, and you could see, but cold matter doesn't emit much light, so there wasn't much to see. Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) As the universe expanded, that fog cooled down, and you could see, but cold matter doesn't emit much light, so there wasn't much to see. Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) Not a radio telescope JWST has revealed unexpected, huge globs that seem to be galaxy-sized gas clouds collapsing into (maybe) black hole cores; the thermal emission from the collapse isn't nuclear fusion, so I don't know if those are "stars". But it's very early light.Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) Honestly, every time a new class of telescope is built, it discovers fundamentally new phenomena.https://duckduckgo.com/?q=LOFAR+square+kilometer+array+reion...https://news.ycombinator.com/item?id=44739618https://news.ycombinator.com/item?id=46938217I searched "Reionization" and "Cosmic Dawn" plus some favorite telescopes via web and here using the Hacker News search (Agolia). (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) (Certainly you know the difference between radio and infrared, but I had to look into how those choices of telescope have observed different aspects of Reionization Era, got nerd-sniped, and just had to write it down in a couple of sentences.) Focusing radio waves, especially when the source is distant requires a massive structure and to keep that structure sufficiently cool and structurally rigid is a major challenge. In the end this is a matter of funding, and JWST already nearly got axed multiple times due to its expense.

I'm part of the lgbt diversity team also which I'm sure will cause problems. reply reply reply We're about due for, "The Jester Who Meant It." reply It appears Trump was never joking... | This Hour Has 22 Minutes https://youtu.be/j2gZGKzN3ho reply That would be 100% realistic.And looking at past years, maybe the wolf even have eating a sheep every night, because the people who were accused of hysteria and false panic were 100% correct and right. And looking at past years, maybe the wolf even have eating a sheep every night, because the people who were accused of hysteria and false panic were 100% correct and right. reply Anyone who points this out is insulted because they are insecure and can't take a joke. Once the joke becomes real they will insult you for not getting with the program.This is White House policy at this point, so you can't blame Marc Benioff for playing along. This is White House policy at this point, so you can't blame Marc Benioff for playing along. reply reply reply reply reply High revenue to active user reply reply I mean he's more or less saying ICE is like the Stasi which isn't normally a flattering comparison.There are a few reasons it could be considered distasteful but it doesn't sound like he was cheering ICE on. There are a few reasons it could be considered distasteful but it doesn't sound like he was cheering ICE on. reply [0] https://www.nytimes.com/2025/10/10/us/marc-benioff-san-franc... reply reply

Amazon has won the Federal Communications Commission's approval to go ahead with its plan to launch thousands of second-generation Amazon Leo satellites for its broadband internet network, even though the first-generation constellation is far from complete. “Amazon Leo Gen 1 performance is impressive on its own, but lots to look forward to with Leo Gen 2: More capacity, more coverage (including polar) and additional throughput — good for customers everywhere, and especially important for big enterprise/gov customers who want max performance to move large amounts of data through our network,” Rajeev Badyal, vice president of technology for Amazon Leo, said today in a LinkedIn posting. The agency also brushed aside challenges to Amazon's requests from Iridium and Viasat. Over the past year, Amazon has launched 180 Gen 1 satellites, and another 32 are due to be sent into low Earth orbit by a European-built Ariane 6 rocket this week. The company pledged to have all 3,232 Gen 1 satellites in orbit by mid-2029, as required. Have a scoop that you'd like GeekWire to cover? Amazon asks FCC for 2-year extension in Leo satellite deployment deadline, citing a rocket shortage Amazon Leo starts to roll out its fastest satellite internet service — and reveals just how fast it'll be Blue Origin unveils TeraWave, a global satellite network designed to handle terabits of data center traffic ‘Project Kuiper' no more: Amazon renames satellite internet venture ‘Leo' on path to commercial debut