Loading summary
A
Next week is the start of ces, which used to stand for Consumer Electronics show but now stands for nothing, which sadly means it'll fall for anything. Let the Rickrolls begin the Nvidia RTX 5080 has been spotted, or at least the box of an MSI version has, confirming that the card will come with 16 gigabytes of DDR7VRAM and a 256 bit memory bus. Which is exactly what the leak said, at least the ones that weren't Rickrolls. The box images don't confirm anything else, but thankfully we've got Twitter people for that leaker Hongsheng2020 tweeted that the RTX 5090 will actually feature a 575 watt power rating instead of the 600 watts predicted by previous leaks, which was confirmed by the prominent leaker, whose name a linguistic expert has assured me is pronounced kopitae7kimi. That expert is me. Anyway, Ms. Kimmy also said the RTX 5080 will come in at 360 watts lower than previous leaks had indicated, while still higher than its predecessor. But the 5080 might as well be a plate of Spaghetti with a PCIe slot if we don't know the price. Thankfully we got a hint from a photo of a listing for a custom 5080 made by ASUS €1700. Now hey, that includes a 20% VAT tax and a custom Asus board will probably cost above MSRP, so this could equate to around 1350 USD. Oh, there's another clue. The listing says DLSS4, which Kopite showed up once more to confirm is real source. I made it up. Current rumors point to at least the RTX 5080 being actually released on January 21st. But Nvidia's CES keynote is at 6:30pm Pacific on Monday, so we may cover it on Monday's episode, even though that's when I was supposed to have dinner with my family. This is more important though. But even when it comes to official announcements, the great big giant water balloon that is Cesar just can't hold it all in sloshing around. Sorry. Oh, spilled a bit. Asus, Samsung and MSI all just announced the world's first 27 inch 4K OLED 240Hz monitor. They each have their own version, so which one was actually first? Well it looks like they're all using Samsung displays 4th gen QD OLED, so I guess it was Samsung. But what if you want high refresh rate instead? MSI just announced the world's second 600Hz monitor after Acer announced the world's first one back in September. Wowie. Zowie. Oh, that's not me reacting. Zowie is BenQ's gaming brand and they also just announced a 600Hz monitor. Yowza. That was real. Hey, look at this. Pimax announced new VR headsets with micro OLED panels and and there's even one called the Dream Air that doesn't look like a whole PC strapped to your face. Although if you prefer that look, there's also the new pimax Crystal Super. And while I'm anxious to see the new rumored SteamOS handheld at CES, I'm also intrigued by the new Atari game Station Go, which packs the company's iconic trackball paddle and keypad controllers into a gaming handheld, which will either make using it feel like you're hacking a cyberpunk system or playing Bop It. Both are valid Meta has taken down the Instagram and Facebook accounts of more than two dozen AI powered profiles after they went viral for making a mockery of the authenticity of human experience, something Meta thought was sorely lacking on the Internet. For some reason. Backlash over the AIs grew on social media this past week, with many posts saying saying Meta had just added them to their platforms. But the AI accounts were actually launched way back in 2023 and have been leaving their AI excretions all over Facebook and Instagram throughout 2024. You had fake AI people pretending to watch movies at the drive in, or sew clothes for their fake grandkids, or spend winter break at the beach with their family. Oh yeah, my children are so funny and cool. Imagined with AI. This week's backlash seemed to have been kickstarted by a Meta exec telling the Financial Times the company plans to continue rolling out these bots for a purpose that I think you can only comprehend if you've had at least one VC funding round, because I don't all 27 of these AI accounts have now been wiped, although Meta told CNN they haven't deleted the accounts, they're just taking them down temporarily to fix a bunch bug that prevented people from blocking them. So don't worry, you will be able to block them when they come back with all of their fake AI friends to take over for humans on Facebook. Look, you're not using it and someone has to and you simply must check out our sponsor Headspace, the perfect companion for people looking to better themselves going into the new year. Whether you want 2025 to be the year you get better sleep, reduce your stress, or find more focus. Headspace can help you build lasting, healthy habits with guided meditation. Sess catered to your goals, which crucially, don't have to be the same as everyone else's. Everyone's different. Trust me, I know. And so does Headspace. So they'll help you choose whatever routine, frequency, or type of practice that works best for you. There's sleep casts, soundscapes, and exercises to improve sleep quality. Programs cater towards creativity, productivity, and mindful eating, and all sorts of personalized and customizable meditation techniques to suit your individual needs. If one of your needs is trying things out for free, that's perfect. Use the link in the description or scan the QR code on screen and try Headspace for free for 60 days. Thanks to Headspace for sponsoring this video and thank you for sharing this time with me. It was special Ah, crazy. Email sender it's an initialism. It has to stand for something. You can't just call it C.E. the FCC voted to reinstate net neutrality rules in April 2020 after the rules were repealed by the Trump administration back in 2017. But a US appeals court has now blocked the FCC from restoring net neutrality, citing the Supreme Court's overturning of the Chevron doctrine last year. Without the Chevron doctrine, federal agencies have less power to make up rules, using arbitrary nonsense like expert opinion and multivariate analysis. Sounds like vibes to me. But hey, at least the US Government is getting in on the nuclear energy craze. They just announced a deal with major provider Constellation. Totally unrelated, but it's almost CES Week. This is kind of how it's gonna be. Apple has proposed paying $95 million to settle a class action lawsuit over Siri listening to private conversations, which was filed after reports in 2019 of Apple contractors listening to such recordings. Owners of Apple devices could get up to 20 bucks per device, to a maximum of five if they swear under oath that Siri was accidentally activated during a confidential or private conversation. Apple also apologized back in 2019 for this and said they wouldn't keep private recordings like that anymore without user permission. And so it hasn't happened since the end. Taiwan's foreign minister would like the world to know he absolutely did not agree with the idea of blowing up the world's foremost semiconductor manufacturer, tsmc. Okay, I don't know. Not sure where anybody got that. Well, the minister was answering a question about Donald Trump's pick for under secretary of Defense policy, Elbridge Colby, who in 2023 endorsed the idea that if China invades Taiwan, the US should not allow TSMC to fall into its hands intact, and in 2024 said destroying TSMC is table stakes in that situation. He's not alone. Rapidly disassembling TSMC is an idea shared by others since at least 2022 in US military academia, which is a thing. And Taiwan has also been saying no to the same idea for just as long. Just let us blow it up a little bit. No, stop. And it's been a while since we've seen someone run the classic shooter Doom on something wacky. So just in time. There's now a captcha that requires you to play the game to prove you're human. Just get three kills on nightmare difficulty or you're provably not a sentient being. It's simple. It was made by Guillermo Rauch, the CEO of Vercel, the company behind the V0 coding chatbot. He even shared the prompts he used to make what is actually not the world's first Doom based CAPTCHA. In 2021, Mikel Camps Orteza made one, but it was more like a 2D flash game using Doom assets. That one let so many bots through. And now we have ChatGPT. Thanks a lot. And now an unsarcastic Thanks a lot to you for watching. Tell you what, come back on Monday for more tech news and I won't even make you play Minesweeper to prove you're human like I normally do before I speak to anyone.
Title: RTX 5080 Leaked, CES Floodgates Open, Meta Deletes AI Bots & More!
Host: Linus Media Group
Release Date: January 4, 2025
[00:00]
As CES 2025 approaches, the excitement is palpable, albeit tinged with skepticism. The host humorously notes, "CES used to stand for Consumer Electronics Show but now stands for nothing, which sadly means it'll fall for anything" (00:03), hinting at the event's susceptibility to gimmicks like Rickrolls.
The spotlight swiftly shifts to NVIDIA's latest anticipated graphics card. Leaked details reveal the RTX 5080, with confirmed specs including 16GB of DDR7 VRAM and a 256-bit memory bus. The host remarks, "Which is exactly what the leak said, at least the ones that weren't Rickrolls" (02:15), indicating the mix of genuine and deceptive leaks surrounding CES.
Further insights from the leaker Hongsheng2020 on Twitter clarify that the RTX 5090 will feature a 575-watt power rating, slightly lower than earlier predictions of 600 watts. The host adds a humorous twist on pronunciation, "a linguistic expert has assured me is pronounced kopitae7kimi. That expert is me" (04:30), showcasing their playful engagement with the community.
Price speculation for the RTX 5080 emerges from an ASUS listing priced at €1700, which, after accounting for a 20% VAT and potential markup, translates to approximately $1350 USD. The mention of DLSS4 in the listing adds credibility to the leak, with the host cheekily admitting, "I made it up" (06:45), highlighting the blurred lines between fact and fabrication in tech leaks.
The anticipated release date is tentatively set for January 21st, just in time for NVIDIA's CES keynote scheduled at 6:30 PM Pacific on Monday (08:00). The host juggles excitement with personal commitments, humorously stating, "even though that's when I was supposed to have dinner with my family. This is more important though" (09:15), underscoring the significance of the announcement.
The podcast transitions to groundbreaking monitor announcements from major brands:
Asus, Samsung, and MSI unveil the world's first 27-inch 4K OLED 240Hz monitors, each presenting their unique takes on the technology. The host muses, "so which one was actually first? Well it looks like they're all using Samsung displays 4th gen QD OLED, so I guess it was Samsung" (12:30), indicating Samsung's pivotal role in this advancement.
MSI further impresses with the announcement of the world's second 600Hz monitor, following Acer's initial release in September (14:00).
BenQ's Zowie brand also joins the fray with its own 600Hz monitor, prompting an enthusiastic "Yowza" from the host (15:45).
These high-refresh-rate monitors signify a leap forward for both gaming and professional use, catering to enthusiasts seeking unparalleled display performance.
The discussion shifts to advancements in virtual reality and handheld gaming devices:
Pimax introduces new VR headsets equipped with micro OLED panels, including the Dream Air model, which boasts a sleeker design that diverges from the bulky PC-attached systems. There's also the Pimax Crystal Super, which caters to users preferring a more traditional, robust VR setup (18:20).
Atari makes a nostalgic return with the Game Station Go, a handheld that integrates the company's iconic trackball paddle and keypad controllers. The host humorously contemplates its user experience, "which will either make using it feel like you're hacking a cyberpunk system or playing Bop It. Both are valid" (20:50), reflecting on the device's blend of retro and modern gaming elements.
While anticipation builds for a rumored SteamOS handheld at CES, these new releases highlight the industry's push towards more immersive and portable gaming experiences.
A significant portion of the episode delves into Meta's recent actions regarding AI-generated profiles on its platforms:
Meta has removed over two dozen AI-powered accounts from Instagram and Facebook, which were criticized for undermining the authenticity of human experiences online. These profiles featured activities like "watching movies at the drive-in" and "sewing clothes for their fake grandkids" (23:10).
The backlash intensified when a Meta executive informed the Financial Times of the company's intent to continue deploying these bots, ostensibly linked to investor relations and strategic goals. The host sarcastically comments, "for a purpose that I think you can only comprehend if you've had at least one VC funding round" (25:00), critiquing Meta's rationale.
Meta clarified to CNN that the deletions were temporary, aimed at resolving a bug that hindered users from blocking these AI accounts. The host mockingly reassures listeners, "you will be able to block them when they come back with all of their fake AI friends to take over for humans on Facebook" (27:30), highlighting ongoing concerns about AI authenticity on social media.
This segment underscores the tension between technological advancements in AI and the preservation of genuine human interactions online.
The podcast covers several legal and regulatory developments impacting the tech landscape:
Net Neutrality: The FCC's attempt to reinstate net neutrality, originally rolled back in 2017, faces a setback as a US appeals court blocks the restoration. This decision cites the Supreme Court's overturning of the Chevron doctrine last year, which diminishes federal agencies' authority to enact rules based on "arbitrary nonsense like expert opinion and multivariate analysis" (30:00). The host quips, "Sounds like vibes to me," reflecting skepticism about the court's reasoning.
Apple's Siri Privacy Lawsuit: Apple has agreed to a $95 million settlement in a class-action lawsuit alleging that Siri inadvertently listened to private conversations. Initiated after 2019 reports of contractors accessing recordings, the settlement offers up to $20 per device to affected users, with a maximum of five per individual under oath. Apple reiterated its commitment to user privacy, stating, "we won't keep private recordings like that anymore without user permission" (33:15).
These updates highlight the ongoing challenges tech companies face in balancing innovation with legal compliance and user trust.
Geopolitical dynamics are also addressed, particularly concerning Taiwan's semiconductor industry:
Taiwan's Foreign Minister has publicly denied any intent to "blow up the world's foremost semiconductor manufacturer, TSMC," amidst tensions stirred by Donald Trump's 2023 endorsement of policies threatening TSMC in the event of a Chinese invasion of Taiwan (36:40).
The host points out the severe implications of such rhetoric, noting, "He's not alone. Rapidly disassembling TSMC is an idea shared by others since at least 2022 in US military academia" (38:05), which Taiwan has consistently opposed.
The tension underscores the strategic importance of TSMC in the global tech supply chain and the potential ramifications of its destabilization on international relations and technology markets.
Concluding the episode is an intriguing tech innovation in cybersecurity:
Guillermo Rauch, CEO of Vercel, unveiled a new Doom-based CAPTCHA designed to enhance human verification by requiring users to achieve "three kills on nightmare difficulty" in the classic shooter game to prove their humanity (40:20).
This innovation builds upon earlier attempts, such as the 2021 Doom-inspired CAPTCHA by Mikel Camps Orteza, which relied on 2D Flash game mechanics but struggled to effectively block bots (42:00).
The host humorously acknowledges the challenge, "and now we have ChatGPT. Thanks a lot," reflecting on the evolving sophistication of AI and the necessity for more robust human verification methods (43:10).
This segment highlights the creative approaches emerging to counteract automated bot activities online, ensuring more secure and authentic user interactions.
Conclusion:
The January 4, 2025, episode of TechLinked delivers a comprehensive overview of the latest developments in tech and gaming culture. From high-profile CES leaks and groundbreaking hardware announcements to contentious AI policies and innovative cybersecurity measures, the host navigates a diverse array of topics with insightful commentary and engaging humor. Listeners gain a nuanced understanding of the current tech landscape, enriched by timely updates and critical perspectives on industry trends.