4:41
Ladies and gentlemen, in the red corner weighing in at several trillion dollars, the entire tech industry. And in the blue corner weighing in at 980 words, THE TECH NEWS. LET'S GET READY TO RUMBLE! THROUGH SOME HEADLINES. Valve says the Steam Machine and Steam Frame are now shipping sometime this summer, which means Valve hardware has advanced from early 2026 to uh-oh, ram prices to get YOUR SUNSCREEN ON, FOLKS. HA! THE NEW launch window showed up in a Steamworks blog post about expanding the Steam Verified program beyond the Steam Deck, because apparently Valve announces hardware updates like it's hiding a side quest. There's some detail in there.
What could it be? But this is still the clearest timing we've received on the Steam Machine's launch yet, after February's component-cost-driven delay into a liminal space between early 2026 and sometime this year. I think it hit the backroom. Steam Machine Verified will mark games that should run well on the Gabe Cube, while Steam Frame standalone verified covers games that run natively on the headset. So, developers are clearly being told to get ready. Valve still hasn't announced pricing, which matters a little bit when the delay was partly about the stuff inside the box, but at least Valve time now has a season. A time to buy a million games on a Steam sale that you'll never play. A time to die.
Anthropic has issued a plea for the AI industry to please slow down development of frontier models so safety research can catch up. We're all trying to find the guy who did this. In a blog post released yesterday, the company warns that without a brake pedal, society is heading for a Skynet scenario where humanity loses control of its own creations. We're not going to stop making it though." This plea was part of a larger announcement that Anthropic researchers have observed their models showing signs of recursive self-improvement where AI can autonomously design and train its own successors. Anthropic says it plans to convene policy makers and rival AI labs in the coming months to develop an approach to safety modeled on nuclear
arms treaties. Critics, however, are skeptical given that Anthropic just filed for an IPO. We've got to do this the safest way possible. Get those shareholders in here. And then there's the fact that CEO Dario Amodei has a history of being a bit of a hype beast about his own products. Like in April when he refused to release Claude 2 publicly because it was so dangerous, only giving it to a special council of nerds he called the Glasswing. Then a couple of months later, gave it to 150 more companies. What could have possibly changed in the last 2 months, Dario? Was it maybe a trillion-dollar valuation? Nah. Meta is under fire after Wired reporters found facial recognition software buried in the companion app for Meta's smart glasses. Now the code,
which was confirmed to be dormant, it's not active, is part of an unreleased feature called Nametag that identifies faces captured through the glasses camera and notifies the wearer when there's a match. We reported on this back in February when a leaked memo revealed Meta plan to use the dynamic political environment as cover for the release of the feature. That creeped everyone out and led over 70 organizations to sign a joint statement warning the feature could be used by stalkers and caused a number of senators to ask Meta to defend its position. Meta responded saying that they don't offer that feature, but if they were to release something like it, they'd take a thoughtful approach. We'd be super careful, guys. I was like, trust me, reptile man. So, I
guess they thoughtfully decided to move forward with this feature that allows Ray-Ban Meta wearers, who are already non-consensually filming people, to also non-consensually collect biometric data. Wow, that's really thoughtful. Too thoughtful, maybe. Maybe stop thinking about me so much, Meta, and instead think about our sponsor, TryHackMe. Basically, the easiest way to learn cybersecurity. Everything runs right in your browser, so no complicated setup, no expensive hardware, and you can jump straight into hands-on labs. Like well, it's like they was jumping in. Um, there's a reason more than 7 million people are learning on the platform. Another reason it's special,
it guides you with structured learning paths, so you always know what to learn next instead of looking up tutorials at 2:00 a.m. This is you, oh, what do I know? Just like the name suggests, you'll be telling your friends, "TryHackMe, bro." But, they won't be able to. There are over 1,100 rooms to learn in, and new challenges every week. In the offensive security room, you'll start things off by hacking into a fake web application. Like you're Hugh Jackman in Swordfish, the coolest movie ever. But, if you don't want to be some sort of hacker on the attack, TryHackMe can also teach you about defensive cybersecurity with labs based on real attacks and defenses used in the industry. That's only a few of them.
Speaking of the industry, the certifications and careers hub prepares you for interviews in cybersecurity jobs, if you want those. And now, you're River Phoenix in Sneakers. Look, I'm trying to free your mind, but I can only show you the door. So, check out TryHackMe at the link in the description and sign up for free. Oh, you can also upgrade to premium for 25% off using our link, at checkout. And then, you'll be in. And here come the quick bits off the top rope. ALL FIVE AT ONCE. OH, NO. OH, GOD. This was a huge mistake. Nvidia's rumored RTX 50 Super cards are reportedly back on the menu, contrary to what previous leaks might have made you think. The refresh was originally expected to bring bigger VRAM pools to the cards, but then it got delayed when
the chips required for that became hard to find. Now, leaker Megasize GPU says the 5070 and 5080 Super are alive again, with a possible 12 GB RTX 5060 joining in. Nvidia hasn't confirmed anything, so for now, the Super lineup exists in ray-traced Schrödinger mode. Leaked Microsoft documents obtained by 404 Media reportedly said phase one for the company's new Scout AI assistant was to make people addicted, which is a bold thing to write down when make Clippy sexier for nerds was right there. Put Clippy in a bikini. What? Are [clears throat] we Scout is an always-on Microsoft 365 agent based on Open Claw, the viral agent tool CEO Satya Nadella once said Microsoft couldn't ship because it would
look like they were launching a virus. He later denied that addiction was the goal, so apparently, Microsoft just took its previously announced AI assistant Miko that everyone forgot about, gave it claws, a new logo, and a freaking lobster tank for no reason at all. We just wanted to do that. Just felt like it that day. YouTube Premium's price increases announced back in April come into effect this month, and users are still pissed. I thought they would have gotten over it by now. The hike has many subscribers considering whether the service is worth it, with several speculating that the advent of 30-second unskippable ads are a tactic to force people to upgrade. One tiny sliver of a
silver lining, however, other than the fact that YouTube Premium is pretty good, I have it, is that the increases do come with some improved features, quote-unquote improved, including auto speed playback and a new podcast finding tool. It's so hard to find pod Where are all the podcasts? So, while you will be paying more, you could make that money back by listening to those boring finance bro podcasts way faster. Just listen at three times speed, you're still absorbing it, and you don't have to actually experience it. Bots now generate more web traffic than humans, according to Cloudflare, which means the dead internet theory just walked into the room wearing our skin. Feels like home. Cloudflare's radar dashboard shows
bots making the majority of HTTP requests, with CEO Matthew Prince saying agentic traffic crossed the line faster than expected. So, video killed the radio star, the internet killed the video star, and now the AI agents have turned the web into a cyber wasteland. Didn't see that coming. And a San Francisco burglar reportedly used Waymo as a getaway car after stealing activewear from a yoga studio, making the future of crime breathable fabric and five-star passenger ratings. It's the most San Francisco crime ever. Police say the suspect escaped in January and still hasn't been caught, even though Waymo's taxis have cameras, accounts, and payment info. The interior footage was already gone by the time police got a warrant. So,
congratulations to Waymo for inventing Delamain, the self-aware taxi from Cyberpunk 2077, featuring plausible deniability. Why didn't they interrogate the Waymo? Get him into that dark room with the light. What do you know? THE TECH INDUSTRY HAS BEEN declared the winner by technical knockout and a wild amount of government lobbying. A rematch has been called for Monday. Let's get the tech news patched up and ice those stories. So, that was Yikes.
