WEBVTT 00:00:00.000 --> 00:00:05.600 The iPhone 15 Pro and Pro Max, I've had a heck of a launch week with users and reviewers alike 00:00:05.600 --> 00:00:10.080 testing the phone's new titanium construction. And the scores are not looking great, they really 00:00:10.080 --> 00:00:13.520 should have paid attention in class. I'm James Shrive, this is TechLinked, and this is the 00:00:13.520 --> 00:00:21.040 Repairability Score awarded to the iPhone 15 lineup by iFixit. 4 out of 10, matching the iPhone 14 00:00:21.040 --> 00:00:27.200 score after iFixit downgraded it from a 7. You see, the 14 added the ability to open up from 00:00:27.200 --> 00:00:33.680 the front or the back, but Apple still heavily limits user repair through software. Kind of like 00:00:33.680 --> 00:00:37.600 inviting your friend over to stand in your entryway, but proceed no further. 00:00:37.600 --> 00:00:45.600 And value your friendship this much. I get cookies. Similar to many early iPhone 15 Pro owners, iFixit found the titanium shell 00:00:45.600 --> 00:00:50.080 scratches really easily, something also discovered by YouTuber JerryRig. Everything 00:00:50.080 --> 00:00:56.160 shortly before he cracked the glass back of the 15 Pro Max about two seconds into his bend test. 00:00:56.160 --> 00:00:59.520 It might be time for iPhoneers to consider switching to the 00:00:59.520 --> 00:01:04.960 Pixel 8 Pro when it launches on October 4th. Google is now promising seven years of software 00:01:04.960 --> 00:01:09.360 updates which should really extend the life of your phone as long as it doesn't pull a Pixel 7 00:01:09.360 --> 00:01:13.280 and have its buttons randomly fall off. I mean, you barely use them. 00:01:13.280 --> 00:01:18.640 Redditors can now earn real money based on the amount of karma and gold they receive on eligible 00:01:18.640 --> 00:01:23.760 posts, validating the oceans of blood, sweat, and tears excreted by millions of users who 00:01:23.760 --> 00:01:27.120 have spent years toiling in the fake internet point fields. 00:01:28.720 --> 00:01:33.360 It's hard not to draw comparisons between the new Reddit contributor program and Twitter's 00:01:33.360 --> 00:01:39.200 recent rollout of ad revenue sharing for Twitter Blue. I mean, X-premium subscribers. 00:01:39.200 --> 00:01:43.840 If only because social media platforms paying users living wages for a shitpost written 00:01:43.840 --> 00:01:47.280 while shitting on the literal shitter is a fairly new thing. 00:01:47.280 --> 00:01:51.440 Unlike BlueShack, Redditors won't have to front eight bucks a month to get paid, 00:01:51.520 --> 00:01:55.360 but they will have to verify that they meet certain eligibility requirements 00:01:55.360 --> 00:02:00.960 like being over 18, living in the US, and having an account in good standing. 00:02:00.960 --> 00:02:05.920 They can start farming dollar-dollar bills on eligible contributions which don't include posts 00:02:05.920 --> 00:02:10.720 with sexually explicit content or content related to violence, drugs, farms, gambling, 00:02:10.720 --> 00:02:14.480 scams, crypto or otherwise, and of course, donations to chair. 00:02:16.640 --> 00:02:20.960 References to the new program were actually first spotted by Android Authority back in July 00:02:20.960 --> 00:02:26.160 when Reddit announced a phase out of Reddit coins and gold shortly before launching more 00:02:26.160 --> 00:02:30.960 NFT avatars which have somehow survived the crypto collapse. Yes, they're still cringe, 00:02:30.960 --> 00:02:35.520 but it's Reddit, so it doesn't matter. NVIDIA envisions a future where games are mostly rendered 00:02:35.520 --> 00:02:40.480 from AI-generated frames as opposed to frames produced with traditional rasterization at 00:02:40.480 --> 00:02:45.440 native resolution because they'd look better anyway. Like how any picture pre-2022 pales in 00:02:45.440 --> 00:02:49.760 comparison to the anime wife whose mid-journey creates for me to pretend to cuddle with. 00:02:53.040 --> 00:02:57.440 The AI-rendered games comment comes from NVIDIA's vice president of applied deep learning research, 00:02:57.440 --> 00:03:03.680 Brian Catanzaro, who told Digital Foundry that in Cyberpunk 2077's latest update, 00:03:03.680 --> 00:03:08.560 DLSS 3.5's frame generation and ray reconstruction features result in 00:03:09.120 --> 00:03:14.000 real-er, better-looking frames, despite being the result of an upscaling process. 00:03:14.080 --> 00:03:20.720 Brian went so far as to suggest that in the far future, DLSS 10 will enable a completely 00:03:20.720 --> 00:03:25.920 neural rendering system which won't use rasterization at all. Whether he made the comments 00:03:25.920 --> 00:03:33.840 as a cope for the fact that AMD's Radeon RX 7900 XT can power the 7680x2160 240Hz display on 00:03:33.840 --> 00:03:40.240 Samsung's latest super-ultra-wide Odyssey Neo G9 while the RTX 4090 cannot is anyone's guess. 00:03:40.320 --> 00:03:43.920 Sadly, the quick bits got a bit squished in transit, so they're really more like 00:03:45.200 --> 00:03:51.200 no, still good, but just smooshed. A ransomware group is claiming that all of Sony's bases are 00:03:51.200 --> 00:03:57.840 belonged to them as they have breached quote all Sony systems and quote everyone. The proof of 00:03:57.840 --> 00:04:02.800 the hack posted by the group is apparently not particularly compelling according to cybersecurity 00:04:02.800 --> 00:04:07.280 experts, but we'll know for sure sooner or later as the group says that Sony won't pay and the data 00:04:07.280 --> 00:04:12.240 will be published in three days if nobody buys it first. Though if this is anything like the 00:04:12.240 --> 00:04:17.360 2014 Sony hack, we might get a sneak peek into future films and terrorist threats targeting 00:04:17.360 --> 00:04:23.440 Seth Rogen. Remember that? Wait, how do we know this wasn't him? Kim's back. OpenAI has released 00:04:23.440 --> 00:04:29.440 an update for chatGPT that allows the chatbot to see, hear, and speak, removing one of the last 00:04:29.440 --> 00:04:33.920 things really throwing a wrench in our romantic relationship. It already has that sweet nothings 00:04:33.920 --> 00:04:39.440 extension. You can now audibly talk to it in a conversational style as well as include images 00:04:39.440 --> 00:04:43.200 in your prompts so it knows what the hell you're talking about if your parents told you to use 00:04:43.200 --> 00:04:48.800 your words a bit too much. Now you can show chatGPT a picture of your barbecue to troubleshoot why it 00:04:48.800 --> 00:04:54.560 won't work if you'd like to hear that it's infested by gremlins. The update gives chatGPT 00:04:54.560 --> 00:04:58.960 the tools it will need to go against Meta's planned sassy chatbot for teens, but it's 00:04:58.960 --> 00:05:06.000 left me missing our old much worse AI assistants like ALXE-A. ChatGPT, play despacito. Getty images 00:05:06.000 --> 00:05:10.800 who has previously sued AI art generators for using its copyrighted images in their training 00:05:10.800 --> 00:05:16.320 data has now announced their own AI art generator because this is the way of things built by NVIDIA 00:05:16.320 --> 00:05:21.600 and reportedly trained entirely on licensed images. A claim Getty is so confident in, 00:05:21.600 --> 00:05:25.920 they're offering to get sued on your behalf. It's like the 30 minutes are free guarantee, 00:05:25.920 --> 00:05:31.120 but if the driver was explicitly willing to go to jail to get your pizza to you on time, 00:05:31.120 --> 00:05:36.480 I don't care for the school zone. Getty also says that images generated this 00:05:36.480 --> 00:05:41.360 way will not be added into the library, thereby preventing the lossy JPEG problem of training 00:05:41.360 --> 00:05:46.640 in AI using their own output that threatens to turn all future AI art into an uncanny 00:05:46.640 --> 00:05:53.440 valley soup or hellscape. Head of AMD's Radeon Division, Scott Hurtlman, 00:05:53.520 --> 00:05:57.760 has announced his departure from the company after seven years following the launch of the RX 00:05:57.760 --> 00:06:03.680 7800 XT and 7700 XT. Hurtlman was known as a friendly and approachable face willing to engage 00:06:03.680 --> 00:06:09.040 with the community. Via his socials, Hurtlman said goodbye to the Radeon team, saying he will 00:06:09.040 --> 00:06:18.400 miss every single one of them. And Tesla has released a new video ostensibly 00:06:18.400 --> 00:06:24.640 showcasing the new capabilities of their Neuronet-trained Optimus robot, first unveiled two 00:06:24.640 --> 00:06:29.680 years ago. Though I don't see it taking on the Decepticons anytime soon, some have speculated 00:06:29.680 --> 00:06:34.160 that parts of the video look like CGI, but it's equally possible that they simply fast forwarded 00:06:34.160 --> 00:06:38.640 sections to make the robot look less like a concussed toddler slowly sorting blocks, 00:06:38.640 --> 00:06:42.480 which would explain why its human assistant appears to move uncannily fast. 00:06:43.440 --> 00:06:48.560 Hori's got a lot of coffee. In another scene, the same robot also demonstrated its ability to 00:06:48.560 --> 00:06:53.760 balance on one leg through some beginner yoga, hopefully labeled Namaste. 00:06:56.560 --> 00:07:00.000 And you should Namaste tuned for Wednesday when we'll be back!