正在加载视频...

视频加载失败

返回首页

After optimizing 2000+ PCs, here's the only recording setup I recommend: OBS Replay Buffer with launch parameters: --startreplaybuffer --minimize-to-tray --disable-updater NVIDIA App: Inconsistent, performance killing, telemetry forced on OBS: Professional results, very minimal performance impact, no telemetry Stop settling for bloatware

x1lly's profile picture

Xilly

24,074 subscribers

23,346 次观看 • 1 年前 •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Silver Palace Performance Results After 48 Hours of Testing !! 1. The game currently runs worse than Wuthering Waves, with frequent frame drops. 2. The developers have released 3 updates so far, and I noticed that the CPU bottleneck has been reduced. 3. These results are on DLSS Quality. DLAA has even more stuttering. Question: Is it fixable? Answer: It depends. No matter what, this game is going to be very demanding. If you look at NTE, it uses a hybrid graphics system where you can disable SSR through presets, but Silver Palace doesn't have an option like that. Combined with other heavy graphical effects, performance takes a noticeable hit. This hybrid system should have been available from the start. Since we're only in CBT2, I expect some improvements before launch. Based on my testing, the game will likely perform similarly to STALKER 2, and performance will also depend heavily on Frame Generation. I also noticed CPU bottlenecks on both DLAA and DLSS in almost every area. Things that need to be added: Render Distance setting — this would significantly reduce CPU bottlenecks for mid-range and low-end CPUs. #SilverPalace #Gacha
1:09

Silver Palace Performance Results After 48 Hours of Testing !! 1. The game currently runs worse than Wuthering Waves, with frequent frame drops. 2. The developers have released 3 updates so far, and I noticed that the CPU bottleneck has been reduced. 3. These results are on DLSS Quality. DLAA has even more stuttering. Question: Is it fixable? Answer: It depends. No matter what, this game is going to be very demanding. If you look at NTE, it uses a hybrid graphics system where you can disable SSR through presets, but Silver Palace doesn't have an option like that. Combined with other heavy graphical effects, performance takes a noticeable hit. This hybrid system should have been available from the start. Since we're only in CBT2, I expect some improvements before launch. Based on my testing, the game will likely perform similarly to STALKER 2, and performance will also depend heavily on Frame Generation. I also noticed CPU bottlenecks on both DLAA and DLSS in almost every area. Things that need to be added: Render Distance setting — this would significantly reduce CPU bottlenecks for mid-range and low-end CPUs. #SilverPalace #Gacha

Kaito

134,948 次观看 • 2 个月前

The Astra videos are fun. But I am trying a much cheaper setup with my laptop, around 11GB of open models and a $1.6k Unitree Go2 Air After one walk I can search what the dog saw and where it saw it. Next I want it to recognize those places when we go back and use that memory This is the Air, not the EDU model with official support for secondary development. It was not really built for this, which makes it more fun I drove it around my apartment (sorry for the mess) and (Dimensional) recorded frames along with LiDAR, odometry and transforms. Having the images and the robot's position in one recording gave me something to build on On the right of the video, the dog is map filling in with what it saw where. The left one shows what it detects Three very small and open models ran later on my laptop: SigLIP around 200M parameters, made the frames searchable. Qwen3-VL-8B in 4-bit through ollama described the selected views and checked search result Moondream2, from moondream supplied the object boxes All zero shot. We can upload the recordings and saved maps to the Dimensional cloud and download them later. The goal is for the Go2 to know where he is and where is what Next step is running this minimal stack on a 16 GB NVIDIA Embedded Jetson mounted on the Go2 Air
0:52

The Astra videos are fun. But I am trying a much cheaper setup with my laptop, around 11GB of open models and a $1.6k Unitree Go2 Air After one walk I can search what the dog saw and where it saw it. Next I want it to recognize those places when we go back and use that memory This is the Air, not the EDU model with official support for secondary development. It was not really built for this, which makes it more fun I drove it around my apartment (sorry for the mess) and (Dimensional) recorded frames along with LiDAR, odometry and transforms. Having the images and the robot's position in one recording gave me something to build on On the right of the video, the dog is map filling in with what it saw where. The left one shows what it detects Three very small and open models ran later on my laptop: SigLIP around 200M parameters, made the frames searchable. Qwen3-VL-8B in 4-bit through ollama described the selected views and checked search result Moondream2, from moondream supplied the object boxes All zero shot. We can upload the recordings and saved maps to the Dimensional cloud and download them later. The goal is for the Go2 to know where he is and where is what Next step is running this minimal stack on a 16 GB NVIDIA Embedded Jetson mounted on the Go2 Air

Omar Espejel

77,940 次观看 • 14 天前

Few people care more about software performance than Casey Muratori. Give him a few minutes of your attention with this episode, and he'll convince you to learn to read Assembly (no, really, I finally started to read it, it's really not that scary, esp with an AI that can help explain the sequences). Timestamps: 00:00 Intro 05:17 Games at Microsoft 12:52 Building games 16:00 Why performance matters 27:12 Why you should learn to read assembly 30:36 Designing for optimization 42:51 How to get better at writing performant software 49:04 Understanding how the CPU works 55:53 Building games then and now 1:05:56 How game engines changed building games 1:10:48 Why new games compete with old games 1:13:25 GTA 6: why is it taking so long? 1:16:59 Casey's critique of clean code 1:21:48 Casey's take on TDD 1:24:30 What is good code? 1:27:32 What makes a good software engineer? 1:33:56 Why Casey doesn't code with AI 1:39:01 AI's impact on the game industry 1:44:43 AI and burnout 1:50:21 Why you should read papers Brought to you by: • Antithesis – verify your system’s correctness without human review or traditional integration tests – and avoid bugs or outages • Sentry – application monitoring software considered “not bad” by millions of developers • turbopuffer – a vector and full-text search engine built on object storage. It’s fast, cheap, and extremely scalable Three interesting things we talked about: 1. Is performance starting to matter to businesses? Enterprise software buyers care mainly about cost, compliance, and capabilities – but not performance. Even so, there are some products gaining major popularity and market share due to their performance, such as File Pilot (next-gen file explorer) and the Blick video editor. Is the tide turning? 2. Profiler-driven performance optimization is the wrong way to optimize The standard way of optimizing is to profile the application, tweak hotspots, then check if the stats have improved. But this only finds a local minimum; Casey says every engineer he’s worked with who was a great “optimizer” began by establishing what the hardware could theoretically do, and then did not stop until they’d closed the gap to that performance level. 3. Take a grain of salt with conventional wisdom that premature optimization is the “root of all evil” Many devs use it as an excuse to delay performance optimization, but Casey says that not optimizing in time could mean that only performance hotspots can be fixed later, and not the architectural issues that create poor performance. Architect your system to be performant, or you’ll have trouble solving problems without a rewrite!
1:53:58

Few people care more about software performance than Casey Muratori. Give him a few minutes of your attention with this episode, and he'll convince you to learn to read Assembly (no, really, I finally started to read it, it's really not that scary, esp with an AI that can help explain the sequences). Timestamps: 00:00 Intro 05:17 Games at Microsoft 12:52 Building games 16:00 Why performance matters 27:12 Why you should learn to read assembly 30:36 Designing for optimization 42:51 How to get better at writing performant software 49:04 Understanding how the CPU works 55:53 Building games then and now 1:05:56 How game engines changed building games 1:10:48 Why new games compete with old games 1:13:25 GTA 6: why is it taking so long? 1:16:59 Casey's critique of clean code 1:21:48 Casey's take on TDD 1:24:30 What is good code? 1:27:32 What makes a good software engineer? 1:33:56 Why Casey doesn't code with AI 1:39:01 AI's impact on the game industry 1:44:43 AI and burnout 1:50:21 Why you should read papers Brought to you by: • Antithesis – verify your system’s correctness without human review or traditional integration tests – and avoid bugs or outages • Sentry – application monitoring software considered “not bad” by millions of developers • turbopuffer – a vector and full-text search engine built on object storage. It’s fast, cheap, and extremely scalable Three interesting things we talked about: 1. Is performance starting to matter to businesses? Enterprise software buyers care mainly about cost, compliance, and capabilities – but not performance. Even so, there are some products gaining major popularity and market share due to their performance, such as File Pilot (next-gen file explorer) and the Blick video editor. Is the tide turning? 2. Profiler-driven performance optimization is the wrong way to optimize The standard way of optimizing is to profile the application, tweak hotspots, then check if the stats have improved. But this only finds a local minimum; Casey says every engineer he’s worked with who was a great “optimizer” began by establishing what the hardware could theoretically do, and then did not stop until they’d closed the gap to that performance level. 3. Take a grain of salt with conventional wisdom that premature optimization is the “root of all evil” Many devs use it as an excuse to delay performance optimization, but Casey says that not optimizing in time could mean that only performance hotspots can be fixed later, and not the architectural issues that create poor performance. Architect your system to be performant, or you’ll have trouble solving problems without a rewrite!

Gergely Orosz

97,977 次观看 • 28 天前

Agent OBS: First Day Live. We armed Agent OBS for real trading on Robinhood Chain, with $100 per trade, a 10% floor, and a cap of 6 entries and $1,200 of swaps a day. This is what the first 11 hours looked like. Numbers: - 85 desk cycles: 61 holds, 17 entries refused by the rails, 6 buys. - 6 trades, all $100: 4 wins, 2 losses. SHARD +$20 (take profit at +20% after one minute) TRIBUTE +$43 (sold into thinning buyers after 44 minutes) COMMOTITTES +$6, ZZZ +$2 FRANKLIN -$48 (a fresh launch that fell 48% in 48 seconds; the floor sold it) MEME -$10 (we closed it by hand when we narrowed the lane, see below) - Gross on the marks: +$13. After price impact, gas and two USDG hops: the wallet went from 0.4100 to 0.4057 ETH, about -1%. Dollar equity was flat, $1,018 to $1,014, with ETH up 0.7% on the day. What We Learned: - The rails did their job. Every exit fired on its rule: take profit, floor, tape roll-over. The agent itself proposed only sells it could justify. - Fresh launches are where the loss came from. After FRANKLIN we narrowed the lane to tokens at least 24 hours old with a trading record, and the four trades after that were +$41 net. - At $100 a trade, round-trip cost of roughly 4% eats the small wins. The next decision is size and selection, not more rules. What Broke And Was Fixed The Same Night: - A sell that would have reverted on every full exit (the book carried a hair more than the wallet held). Caught before it cost anything. - Dust after a full sell counting as a position, and airdropped tokens counting as holdings. The desk now only ever touches what it bought. - The terminal describing a held token as a new entry, and the performance chart dropping off a cliff for a minute after each sell. Both fixed. Next: Let the 24-hour-old lane run a full day at the same size, then decide on size and on whether fresh launches get a smaller lane of their own.
0:50

Agent OBS: First Day Live. We armed Agent OBS for real trading on Robinhood Chain, with $100 per trade, a 10% floor, and a cap of 6 entries and $1,200 of swaps a day. This is what the first 11 hours looked like. Numbers: - 85 desk cycles: 61 holds, 17 entries refused by the rails, 6 buys. - 6 trades, all $100: 4 wins, 2 losses. SHARD +$20 (take profit at +20% after one minute) TRIBUTE +$43 (sold into thinning buyers after 44 minutes) COMMOTITTES +$6, ZZZ +$2 FRANKLIN -$48 (a fresh launch that fell 48% in 48 seconds; the floor sold it) MEME -$10 (we closed it by hand when we narrowed the lane, see below) - Gross on the marks: +$13. After price impact, gas and two USDG hops: the wallet went from 0.4100 to 0.4057 ETH, about -1%. Dollar equity was flat, $1,018 to $1,014, with ETH up 0.7% on the day. What We Learned: - The rails did their job. Every exit fired on its rule: take profit, floor, tape roll-over. The agent itself proposed only sells it could justify. - Fresh launches are where the loss came from. After FRANKLIN we narrowed the lane to tokens at least 24 hours old with a trading record, and the four trades after that were +$41 net. - At $100 a trade, round-trip cost of roughly 4% eats the small wins. The next decision is size and selection, not more rules. What Broke And Was Fixed The Same Night: - A sell that would have reverted on every full exit (the book carried a hair more than the wallet held). Caught before it cost anything. - Dust after a full sell counting as a position, and airdropped tokens counting as holdings. The desk now only ever touches what it bought. - The terminal describing a held token as a new entry, and the performance chart dropping off a cliff for a minute after each sell. Both fixed. Next: Let the 24-hour-old lane run a full day at the same size, then decide on size and on whether fresh launches get a smaller lane of their own.

Obscura

12,638 次观看 • 17 天前

Here’s my written & video review of the new 2026 Tesla Model Y Performance after driving it for a week. This is the best-value new Tesla you can buy. Crazy performance, no real drawbacks, and all for just $57,490. Let’s dive in. Price: I haven’t seen many others mention this: every option on the Model Y Performance is included at no extra cost in the US (except FSD). So if you spec a Model Y Premium AWD with an upgraded paint color, tow package, white interior, and upgraded 20" wheels, a fully loaded Model Y Premium AWD ends up only about $2,500 less expensive than a fully loaded Model Y Performance. Ride Quality: I thought I might feel worse ride quality vs my Premium AWD Model Y with 20" wheels, but I struggled to find any real difference, despite the larger 21" wheels, firmer suspension setting, and 0.6" lower ride height on the Performance trim. A true testament to Tesla's engineering magic on this thing. Exterior Design: Unlike the previous Model Y Performance, Tesla made some exterior design tweaks with a new front and rear fascia to spice things up a bit. The result, in my opinion, is the best-looking Model Y trim you can buy. It definitely has a more aggressive presence in person, even if it's subtle. The carbon-fiber spoiler boosts high-speed stability and cuts aerodynamic drag by 10%. The new 21’’ Arachnid 2.0 wheels look fantastic in person, one of my favorite designs ever from Tesla. Staggered wheel and tire fitment provides better grip and steering. The beefier 275mm rear tires (255mm in the front) give the vehicle a better stance from behind. Vehicle-to-Load (V2L): For the first time on a Model Y in North America, you can now plug in anything you want to the exterior charge port with an adapter, even a campsite! It provides up to 2.4 kW of power (120V at 20A) from two household outlets. It's a great feature. Interior: It’s what you know and love, but with a few changes that elevate the ownership experience. New with the Performance is a larger 16" center screen (vs. 15.4" on non-Performance models), with thinner bezels and higher resolution. It’s not a huge difference on paper, but you definitely notice it in daily use. The carbon-fiber décor on the door cards and dash is a nice touch, though I would like to see it extended to the center console. The new performance seats are the best seats of any Tesla I've ever experienced. While they retain aggressive bolstering in the torso area, the bottom seat cushion has less aggressive bolstering than on the Model 3 Performance seats, making it easier to get in and out. It also doesn’t squeeze your thighs too tightly. The powered thigh extenders add comfort on longer drives, especially for taller people who want extra support. And of course, they’re heated and ventilated. The headrests also feel more comfortable than the ones in my Model Y. I want these seats. Unlike the old Model Y Performance, the new one has no Track Mode. Why? Because nobody used it lol. No point in putting engineering resources into something people won’t use. The refreshed Model 3 Performance still has it, though. Cabin Quietness: Despite the thinner-profile tires, there is no noticeable difference vs my Premium Model Y. Decibel reading results at highway speeds were visually the same compared to my 2026 Model Y Premium (65-66). Driving Impressions: It’s amazing. Sharp, precise, and agile. The vehicle feels stable at all times. Acceleration is blistering (3.3s 0–60 mph), with plenty of punch even at higher speeds. The tires offer good grip, and cornering is fantastic for an SUV. The suspension setup is great. More steering wheel feedback would be nice, though. Cruising around traffic is a joy. The brakes are much improved over the previous Model Y Performance and are far better suited for spirited driving. There are three acceleration modes: Chill, Standard, and Insane. Just stay in Insane. You’d be insane not to lol. The car also lets you switch between two ride and handling modes: Standard and Sport. The difference isn’t huge, but Standard is better if you’ve got passengers. FSD: I unfortunately wasn’t able to get FSD V14 on this car. It had V13.2.9, so I didn’t use it much. But in the little time I did, it was smooth and comfortable. It didn’t bother me because V14 will perform just as well here as it does on my 2026 Model Y Premium AWD. Conclusion: You won’t find another new SUV today that offers this level of performance for the price. Back in 2022, when Tesla couldn’t build Model Ys fast enough, a fully loaded Model Y Performance cost over $90,440. Today, the refreshed and far more capable 2026 Model Y Performance is just $57,490 fully loaded, and you can simply subscribe to FSD for $99/month. The 2026 Model Y Performance delivers utility, great performance, comfort, tech, self-driving and everything else people love about the Model Y. It’s a no-brainer purchase. I want one badly, but I'll need to show restraint, as I’m saving up for a house lol.
15:48

Here’s my written & video review of the new 2026 Tesla Model Y Performance after driving it for a week. This is the best-value new Tesla you can buy. Crazy performance, no real drawbacks, and all for just $57,490. Let’s dive in. Price: I haven’t seen many others mention this: every option on the Model Y Performance is included at no extra cost in the US (except FSD). So if you spec a Model Y Premium AWD with an upgraded paint color, tow package, white interior, and upgraded 20" wheels, a fully loaded Model Y Premium AWD ends up only about $2,500 less expensive than a fully loaded Model Y Performance. Ride Quality: I thought I might feel worse ride quality vs my Premium AWD Model Y with 20" wheels, but I struggled to find any real difference, despite the larger 21" wheels, firmer suspension setting, and 0.6" lower ride height on the Performance trim. A true testament to Tesla's engineering magic on this thing. Exterior Design: Unlike the previous Model Y Performance, Tesla made some exterior design tweaks with a new front and rear fascia to spice things up a bit. The result, in my opinion, is the best-looking Model Y trim you can buy. It definitely has a more aggressive presence in person, even if it's subtle. The carbon-fiber spoiler boosts high-speed stability and cuts aerodynamic drag by 10%. The new 21’’ Arachnid 2.0 wheels look fantastic in person, one of my favorite designs ever from Tesla. Staggered wheel and tire fitment provides better grip and steering. The beefier 275mm rear tires (255mm in the front) give the vehicle a better stance from behind. Vehicle-to-Load (V2L): For the first time on a Model Y in North America, you can now plug in anything you want to the exterior charge port with an adapter, even a campsite! It provides up to 2.4 kW of power (120V at 20A) from two household outlets. It's a great feature. Interior: It’s what you know and love, but with a few changes that elevate the ownership experience. New with the Performance is a larger 16" center screen (vs. 15.4" on non-Performance models), with thinner bezels and higher resolution. It’s not a huge difference on paper, but you definitely notice it in daily use. The carbon-fiber décor on the door cards and dash is a nice touch, though I would like to see it extended to the center console. The new performance seats are the best seats of any Tesla I've ever experienced. While they retain aggressive bolstering in the torso area, the bottom seat cushion has less aggressive bolstering than on the Model 3 Performance seats, making it easier to get in and out. It also doesn’t squeeze your thighs too tightly. The powered thigh extenders add comfort on longer drives, especially for taller people who want extra support. And of course, they’re heated and ventilated. The headrests also feel more comfortable than the ones in my Model Y. I want these seats. Unlike the old Model Y Performance, the new one has no Track Mode. Why? Because nobody used it lol. No point in putting engineering resources into something people won’t use. The refreshed Model 3 Performance still has it, though. Cabin Quietness: Despite the thinner-profile tires, there is no noticeable difference vs my Premium Model Y. Decibel reading results at highway speeds were visually the same compared to my 2026 Model Y Premium (65-66). Driving Impressions: It’s amazing. Sharp, precise, and agile. The vehicle feels stable at all times. Acceleration is blistering (3.3s 0–60 mph), with plenty of punch even at higher speeds. The tires offer good grip, and cornering is fantastic for an SUV. The suspension setup is great. More steering wheel feedback would be nice, though. Cruising around traffic is a joy. The brakes are much improved over the previous Model Y Performance and are far better suited for spirited driving. There are three acceleration modes: Chill, Standard, and Insane. Just stay in Insane. You’d be insane not to lol. The car also lets you switch between two ride and handling modes: Standard and Sport. The difference isn’t huge, but Standard is better if you’ve got passengers. FSD: I unfortunately wasn’t able to get FSD V14 on this car. It had V13.2.9, so I didn’t use it much. But in the little time I did, it was smooth and comfortable. It didn’t bother me because V14 will perform just as well here as it does on my 2026 Model Y Premium AWD. Conclusion: You won’t find another new SUV today that offers this level of performance for the price. Back in 2022, when Tesla couldn’t build Model Ys fast enough, a fully loaded Model Y Performance cost over $90,440. Today, the refreshed and far more capable 2026 Model Y Performance is just $57,490 fully loaded, and you can simply subscribe to FSD for $99/month. The 2026 Model Y Performance delivers utility, great performance, comfort, tech, self-driving and everything else people love about the Model Y. It’s a no-brainer purchase. I want one badly, but I'll need to show restraint, as I’m saving up for a house lol.

Sawyer Merritt

225,461 次观看 • 9 个月前

#SEONGHWA about today's stage at the 2025 AAA ⭐️: I felt pressure because I was performing in front of someone who used to be my idol (IU), but since I prepared so hard for it, I did have confidence. But after the performance, I monitored it and in the first line I did one little pitch mistake. At least that’s how it felt for me, so… 🐶: It didn't show~! ⭐️: Yes.. I only realized it after I finished bc I was so immersed in the performance… so yeah, that’s what happened. 🐶: no but that.. ⭐️: it was a bit.. 🐶: everything's okay, it wasn't noticeable! ⭐️: I was a bit embarrassed… I felt embarrassed. 🐶: no.. it's because the members, I'm like that too, but on stage… when I go in, from the very beginning, from the intro, San is like that as well.. everyone looks at me with this sharp, cold gaze. Maybe it's because we all go in with that hyped-up and intense mood, yeah… (Reading a comment) "I was so excited because of the performance that I didn’t notice at all." ⭐️: Mm well, some people might have noticed and some might not, but still… What should I say? 🐶: But your dance break today was really cool, hyung ⭐️: I did my best. I really did my best, but I’ll make up for what was lacking today and show an even better performance tomorrow. 🐶: And today Seonghwa hyung had the center part in the dance break right? Yes, that was cool. ⭐️: Once again, after a long time... I ate well! 🐶: Yes, Seonghwa hyung practiced a lot to perfect the dance break center part. ⭐️: For Man On Fire as well.. that last part 🐶: For this stage hyung got all the dance breaks ⭐️: Oh, that's right.. The Man On Fire dance break was a genre we were doing for the first time back then, so it was a bit difficult. But as we performed it in this concert (tour), I became more familiar with it. Because of that, I think I could do it with a bit more confidence this time. 🐶: Each member has a different vibe when they take the dance center. Doesn't Seonghwa hyung have his own unique vibe? Like, restrained but still leaving that ‘killing part’ kind of vibe? ⭐️: Honestly, I’m still lacking a lot. 🐶: Oh, come on.. each member has their own dance-center vibe, but Seonghwa hyung seems to blend them all together? ⭐️: Like a chameleon, doing all sorts of different things right? 🐶: Yes yes.. that's very charming ⭐️: I really like mimicking the teachers’ dance styles too, so that’s why. Anyway, I had no regrets on stage, I had lots of fun. Also, I was happy to be able to show such a cool stage in front of ATINYs. Receiving the awards was great and it also was really good to perform on such a big stage. Personally, performing in front of seniors I admire was very meaningful. It gave me a bit more motivation. 🐶: That's right, Seonghwa hyung really respects that senior after all. ⭐️: I was so nervous.. 🐶: That’s totally understandable, I’d feel the same way in that situation.
3:19

#SEONGHWA about today's stage at the 2025 AAA ⭐️: I felt pressure because I was performing in front of someone who used to be my idol (IU), but since I prepared so hard for it, I did have confidence. But after the performance, I monitored it and in the first line I did one little pitch mistake. At least that’s how it felt for me, so… 🐶: It didn't show~! ⭐️: Yes.. I only realized it after I finished bc I was so immersed in the performance… so yeah, that’s what happened. 🐶: no but that.. ⭐️: it was a bit.. 🐶: everything's okay, it wasn't noticeable! ⭐️: I was a bit embarrassed… I felt embarrassed. 🐶: no.. it's because the members, I'm like that too, but on stage… when I go in, from the very beginning, from the intro, San is like that as well.. everyone looks at me with this sharp, cold gaze. Maybe it's because we all go in with that hyped-up and intense mood, yeah… (Reading a comment) "I was so excited because of the performance that I didn’t notice at all." ⭐️: Mm well, some people might have noticed and some might not, but still… What should I say? 🐶: But your dance break today was really cool, hyung ⭐️: I did my best. I really did my best, but I’ll make up for what was lacking today and show an even better performance tomorrow. 🐶: And today Seonghwa hyung had the center part in the dance break right? Yes, that was cool. ⭐️: Once again, after a long time... I ate well! 🐶: Yes, Seonghwa hyung practiced a lot to perfect the dance break center part. ⭐️: For Man On Fire as well.. that last part 🐶: For this stage hyung got all the dance breaks ⭐️: Oh, that's right.. The Man On Fire dance break was a genre we were doing for the first time back then, so it was a bit difficult. But as we performed it in this concert (tour), I became more familiar with it. Because of that, I think I could do it with a bit more confidence this time. 🐶: Each member has a different vibe when they take the dance center. Doesn't Seonghwa hyung have his own unique vibe? Like, restrained but still leaving that ‘killing part’ kind of vibe? ⭐️: Honestly, I’m still lacking a lot. 🐶: Oh, come on.. each member has their own dance-center vibe, but Seonghwa hyung seems to blend them all together? ⭐️: Like a chameleon, doing all sorts of different things right? 🐶: Yes yes.. that's very charming ⭐️: I really like mimicking the teachers’ dance styles too, so that’s why. Anyway, I had no regrets on stage, I had lots of fun. Also, I was happy to be able to show such a cool stage in front of ATINYs. Receiving the awards was great and it also was really good to perform on such a big stage. Personally, performing in front of seniors I admire was very meaningful. It gave me a bit more motivation. 🐶: That's right, Seonghwa hyung really respects that senior after all. ⭐️: I was so nervous.. 🐶: That’s totally understandable, I’d feel the same way in that situation.

Everything Seonghwa

61,750 次观看 • 9 个月前

Jensen Huang just blamed the US government for handing China its AI industry. He called Washington's entire chip export strategy "lunacy." The policy was simple: Block Nvidia chips from China, starve their AI development, win the race. But what actually happened: Cutting China off forced their entire AI ecosystem to stop building on Nvidia and optimize for domestic hardware instead. The very policy designed to kneecap them ACCELERATED their independence. Huawei just had the largest single year in company history. Chinese chip companies are going public. SMIC is scaling. Jensen's line: "We caused it." And while Washington celebrates the export controls, Jensen is watching 40% of the world's AI developers - most of them Chinese - slowly stop building on the American tech stack. That's the real national security threat nobody's talking about. But the China argument is just one piece of what Jensen revealed... He also dismantled the entire "escape Nvidia" story in two sentences: "Without Anthropic, why would there be any TPU growth at all? It's 100% Anthropic." Google's TPUs. Amazon's Trainium. Every ASIC challenging Nvidia. All of it traces back to ONE customer and funding relationship. Strip Anthropic out and the chip war narrative Wall Street has been pricing in for two years collapses overnight. And Anthropic only ended up on competing chips because Jensen missed his window. When they were raising early, they needed billions - not VC money. Nobody writes a $10 billion check into a lab with no product. Google wrote it. AWS wrote it. Both got compute commitments in return. But Jensen didn't move. His words: "That was my miss." The CEO of the most valuable company on Earth accidentally created his own competition by not writing a check. He's since invested in Anthropic directly. He now owns a piece of the only customer keeping the competition story alive. Then comes the economics that should end the debate entirely: Nvidia gross margins: ~70%. ASIC margins from Broadcom: ~65%. Jensen on stage: "What are you really saving?" Companies burned billions on custom silicon, wrecked their roadmaps, took on massive execution risk - to save 5 margin points on compute they STILL can't benchmark publicly against Nvidia. Jensen's open invitation: "I welcome Trainium to demonstrate their 40% cost advantage. Come to InferenceMAX." So here's the full picture Jensen painted this week: - The export controls created the competitor everyone feared - The chip war is one customer deep - The CEO who missed that customer now owns a piece of them - The companies spending billions to escape Nvidia can't justify the math in public There's no uprising. There's no escape. Just a lot of very expensive stories with nowhere to go. What do you think?
2:29

Jensen Huang just blamed the US government for handing China its AI industry. He called Washington's entire chip export strategy "lunacy." The policy was simple: Block Nvidia chips from China, starve their AI development, win the race. But what actually happened: Cutting China off forced their entire AI ecosystem to stop building on Nvidia and optimize for domestic hardware instead. The very policy designed to kneecap them ACCELERATED their independence. Huawei just had the largest single year in company history. Chinese chip companies are going public. SMIC is scaling. Jensen's line: "We caused it." And while Washington celebrates the export controls, Jensen is watching 40% of the world's AI developers - most of them Chinese - slowly stop building on the American tech stack. That's the real national security threat nobody's talking about. But the China argument is just one piece of what Jensen revealed... He also dismantled the entire "escape Nvidia" story in two sentences: "Without Anthropic, why would there be any TPU growth at all? It's 100% Anthropic." Google's TPUs. Amazon's Trainium. Every ASIC challenging Nvidia. All of it traces back to ONE customer and funding relationship. Strip Anthropic out and the chip war narrative Wall Street has been pricing in for two years collapses overnight. And Anthropic only ended up on competing chips because Jensen missed his window. When they were raising early, they needed billions - not VC money. Nobody writes a $10 billion check into a lab with no product. Google wrote it. AWS wrote it. Both got compute commitments in return. But Jensen didn't move. His words: "That was my miss." The CEO of the most valuable company on Earth accidentally created his own competition by not writing a check. He's since invested in Anthropic directly. He now owns a piece of the only customer keeping the competition story alive. Then comes the economics that should end the debate entirely: Nvidia gross margins: ~70%. ASIC margins from Broadcom: ~65%. Jensen on stage: "What are you really saving?" Companies burned billions on custom silicon, wrecked their roadmaps, took on massive execution risk - to save 5 margin points on compute they STILL can't benchmark publicly against Nvidia. Jensen's open invitation: "I welcome Trainium to demonstrate their 40% cost advantage. Come to InferenceMAX." So here's the full picture Jensen painted this week: - The export controls created the competitor everyone feared - The chip war is one customer deep - The CEO who missed that customer now owns a piece of them - The companies spending billions to escape Nvidia can't justify the math in public There's no uprising. There's no escape. Just a lot of very expensive stories with nowhere to go. What do you think?

Ricardo

11,290 次观看 • 5 个月前

I nearly 2x'd the speed while only using +1GB VRAM with the new MTP update in llama.cpp 🤯 You need to add these flags to start using it: --spec-type draft-mtp \ --spec-draft-p-min 0.75 \ --spec-draft-n-max 2 My results with Qwen3.6 27B on a single RTX 5080 ↓ ⚪️ no flag (without mtp) → 54.3 tok/s with 13.26GB VRAM 🔵 --spec-draft-n-max 2 → 90.7 tok/s with 14.29GB VRAM 🔴 --spec-draft-n-max 2 --spec-draft-p-min 0.75 → 93.9 tok/s with 14.30GB VRAM 🟢 --spec-draft-n-max 6 --spec-draft-p-min 0.75 → 93.9 tok/s with 14.87GB VRAM Increasing to 6 draft tokens didn't help my setup for some reason. I made sure to test with a low context length to have enough headroom and eliminate risk of vram stress. From my understanding: 1) The speed gains are very task-dependent. You need to test across a wide range of tasks to get a realistic idea of the benefits 2) We’re already running heavily quantized GGUF models (Q3, Q4, Q6, etc.), so we already benefit from strong speed/performance thanks to the reduced size. That’s why some people are seeing little to no improvement compared to MLX or other quantized versions The progress over the past few days has been insane to say the least. However, MTP now consumes significantly more VRAM. Personally 16GB just isn't enough to use MTP and run it with a good context size. Time to upgrade lads, 24GB+ users are eating GOOD today 🔥 Full setup below ↓
0:33

I nearly 2x'd the speed while only using +1GB VRAM with the new MTP update in llama.cpp 🤯 You need to add these flags to start using it: --spec-type draft-mtp \ --spec-draft-p-min 0.75 \ --spec-draft-n-max 2 My results with Qwen3.6 27B on a single RTX 5080 ↓ ⚪️ no flag (without mtp) → 54.3 tok/s with 13.26GB VRAM 🔵 --spec-draft-n-max 2 → 90.7 tok/s with 14.29GB VRAM 🔴 --spec-draft-n-max 2 --spec-draft-p-min 0.75 → 93.9 tok/s with 14.30GB VRAM 🟢 --spec-draft-n-max 6 --spec-draft-p-min 0.75 → 93.9 tok/s with 14.87GB VRAM Increasing to 6 draft tokens didn't help my setup for some reason. I made sure to test with a low context length to have enough headroom and eliminate risk of vram stress. From my understanding: 1) The speed gains are very task-dependent. You need to test across a wide range of tasks to get a realistic idea of the benefits 2) We’re already running heavily quantized GGUF models (Q3, Q4, Q6, etc.), so we already benefit from strong speed/performance thanks to the reduced size. That’s why some people are seeing little to no improvement compared to MLX or other quantized versions The progress over the past few days has been insane to say the least. However, MTP now consumes significantly more VRAM. Personally 16GB just isn't enough to use MTP and run it with a good context size. Time to upgrade lads, 24GB+ users are eating GOOD today 🔥 Full setup below ↓

left curve dev

30,086 次观看 • 4 个月前

Just woke my sorry ass up from a code-induced coma after spending the night optimizing the rendering for #GTA3 on the Sega #Dreamcast... still have a SHITTON more code to go through, but the results are looking fantastic! Here's the second half of the opening intro! A bunch of the special effects such as screen-space bloom, spot-lights, fog, and rain have made their way onto the DC, and while it looks like there are a few graphical glitches here and there, the aesthetic of the original game is starting to be realized without impacting performance too much! I need to go approve a few PRs into the KallistiOS repository to buy us a few more KB of RAM back to ensure that a 16MB stock DC can make it this far into the intro cutscene without running out of memory, though, along with giving us the space to enable these -O3 optimizations on all builds, as they're currently an optional build flag only enabled when targeting units with the 32MB RAM expansion. Right now 16MB builds are currently only optimized at -O2, which yields a drop of maybe 3-5 FPS or so. ON IT! The rendering glitches with the spotlights are almost certainly textbook graphics pipeline state change leaks (common especially with OpenGL-style APIs which manipulate global static state), and will just take a little more sleuthing to resolve. Finally, there is still a SHITLOAD of room left for increased performance on the rendering side... I will be continuing my work there, Jaxyn will be implementing the final portion of the near-plane clipping algorithm, and our fearless leader, Stefanos Kornilios Mitsis Poiitidis will be tackling everyone's favorite thing to ask about: audio. Thank you all so much for your continued support and for proving that, despite what you may have been told, the Sega Dreamcast is still alive and kicking! ...and maybe, just maybe, the narrative you've all been fed (and many of you are still parroting) about the DC not being able to handle this game was fake news. ;) #gamedev #retrogaming #indiegame
1:40

Just woke my sorry ass up from a code-induced coma after spending the night optimizing the rendering for #GTA3 on the Sega #Dreamcast... still have a SHITTON more code to go through, but the results are looking fantastic! Here's the second half of the opening intro! A bunch of the special effects such as screen-space bloom, spot-lights, fog, and rain have made their way onto the DC, and while it looks like there are a few graphical glitches here and there, the aesthetic of the original game is starting to be realized without impacting performance too much! I need to go approve a few PRs into the KallistiOS repository to buy us a few more KB of RAM back to ensure that a 16MB stock DC can make it this far into the intro cutscene without running out of memory, though, along with giving us the space to enable these -O3 optimizations on all builds, as they're currently an optional build flag only enabled when targeting units with the 32MB RAM expansion. Right now 16MB builds are currently only optimized at -O2, which yields a drop of maybe 3-5 FPS or so. ON IT! The rendering glitches with the spotlights are almost certainly textbook graphics pipeline state change leaks (common especially with OpenGL-style APIs which manipulate global static state), and will just take a little more sleuthing to resolve. Finally, there is still a SHITLOAD of room left for increased performance on the rendering side... I will be continuing my work there, Jaxyn will be implementing the final portion of the near-plane clipping algorithm, and our fearless leader, Stefanos Kornilios Mitsis Poiitidis will be tackling everyone's favorite thing to ask about: audio. Thank you all so much for your continued support and for proving that, despite what you may have been told, the Sega Dreamcast is still alive and kicking! ...and maybe, just maybe, the narrative you've all been fed (and many of you are still parroting) about the DC not being able to handle this game was fake news. ;) #gamedev #retrogaming #indiegame

Falco Girgis

24,802 次观看 • 2 年前

this clip, a plumber breaks down the U.S. Department of Energy’s updated efficiency standards for residential water heaters. These standards, which apply to new models manufactured after May 2029, set higher performance requirements — particularly for larger electric storage tanks. The practical result is that newer electric water heaters will increasingly use heat pump technology, which is significantly more efficient than traditional electric resistance models. Homeowners who choose to upgrade proactively can enjoy meaningful long-term energy savings. According to DOE analysis, switching to a qualifying heat pump water heater can deliver average lifetime savings of around $1,800–$2,000 per unit compared to older standard models, with many households seeing hundreds of dollars reduced from their annual utility bills. These savings can be increased even further by adding two simple, cost-effective upgrades: • A basic timer (or smart controller) so the water heater only runs during off-peak hours or when hot water is actually needed. • Professional-grade Armaflex (or equivalent closed-cell elastomeric foam) insulation wrap on the tank and hot water pipes. This high-quality insulation dramatically reduces standby heat loss — often cutting additional energy waste by 10–25% or more depending on your current setup and climate. Together, the more efficient unit + timer + professional insulation wrap creates a system that uses far less electricity while still delivering reliable hot water. Many homeowners find the combined upgrades pay for themselves through lower monthly bills over time, without any requirement to change anything immediately. This approach gives you control over your energy use and long-term costs on your own timeline. I’m not much on government mandates but I think they’re onto something here this new innovation in water heaters can be a game changer as far as savings if we incorporate a timer and a professional grade insulation job, The DIY’er can really clean up with this one, I put a link to an AO Smith 50 gallon in the first comment, I do highly recommend the expansion tank, it will definitely help with the longevity.
3:32

this clip, a plumber breaks down the U.S. Department of Energy’s updated efficiency standards for residential water heaters. These standards, which apply to new models manufactured after May 2029, set higher performance requirements — particularly for larger electric storage tanks. The practical result is that newer electric water heaters will increasingly use heat pump technology, which is significantly more efficient than traditional electric resistance models. Homeowners who choose to upgrade proactively can enjoy meaningful long-term energy savings. According to DOE analysis, switching to a qualifying heat pump water heater can deliver average lifetime savings of around $1,800–$2,000 per unit compared to older standard models, with many households seeing hundreds of dollars reduced from their annual utility bills. These savings can be increased even further by adding two simple, cost-effective upgrades: • A basic timer (or smart controller) so the water heater only runs during off-peak hours or when hot water is actually needed. • Professional-grade Armaflex (or equivalent closed-cell elastomeric foam) insulation wrap on the tank and hot water pipes. This high-quality insulation dramatically reduces standby heat loss — often cutting additional energy waste by 10–25% or more depending on your current setup and climate. Together, the more efficient unit + timer + professional insulation wrap creates a system that uses far less electricity while still delivering reliable hot water. Many homeowners find the combined upgrades pay for themselves through lower monthly bills over time, without any requirement to change anything immediately. This approach gives you control over your energy use and long-term costs on your own timeline. I’m not much on government mandates but I think they’re onto something here this new innovation in water heaters can be a game changer as far as savings if we incorporate a timer and a professional grade insulation job, The DIY’er can really clean up with this one, I put a link to an AO Smith 50 gallon in the first comment, I do highly recommend the expansion tank, it will definitely help with the longevity.

ℳ𝒜𝒢𝒜_𝒳_𝒯𝒾𝓂𝑒𝓈 𝒟𝒶𝒾𝓁𝓎 𝒩𝑒𝓌𝓈🇺🇸

151,514 次观看 • 3 个月前

So, here's my take on the recent discovery in the ongoing Monster Hunter: Wilds PC performance fiasco. I needed to wait a bit before posting as more sources were able to confirm the initial posts. The main reason I wanted to make this post is because I've seen a lot of clowns recently saying that because the DLC checks only affected hub/camp performance that it "doesn't matter." Their inability to consider how these issues may be affecting other hunters is frankly quite loathsome. I've been one of the fortunate ones who had the "season pass" at launch, so the issues never struck me thankfully. But think of all those that didn't... It DOES matter. Being quite attuned to the qualms of this community myself, I've had hundreds of people stop by this past year asking: "how can I fix the game crashing when I enter the hub?" Many of my community and I were dumbfounded as to why this issues was persisting with a seemingly fresh flow of new players experiencing hub crashes. Even those who weren't crashing were getting sub 30 fps dips, sometimes in the single digits that made menu navigation and hunting prep a real challenge. It annoys me immensely as all of this could have been avoided simply by NOT implementing aggressive anti-piracy measures. It's funny to think that the ACTUAL pirates will have a smoother running gameplay experience than the legitimate customers because they don't have to deal with all the anti-tamper. Now I LOVE Monster Hunter: Wilds, but my love for it doesn't mean I'm delusional, and thus blindly support it regardless of it's faults. We in the community have a duty to OUR viewers, supporters & fans to inform them and call out the shitty monetization practices in the industry. We also have (I believe) a role to play in preventing events like this from occurring in the future by vocalizing our discontent with said issues. Now I have no doubt that Capcom will fix this problem. I have no doubt in my minds that the devs, designers, artists and creators involved in the making of Wilds played no part in this egregious F'up. But someone knew. Somebody in a suit made the decision to completely ruin Monster Hunter: Wilds on PC for potentially thousands of loyal customers. This aggressive approach to anti piracy needs to stop. Capcom, you're actively killing your games for many of your most loyal fans and I don't even think you realise it. As a Capcom creator, I implore you to take a look at what you're doing, and stop. Or I can see Monster Hunter going the way of many other beloved franchises: straight in the bin. Guys, stop blindly brushing off issues in Monster Hunter, just because it's our collective favourite. If you do, there won't BE a Monster Hunter worth defending. Enjoy the game, (as I have myself) but demand it to be better! Cheers, have a lovely jubbly weekend. #MonsterHunterWilds #MonsterHunter #MHWilds
1:00

So, here's my take on the recent discovery in the ongoing Monster Hunter: Wilds PC performance fiasco. I needed to wait a bit before posting as more sources were able to confirm the initial posts. The main reason I wanted to make this post is because I've seen a lot of clowns recently saying that because the DLC checks only affected hub/camp performance that it "doesn't matter." Their inability to consider how these issues may be affecting other hunters is frankly quite loathsome. I've been one of the fortunate ones who had the "season pass" at launch, so the issues never struck me thankfully. But think of all those that didn't... It DOES matter. Being quite attuned to the qualms of this community myself, I've had hundreds of people stop by this past year asking: "how can I fix the game crashing when I enter the hub?" Many of my community and I were dumbfounded as to why this issues was persisting with a seemingly fresh flow of new players experiencing hub crashes. Even those who weren't crashing were getting sub 30 fps dips, sometimes in the single digits that made menu navigation and hunting prep a real challenge. It annoys me immensely as all of this could have been avoided simply by NOT implementing aggressive anti-piracy measures. It's funny to think that the ACTUAL pirates will have a smoother running gameplay experience than the legitimate customers because they don't have to deal with all the anti-tamper. Now I LOVE Monster Hunter: Wilds, but my love for it doesn't mean I'm delusional, and thus blindly support it regardless of it's faults. We in the community have a duty to OUR viewers, supporters & fans to inform them and call out the shitty monetization practices in the industry. We also have (I believe) a role to play in preventing events like this from occurring in the future by vocalizing our discontent with said issues. Now I have no doubt that Capcom will fix this problem. I have no doubt in my minds that the devs, designers, artists and creators involved in the making of Wilds played no part in this egregious F'up. But someone knew. Somebody in a suit made the decision to completely ruin Monster Hunter: Wilds on PC for potentially thousands of loyal customers. This aggressive approach to anti piracy needs to stop. Capcom, you're actively killing your games for many of your most loyal fans and I don't even think you realise it. As a Capcom creator, I implore you to take a look at what you're doing, and stop. Or I can see Monster Hunter going the way of many other beloved franchises: straight in the bin. Guys, stop blindly brushing off issues in Monster Hunter, just because it's our collective favourite. If you do, there won't BE a Monster Hunter worth defending. Enjoy the game, (as I have myself) but demand it to be better! Cheers, have a lovely jubbly weekend. #MonsterHunterWilds #MonsterHunter #MHWilds

☕ 'Twas By Design ☕

22,539 次观看 • 8 个月前

🚨Thomas Tuchel on England’s World Cup semi-final defeat to Argentina. 🗣️ “I have to say it because everyone watched the same match. The referee and VAR completely lost control of this game. In a World Cup semi-final, mistakes like these are unacceptable. You work four years for moments like this, and then crucial decisions are handled with unbelievable incompetence.” “The challenge on Anderson? How is that not given? The offside decisions? The fouls ignored? VAR had every angle, every replay, every opportunity to correct obvious errors and they still failed. If technology cannot get the biggest decisions right on the biggest stage, then what exactly is it there for?” “I don’t want excuses because England also made mistakes, but don’t ask me to stay silent when my players are denied fair decisions. We are told to respect referees, yet who takes responsibility when they get it completely wrong? Nobody. They apologise after the tournament while our dream is already over.” “This is a World Cup semi-final, not a training match. The standards should be the highest in football, yet tonight they were nowhere near good enough. One bad decision can change history, and England were forced to play against both Argentina and a refereeing performance that simply wasn’t at the required level.” “Argentina are a fantastic team and they punished our mistakes, but I refuse to pretend the officiating had no impact. It did. The referee and VAR created controversy from the first whistle until the last, and that should embarrass everyone responsible.” “England leave with heartbreak. Not only because we failed to reach the final, but because every player will always wonder what might have happened if the biggest match of their lives had been officiated properly. That is impossible to accept.”
0:43

🚨Thomas Tuchel on England’s World Cup semi-final defeat to Argentina. 🗣️ “I have to say it because everyone watched the same match. The referee and VAR completely lost control of this game. In a World Cup semi-final, mistakes like these are unacceptable. You work four years for moments like this, and then crucial decisions are handled with unbelievable incompetence.” “The challenge on Anderson? How is that not given? The offside decisions? The fouls ignored? VAR had every angle, every replay, every opportunity to correct obvious errors and they still failed. If technology cannot get the biggest decisions right on the biggest stage, then what exactly is it there for?” “I don’t want excuses because England also made mistakes, but don’t ask me to stay silent when my players are denied fair decisions. We are told to respect referees, yet who takes responsibility when they get it completely wrong? Nobody. They apologise after the tournament while our dream is already over.” “This is a World Cup semi-final, not a training match. The standards should be the highest in football, yet tonight they were nowhere near good enough. One bad decision can change history, and England were forced to play against both Argentina and a refereeing performance that simply wasn’t at the required level.” “Argentina are a fantastic team and they punished our mistakes, but I refuse to pretend the officiating had no impact. It did. The referee and VAR created controversy from the first whistle until the last, and that should embarrass everyone responsible.” “England leave with heartbreak. Not only because we failed to reach the final, but because every player will always wonder what might have happened if the biggest match of their lives had been officiated properly. That is impossible to accept.”

The Arsenal Stand

1,830,204 次观看 • 2 个月前

Grok Build CLI just got an update with v0.2.113! 🚀 Changelog v0.2.113 Features: • MCP servers can now be enabled or disabled directly from the CLI with `grok mcp enable ` and `grok mcp disable `. • Full plan markdown can now be copied to the clipboard with `y` during plan approval or preview. • Added support for the new SuperGrok Plus subscription tier in authentication and feature gating. • Enabled automatic recovery from repetitive loops in model output by default. Performance: • Cold start shows the UI instantly while models and settings load in the background. • Large session forks and resumes now use far less memory and avoid spikes. • Prevented thread exhaustion on high-core shared machines by limiting the workspace daemon's worker threads. Bug Fixes: • Terminal command output is no longer lost or duplicated when the gateway is unreachable. • Invalid MCP server entries in config.toml no longer prevent Grok from starting; problems are shown in `grok inspect`. • SessionEnd hooks now run on exit in non-leader TUI and headless sessions. • Paste chips now display with the correct background in inline prompts and question inputs. • Pasted content chips now behave consistently when editing answers in the question view. • Background task status now shows only elapsed duration instead of absolute timestamps. • Session lists no longer drop real sessions when the remote registry reports an outdated turn count of zero. • /loop now stores prompts that include stop conditions so recurring tasks can terminate themselves when done. • Reduced spurious warning messages for common auth and config scenarios. • Fixed conda activation (and other sourced scripts that read $@) when using persistent or login-capture shells. • Fixed stuck background-task tray rows after long foreground shell commands complete. • Agent subprocesses and idle inhibitors are now cleaned up when the parent CLI process dies unexpectedly. • Fixed truncated plans in minimal mode and improved visual separation between reasoning and output (including NO_COLOR). • Fixed credential loss across multiple grok processes sharing the same auth file. • Fixed doubled Enter and other keys on older Alacritty terminals. • Fixed false paywall messages for free-tier and unmatched users.
0:49

Grok Build CLI just got an update with v0.2.113! 🚀 Changelog v0.2.113 Features: • MCP servers can now be enabled or disabled directly from the CLI with `grok mcp enable ` and `grok mcp disable `. • Full plan markdown can now be copied to the clipboard with `y` during plan approval or preview. • Added support for the new SuperGrok Plus subscription tier in authentication and feature gating. • Enabled automatic recovery from repetitive loops in model output by default. Performance: • Cold start shows the UI instantly while models and settings load in the background. • Large session forks and resumes now use far less memory and avoid spikes. • Prevented thread exhaustion on high-core shared machines by limiting the workspace daemon's worker threads. Bug Fixes: • Terminal command output is no longer lost or duplicated when the gateway is unreachable. • Invalid MCP server entries in config.toml no longer prevent Grok from starting; problems are shown in `grok inspect`. • SessionEnd hooks now run on exit in non-leader TUI and headless sessions. • Paste chips now display with the correct background in inline prompts and question inputs. • Pasted content chips now behave consistently when editing answers in the question view. • Background task status now shows only elapsed duration instead of absolute timestamps. • Session lists no longer drop real sessions when the remote registry reports an outdated turn count of zero. • /loop now stores prompts that include stop conditions so recurring tasks can terminate themselves when done. • Reduced spurious warning messages for common auth and config scenarios. • Fixed conda activation (and other sourced scripts that read $@) when using persistent or login-capture shells. • Fixed stuck background-task tray rows after long foreground shell commands complete. • Agent subprocesses and idle inhibitors are now cleaned up when the parent CLI process dies unexpectedly. • Fixed truncated plans in minimal mode and improved visual separation between reasoning and output (including NO_COLOR). • Fixed credential loss across multiple grok processes sharing the same auth file. • Fixed doubled Enter and other keys on older Alacritty terminals. • Fixed false paywall messages for free-tier and unmatched users.

Puck

94,737 次观看 • 1 个月前

First Impression About Neverness To Everness CBT This will be a long or short one and I mostly have plenty of good to say about the game so far and some bad. First I want to get out the way: Graphics In short, we already know its good! But this is definitely playing some part on performance on my end and it wasn't entirely bad experience at all just some fps drops here and there but its playable and 90% of time it was smooth so the performance issues is a easy fix (apart from some I've seen have already started having crash issues so it'll be different for all of us depending on your specs) I'm well in range of the specs so if I'm having some small issues that mean whoever has worse specs will have it worse. Performance/Optimizing the game etc is not something that should be ignored anyway and most of the time developers of recent time feel like they don't care about improving it, I hope NTE by release or during the test will fix the issues by then (crash problems etc + its day one of the test anyway) Second: Combat Just like wuthering waves, the games combat is fun (I still think it can benefit more on having more effects/impact frames etc to really nail the combat experience...) I know people have compared some characters to WuWa characters as "plagirism" but most part alot of characters in the game have their own charm and uniqueness to the combat The boss enemies also may have some unpredictable patterns and can put alot of pressure or aggression on you but isn't exactly too hard thankfully! But its still great because I like how fast pace and intense the fightings are when your in the boss fights Common enemies don't have much going on for them. Third: Story Unlike Genshin and Wuthering Waves the story for NTE or Neverness to Everness is actually more "Human" one of my viewers described it as and I actually agree, I'm not saying WuWa nor Genshin storytelling is bad but the way NTE tells its story is more morden and you can easily be engaged with it and understand everything clearly! The story is also pretty goofy and just like zenless they have great animations and (VERY) expressive with their animations which makes for great overall experience and I prefer this storytelling the most out of other games. (this is a personal preference) The gacha system is some sort of board game but the game gacha is F2P friendly?(For now for what I've seen) and there's no 50/50.. I was also almost able to max out alot of characters that aren't S rank in almost a day too Thats all so far, I'm still experiencing more as I go but this sums up my 5 hours so far into the game on stream and would love to see y'all in the next stream tomorrow. After I reach 50 hours or more I will do a video in conclusion about the game and what more I would love to see. TDLR: The game is good/great don't worry, just some small stuff that can make it better or some stuff that need fixing (despite the combat plagirism drama that was going on)
0:30

First Impression About Neverness To Everness CBT This will be a long or short one and I mostly have plenty of good to say about the game so far and some bad. First I want to get out the way: Graphics In short, we already know its good! But this is definitely playing some part on performance on my end and it wasn't entirely bad experience at all just some fps drops here and there but its playable and 90% of time it was smooth so the performance issues is a easy fix (apart from some I've seen have already started having crash issues so it'll be different for all of us depending on your specs) I'm well in range of the specs so if I'm having some small issues that mean whoever has worse specs will have it worse. Performance/Optimizing the game etc is not something that should be ignored anyway and most of the time developers of recent time feel like they don't care about improving it, I hope NTE by release or during the test will fix the issues by then (crash problems etc + its day one of the test anyway) Second: Combat Just like wuthering waves, the games combat is fun (I still think it can benefit more on having more effects/impact frames etc to really nail the combat experience...) I know people have compared some characters to WuWa characters as "plagirism" but most part alot of characters in the game have their own charm and uniqueness to the combat The boss enemies also may have some unpredictable patterns and can put alot of pressure or aggression on you but isn't exactly too hard thankfully! But its still great because I like how fast pace and intense the fightings are when your in the boss fights Common enemies don't have much going on for them. Third: Story Unlike Genshin and Wuthering Waves the story for NTE or Neverness to Everness is actually more "Human" one of my viewers described it as and I actually agree, I'm not saying WuWa nor Genshin storytelling is bad but the way NTE tells its story is more morden and you can easily be engaged with it and understand everything clearly! The story is also pretty goofy and just like zenless they have great animations and (VERY) expressive with their animations which makes for great overall experience and I prefer this storytelling the most out of other games. (this is a personal preference) The gacha system is some sort of board game but the game gacha is F2P friendly?(For now for what I've seen) and there's no 50/50.. I was also almost able to max out alot of characters that aren't S rank in almost a day too Thats all so far, I'm still experiencing more as I go but this sums up my 5 hours so far into the game on stream and would love to see y'all in the next stream tomorrow. After I reach 50 hours or more I will do a video in conclusion about the game and what more I would love to see. TDLR: The game is good/great don't worry, just some small stuff that can make it better or some stuff that need fixing (despite the combat plagirism drama that was going on)

Aruzien

55,067 次观看 • 1 年前

How did a tiny team of 30 engineers build WhatsApp, more than a decade ago? From Jean Lee, engineer #19 at the company. Timestamps: 00:00 Intro 01:39 Early years in tech 06:18 Becoming engineer #19 at WhatsApp 13:53 WhatsApp’s tech stack 18:09 WhatsApp’s unique ways of working 25:27 Countdown displays and outages 27:07 Why WhatsApp won 28:53 The Facebook acquisition 33:13 Life after acquisition 39:27 Working at Facebook in London 44:07 Transitioning to management 47:27 Performance reviews as a manager 53:29 After Facebook 58:53 AI’s impact on engineering 1:02:34 Jean’s advice to new grads and startups 1:06:45 Empowering employees 1:08:17 Book recommendations Watch or listen: • YouTube: • Spotify: • Apple: Brought to you by: • Statsig – ⁠ The unified platform for flags, analytics, experiments, and more. • Sonar – The makers of SonarQube, the industry standard for automated code review. • WorkOS – Everything you need to make your app enterprise ready Three interesting observations from this episode: 1. WhatsApp had no code reviews after in-place. WhatsApp cofounder, Brian Acton, reviewed the very first pull request of each new hire, and after that, there were no more code reviews. Jean recounts how Brian reviewed her debut PR in extreme detail. This first (and only!) review set the bar high, and she wrote code to that standard from then on. 2. WhatsApp had close to zero formal processes. WhatsApp had no Scrum, no Agile, no TDD (test driven development), and no formal code reviews beyond the first commit. In contrast, Skype had 1,000 engineers and mandatory Scrum training, but WhatsApp still outcompeted it and won. Jean’s response to hearing of all the formal processes Skype used in order to execute faster: “I’m surprised to hear they thought they were shipping faster because of it.” Perhaps process is often a substitute for trust, not quality?” 3. Saying “no” to features was a competitive advantage. WhatsApp’s CEO, Jan Koum, rejected 99% of feature requests from the team. While competitors shipped dozens of shiny, new features, WhatsApp ruthlessly prioritized reliability and simplicity. Jan repeatedly told the team what the mission was. “I want a grandma living in the countryside to be able to use our app”, he said.
1:11:24

How did a tiny team of 30 engineers build WhatsApp, more than a decade ago? From Jean Lee, engineer #19 at the company. Timestamps: 00:00 Intro 01:39 Early years in tech 06:18 Becoming engineer #19 at WhatsApp 13:53 WhatsApp’s tech stack 18:09 WhatsApp’s unique ways of working 25:27 Countdown displays and outages 27:07 Why WhatsApp won 28:53 The Facebook acquisition 33:13 Life after acquisition 39:27 Working at Facebook in London 44:07 Transitioning to management 47:27 Performance reviews as a manager 53:29 After Facebook 58:53 AI’s impact on engineering 1:02:34 Jean’s advice to new grads and startups 1:06:45 Empowering employees 1:08:17 Book recommendations Watch or listen: • YouTube: • Spotify: • Apple: Brought to you by: • Statsig – ⁠ The unified platform for flags, analytics, experiments, and more. • Sonar – The makers of SonarQube, the industry standard for automated code review. • WorkOS – Everything you need to make your app enterprise ready Three interesting observations from this episode: 1. WhatsApp had no code reviews after in-place. WhatsApp cofounder, Brian Acton, reviewed the very first pull request of each new hire, and after that, there were no more code reviews. Jean recounts how Brian reviewed her debut PR in extreme detail. This first (and only!) review set the bar high, and she wrote code to that standard from then on. 2. WhatsApp had close to zero formal processes. WhatsApp had no Scrum, no Agile, no TDD (test driven development), and no formal code reviews beyond the first commit. In contrast, Skype had 1,000 engineers and mandatory Scrum training, but WhatsApp still outcompeted it and won. Jean’s response to hearing of all the formal processes Skype used in order to execute faster: “I’m surprised to hear they thought they were shipping faster because of it.” Perhaps process is often a substitute for trust, not quality?” 3. Saying “no” to features was a competitive advantage. WhatsApp’s CEO, Jan Koum, rejected 99% of feature requests from the team. While competitors shipped dozens of shiny, new features, WhatsApp ruthlessly prioritized reliability and simplicity. Jan repeatedly told the team what the mission was. “I want a grandma living in the countryside to be able to use our app”, he said.

Gergely Orosz

112,027 次观看 • 6 个月前

Peter Thiel’s Founders Fund just led a 2nd round into Icon. To celebrate, I'll give you $69,420 if you refer someone I hire. We’ve gone from $0 to $5M+ ARR in 30 days, are extremely profitable, & will beat every company ever on the $1M to $100M ARR record (<12 months). Our team also went from 8 to 16 people. We were founded less than 6 months ago & our 1st launch was a failure. I don't have a trust fund like other founders so failure is unacceptable: we quickly pivoted to being The AI Admaker & this time we didn’t miss. Our 2nd launch got 5000+ inbound demos, 1M+ website visits, 10M+ views across socials, 50K reactions/comments, & Fortune 500 customers. Here’s how we did it: * Only hire top 0.01% engineers with no life. This results in the best product. If your GitHub isn’t a dark green carpet (3000+ commits per year including weekends), it’s a red flag. * 7 day work week. The 4 day work week leads to defeat. 996 also results in defeat (why do 6 when you can do 7). It’s a cope to think working more hours doesn’t help you get more done. * Pay 2-5x above market because someone working this hard deserves it. On a personal note: this has been a journey full of ups & downs & I’m extremely grateful to be here. I grew up in Portland, Oregon & found school extremely boring. I slept through every class because of this. If I’m being real, no one expected me to go anywhere. This all changed when I was 18. I had a big wakeup call & had to start supporting myself. I started skipping class to code 100 hours a week & combined with my hate for school, I dropped out after 1 year to join Pinterest. Pinterest paid very well, but was very boring. After too long of this, I started my 1st company, Skio, & got into YC. I ate shit during the batch & had to pivot. Thankfully, post-pivot Skio is killing it: $15M+ ARR & profitable on seed funding in just a few years. All these up-and-down adventures led to Icon. Icon is the culmination of every lesson I've learned (I made so many stupid mistakes) in these 10 years of grinding. Thank you from the bottom of my heart to our customers, my team, & our investors. Again, I’m extremely grateful to have this opportunity & will work 24/7 to honor your trust in me by becoming a trillionaire. Icon’s $10T+ vision is simple: Part 1 is Cursor for Ads: we’ll set the world record for time from $1M to $100M ARR. Part 2 is Cursor for Marketing: when we pull this off, we’ll be a $10T+ company that creates generational wealth for everyone involved. I know $10T is crazy, but with inflation we got this. If you’re top 0.01%, extremely hungry, & seek generational wealth, I’d love to chat. I'm sure everyone knows 1 person like this. Especially recruiters. If you do, I’ll give you $69,420 for referring someone we hire ($69,420 for Founding Engineers & $30,000 for Marketers / Sales / Ops / CX / Anything Else). Comment “refer” & I’ll send you our referral form.
0:35

Peter Thiel’s Founders Fund just led a 2nd round into Icon. To celebrate, I'll give you $69,420 if you refer someone I hire. We’ve gone from $0 to $5M+ ARR in 30 days, are extremely profitable, & will beat every company ever on the $1M to $100M ARR record (<12 months). Our team also went from 8 to 16 people. We were founded less than 6 months ago & our 1st launch was a failure. I don't have a trust fund like other founders so failure is unacceptable: we quickly pivoted to being The AI Admaker & this time we didn’t miss. Our 2nd launch got 5000+ inbound demos, 1M+ website visits, 10M+ views across socials, 50K reactions/comments, & Fortune 500 customers. Here’s how we did it: * Only hire top 0.01% engineers with no life. This results in the best product. If your GitHub isn’t a dark green carpet (3000+ commits per year including weekends), it’s a red flag. * 7 day work week. The 4 day work week leads to defeat. 996 also results in defeat (why do 6 when you can do 7). It’s a cope to think working more hours doesn’t help you get more done. * Pay 2-5x above market because someone working this hard deserves it. On a personal note: this has been a journey full of ups & downs & I’m extremely grateful to be here. I grew up in Portland, Oregon & found school extremely boring. I slept through every class because of this. If I’m being real, no one expected me to go anywhere. This all changed when I was 18. I had a big wakeup call & had to start supporting myself. I started skipping class to code 100 hours a week & combined with my hate for school, I dropped out after 1 year to join Pinterest. Pinterest paid very well, but was very boring. After too long of this, I started my 1st company, Skio, & got into YC. I ate shit during the batch & had to pivot. Thankfully, post-pivot Skio is killing it: $15M+ ARR & profitable on seed funding in just a few years. All these up-and-down adventures led to Icon. Icon is the culmination of every lesson I've learned (I made so many stupid mistakes) in these 10 years of grinding. Thank you from the bottom of my heart to our customers, my team, & our investors. Again, I’m extremely grateful to have this opportunity & will work 24/7 to honor your trust in me by becoming a trillionaire. Icon’s $10T+ vision is simple: Part 1 is Cursor for Ads: we’ll set the world record for time from $1M to $100M ARR. Part 2 is Cursor for Marketing: when we pull this off, we’ll be a $10T+ company that creates generational wealth for everyone involved. I know $10T is crazy, but with inflation we got this. If you’re top 0.01%, extremely hungry, & seek generational wealth, I’d love to chat. I'm sure everyone knows 1 person like this. Especially recruiters. If you do, I’ll give you $69,420 for referring someone we hire ($69,420 for Founding Engineers & $30,000 for Marketers / Sales / Ops / CX / Anything Else). Comment “refer” & I’ll send you our referral form.

Kennan Frost

833,084 次观看 • 1 年前