Silver Wolf

@JimMcGonigle3 @MurrayHillGuy1 "beautiful mid" girl is one who looks naturally pretty (due to genetics, diet or simple non-gym exercises) BUT does not usually wear much makeup, or style hair too much, or do intensive gym workouts, or wear expensive or fashionable outfits etc. to look very attractive.

@MurrayHillGuy1 I'm old but isn't "beautiful mid" an oxymoron?

My friend has lost 20lbs and is nearly peeled out. (All natural, no Reta) Today, a beautiful mid approached him. He told her he has a girlfriend (he doesn’t). I asked him why he didn’t smash. He replied: “I didn’t put myself in a 1k calorie deficit for 3 months to fuck mid chicks” Lesson in that.

Why did Kimi CEO Yang Zhilin return to China rather than stay in the U.S.? In some of his early interviews, he explained his thinking. His answer was more practical than ideological. In his view, China had become a sufficiently complete environment for pursuing frontier AI: policy support, venture capital, accumulated talent, a strong domestic industry, and a large user base. He summarized one reason in two words: “环境完备。” “The environment was complete.” Yang was not claiming that the U.S. lacked talent or capital. He had studied at Carnegie Mellon and worked at Facebook AI Research and Google Brain. He understood the American ecosystem firsthand. His question was: where could he assemble enough researchers, capital, computing, engineering, product capability, and users to build an AGI company from scratch? In another interview, he put the resource problem bluntly: “它需要人才聚集、资本聚集。” “It requires a concentration of talent and capital.” This was not a small research project. While still in the U.S., Yang calculated that a serious AGI startup would need at least $100 million almost immediately. The second reason was timing. Yang believed that by late 2022, several conditions had converged: the internet had accumulated enormous amounts of data; Transformer architectures had become scalable; semiconductor progress made large-scale training possible; scaling laws suggested that more computation could produce more intelligence; ChatGPT showed that AI could reach ordinary users. The opportunity was no longer merely to publish papers. It was to build a company combining research, engineering, product, and commercialization. Yang described that organization as: “科学、工程和商业的结合。” “A combination of science, engineering, and business.” That explains why Kimi began as a consumer product rather than a pure research lab. Yang wanted to put advanced models in users’ hands, observe how they interacted with them, and use that feedback to shape the technology. He said: “AGI最终会是一个跟所有用户co-work产生的东西。” AGI would ultimately emerge through collaboration with users. China offered a large, concentrated market for that feedback loop—especially around long context, personalization, complex workflows, and agents. But Yang’s ambition was not merely domestic. He explicitly said: “真正AGI肯定是全球化的。” “True AGI will necessarily be global.” So the distinction is important: China was his launch base, not necessarily AGI’s final market. His metaphor was that internet companies could “plant a tree”—define a product and execute toward it. Large-model companies had to “承包一片森林”—take responsibility for an entire forest whose future growth was unpredictable. That required enormous resources, but also an organization willing to tolerate uncertainty. Yang described the founder’s job as: “爬楼梯，而不只是看风景。” “Climb the stairs, rather than merely admire the scenery.” In other words, prioritize improving model capability and exploring the limits of intelligence, while still building a business. So why did he return? Because he believed China had become sufficiently complete to support an ambitious AGI startup. The U.S. gave him world-class training and research experience. China gave him the opportunity to combine capital, talent, engineering, users, and a company into one system. His decision was therefore less “China versus America” than a strategic judgment: Build where the necessary conditions are available. Pursue technology whose ultimate market is global. And let real users—not just benchmarks—help determine what AGI becomes.

@fame_in_focus AI generated?

🫦

Kimi K3 scores 57 on the Artificial Analysis Intelligence Index. Its intelligence is comparable to Opus 4.8 and GPT-5.5 but remains behind Fable 5 and GPT-5.6 Sol. Moonshot AI has expressed plans to release the 2.8T parameter model's weights, which would make it the leading open weights model Key results: ➤ Strong agentic task performance: @Kimi_Moonshot's Kimi K3 reaches an Elo rating of 1668 on GDPval v2. This is a marked improvement over K2.6’s 1190, surpassing GLM-5.2 (1514), GPT-5.5 (1494), and Claude Opus 4.8 (1600). However, it still lags behind Claude Fable 5 (1760). Kimi K3 also scores an impressive 53% and takes the #1 position on AutomationBench-AA, our implementation of Zapier’s Agentic SaaS workflow evaluation. ➤ Second-highest performance on AA-Briefcase (agentic knowledge work): On our private long-horizon knowledge work evaluation, Kimi K3 reaches an overall Elo of 1547, +732 points from Kimi K2.6 and behind only Claude Fable 5. It is well-rounded: its rubric scoring and analytical quality almost reach Claude Fable 5’s scores, while GPT-5.6 Sol continues to outperform other leading models on presentation quality. ➤ Set to lead open weights models once weights are released: Moonshot AI has not yet released the weights but expressed plans to do so. Once available, Kimi K3 would clearly lead other open weights models including GLM-5.2 (51) and DeepSeek v4 Pro (44). However, at 2.8T parameters, it is significantly larger than its open weights peers (eg. GLM-5.2 at 753B params and DeepSeek V4 Pro at 1.6T), as well as the Kimi K2 to K2.6 models (1T params). ➤ Cost per task ($0.94) is similar to GPT-5.6 Sol ($1.04), ~1/2 the price of Opus 4.8 ($1.80) and higher than open weights peers: Moonshot AI’s pricing for K3 is significantly higher than their K2 pricing (K3’s output token price is $15/1M tokens while K2.6 was $4). This positions the model as cheaper on a cost per task basis than Opus 4.8, similar to GPT-5.6 Sol ($1.04) and more expensive than open weights peers, GLM-5.2 ($0.32) and DeepSeek V4 Pro ($0.04) ➤ Improved token efficiency alongside higher intelligence: Kimi K3’s token usage on the Artificial Analysis Intelligence Index decreased significantly, using 21% fewer output tokens than K2.6. The new model used approximately 132M output tokens to complete all nine evaluations, compared to approximately 166M for K2.6, while achieving higher scores. ➤ Native multimodal capabilities: Kimi K3, like K2.6, is released with native image and text multimodal input. If weights are released, this will position Kimi K3 as one of the leading open weights models with multimodal input capabilities Other model details: Context window: 1M Size: 2.8T total parameters Pricing: The first-party API is priced at $3.00/$15.00 per 1M input/output tokens, with cached input discounted 90% to $0.30 per 1M tokens. Modality: Native multimodal input supports text and images, and the model remains text-only for output. Accessibility: Accessible at launch through Moonshot’s first party API. Model weights are not yet released but Moonshot AI has expressed plans to do so.

@mark_k @SpaceXAI Hoping 4.5 comes to Twitter/X and to Grok.com

Major news from @SpaceXAI: Grok 4.5 is now available in Grok Build in Europe! 🇪🇺 You can simply select it from the model selector with /model I expect Grok 4.5 to become available in Cursor later today as well.

@tomgreenhall @MurrayHillGuy1 More than his friends using fraud tactics, it's the women using fraud tactics to screen out honest or average-looking men & accepting dishonest yet attractive men for their casual flings. Anyways, these women are not worthy of a long-term relationship at all, so it's all good.

My friend has been traveling Europe for the last month. He discovered the most broken Tinder/Hinge strategy I’ve ever seen. He’s ONLY matches with American girls on vacation whose bio says: “Looking for a long-term relationship.” They think he lives in Europe. Dude has been going on a different date every other night and basically has pulled every single one of them on first date. Cheat code?

@techdevnotes I believe it can still retrieve data post-Jan 2026, if prompted accordingly.

Grok 4.5 has a pretraining cutoff of January 2026

@__elysianfields @dmoonlives That's nice to hear, but I've heard that many of these girls have boyfriends/husbands back home (on whom they are cheating with), while they go out on these foreign vacations.

@dmoonlives In Tulum & multiple places foreign girls will ask where I’m staying and then offer to be my gf for the rest of my trip.They ask guys that they’re attracted to.They literally will be your gf.They just go to a different guy every couple weeks.Chad-benefits

When women travel, they go on dating apps and schedule a different guy to show them around every day. If they like one, they ghost the others and shack up with foreign chad for the rest of their stay.

Google's Gemini Omni Flash debuts at #1 on the Artificial Analysis Text to Video and Image to Video Leaderboards, edging out ByteDance's Seedance 2.0 on both Gemini Omni Flash is the first model in Google's Gemini Omni family, unveiled at Google I/O in May and opened to developers in public preview on June 30. Google positions Omni as a natively multimodal model that can "create anything from any input", starting with video: it accepts text, images, and video as input, generates clips with native audio, and supports conversational editing, where prompts change a video while preserving the rest of the scene. Gemini Omni Flash generates 3 to 10 second clips at 720p and 24 FPS, in 16:9 or 9:16, with longer durations coming soon. In the Artificial Analysis Video Arena, Gemini Omni Flash debuts at #1 on both the Text to Video and Image to Video Leaderboards, narrowly ahead of ByteDance's Seedance 2.0 on each. Gemini Omni Flash is priced at $0.10 per second of generated video ($6.00 per minute), matching Veo 3.1 Fast. The rate is the same for Text to Video and Image to Video. It is available now in the Gemini API, Google AI Studio, and the Gemini Enterprise Agent Platform, in the Gemini app and Google Flow for consumers, and at no cost in YouTube Shorts and the YouTube Create app. Congratulations to @GoogleDeepMind on the release! See below for comparisons between Gemini Omni Flash and other leading models in the Artificial Analysis Video Arena 🧵

Seedance 2.5 release is imminent from ByteDance

@MurrayHillGuy1 Very average in terms of looks, but very high in terms of attitude? Anyway, she is in a league of her own - a loner...

Fellas?

@OfficeOfCRSY Madelyn Cline?

Golden hour glow in buttery yellow 💛

@venom1s Whether it's sexual harrasment or not, just think - if the BTS guy has a gf/wife & she sees this video uploaded online - imagine how creeped out or hurt she would feel watching other women salivating to his photos? It's ok to have sexual thoughts/fantasies, but this is not ok.

This girl is licking a BTS member's picture. Isn't this sexual harassment? If a man did the same to a female celebrity, girls would be calling him a potential rapist. Why are girls like this? Does male celebrity consent not matter?

@dvorahfr Can u share the prompt used to create the video? The texture & details look really rich & amazing 👍

Meow Grok Imagine

@senb0n22a @techdevnotes I agree that Seedance (ver 2.0) has better output video quality compared to Grok Imagine, however it's more restricted/moderated (no NSFW or R-rated stuff) when compared to Grok Imagine. So overall, I prefer Grok Imagine even with 720p max resolution.

@sprash1982 @techdevnotes it really isn't. doesn't come close to seedance which can actually be used professionally for video, and image is far behind.

We still don’t have a timeline when Grok Imagine 2.0 will be released

@Thrasher66099 @bridgemindai "Grok 4.5 high has them both smoked" - for now. Can SpaceXAI sustain the low pricing for weeks or months from now? If the prices do go up in near future, can they still maintain the cost advantage over GPT 5.6?

@bridgemindai So we're just glazing over the fact that Grok 4.5 high has them both smoked?

GPT 5.6 Sol just hit CursorBench. The economics are brutal for Anthropic. Fable 5 Max: 70.5% at $17.32 per task. 103,525 tokens. GPT 5.6 Sol Max: 67.2% at $5.22 per task. 28,320 tokens. 95% of the performance. A third of the cost. 73% fewer tokens. Fable 5 kept the crown on raw score. It lost on everything that shows up on your invoice. And remember: GPT 5.6 comes with limits you can actually live with. The frontier is not an intelligence war anymore. It is a value war.

GPT-5.6 Sol comes close second to Claude Fable 5 in the Artificial Analysis Intelligence Index at one third of the cost, and leads the Artificial Analysis Coding Agent Index in OpenAI’s Codex harness We supported @OpenAI with pre-release evaluation of GPT-5.6 Sol, Terra, and Luna. GPT-5.6 Sol (max) scores 1 point below Claude Fable 5 (max) in the Artificial Analysis Intelligence Index at 59 points, at approximately one third of the cost. GPT-5.6 Terra (max) and Luna (max) score 55 and 51 respectively in the Intelligence Index, at ~50% and ~80% lower Cost per Task than Sol. GPT-5.6 Sol (max) leads the Artificial Analysis Coding Agent Index at 80 points. Congratulations @OpenAI and @sama on the launch! Key takeaways: ➤ One third of the cost of Claude Fable 5: On max reasoning effort, GPT-5.6 Sol costs $1.04 per task in the Artificial Analysis Intelligence Index - offering a similar level of intelligence to Claude Fable 5 at approximately one third of the cost. Reasoning levels across GPT-5.6 Sol and Luna offer a range of options at the Pareto frontier of Intelligence vs Cost per Task. For example, GPT-5.6 Luna (max) matches or exceeds the intelligence of GLM-5.2 (max) and Gemini 3.5 Flash at a lower cost. GPT-5.6 Terra (max) and Luna (max) cost $0.55 and $0.21 per Intelligence Index task, ~50% and ~80% less than Sol. Across reasoning efforts, each new GPT-5.6 model pushes past GPT-5.5 on the Pareto frontier (excluding non-reasoning). Notably, Luna and Sol are always on the Pareto frontier ahead of Terra. This means that for any Terra effort level, there is a Luna or Sol effort level that is more intelligent at no extra cost, or as intelligent at lower cost. ➤ Leading in all Coding Agent evaluations: The new Artificial Analysis Coding Agent Index pairs models with agentic harnesses and features three frontier coding evaluations - DeepSWE, Terminal-Bench v2, and SWE-Atlas-QnA. GPT-5.6 Sol (max) in Codex scores 80 in the Index, leading in all three evaluations (tying Grok 4.5 in Grok Build for SWE-Atlas-QnA). In addition to scoring higher, its per task cost is ~40% and ~10% cheaper than Claude Fable 5 (max) and Opus 4.8 (max) respectively in Claude Code. GPT-5.6 Terra (max) and Luna (max) score 77 and 75 in the Coding Agent Index respectively, with ~60% and ~80% per-task cost reductions compared to Sol. ➤ Highest Presentation Elo in AA-Briefcase: GPT-5.6 Sol (max) ranks second only to Claude Fable 5 (max) in AA-Briefcase, and has the highest Presentation Elo of any model. AA-Briefcase is a new benchmark for testing models on realistic knowledge work tasks in complex projects built by industry experts. GPT-5.6 Sol (max) has the highest recorded Presentation Elo - its outputs across various file types, including PowerPoint and Excel, are the most visually attractive of any model. Fable 5 (max) still leads AA-Briefcase, largely due to its Rubric Score of 56% vs 42% for GPT-5.6 Sol (max). Fable 5 (max) also scores 1764 in Analytical Quality Elo vs GPT-5.6 Sol (max) at 1592. ➤ First OpenAI models with cache-write pricing: GPT-5.6 introduces cache-write pricing for the first time at OpenAI. Sol, Terra, and Luna are priced at $5/$30, $2.5/$15, and $1/$6 respectively per million input/output tokens. OpenAI has retained its previous discount of 90% for cache reads, but joins Anthropic in introducing a cost premium for cache writes, at 1.25x the price of input tokens. Cache writes occur when input tokens are committed to memory. Charging for a cache write more accurately reflects the model’s cost to serve, as cached tokens occupy memory whether or not they are reused. Also in line with Anthropic's models, GPT-5.6 introduces a max reasoning effort level. ➤ Low token use: GPT-5.6 Sol (max) uses fewer output tokens than most models of comparable intelligence, and defines a new Pareto frontier of Intelligence vs Output Tokens per Task. GPT-5.6 Sol (max) offers a slight improvement in token efficiency with 15k tokens per Intelligence Index task, vs GPT-5.5 at 16k. Notably, it uses fewer tokens and is more intelligent than Claude Opus 4.8 (max), GLM-5.2 (max), and Gemini 3.5 Flash (high).

@farzyness Until it comes to Twitter/X, it won't gain popularity. As of now Twitter is using an older version (4.3, I think).

Grok 4.5 is a banger. Everyone try it. Now!!!!!

@Legionista83 She's girlfriend material 🥰

Natalie Portman