Tested Muse Spark 1.2, Kimi K3, and GLM 5.2... with our FlappyBench 3 models, same prompt with the /design command. Reviewed gameplay features, UX/UI, and cost. 🔹 Kimi K3 → 9.5/10 · $0.0740 🔹 GLM 5.2 → 9/10 · $0.0480 🔹 Muse Spark 1.2 → 8/10 · $0.0187 Muse Spark 1.2 delivered a completely different game style despite being super cheap. Kimi K3 and GLM 5.2 outputs are comparable and close in cost.show more

Fable 5.1 vs Kimi K3 Tested both models with... the same prompt at the highest reasoning level available. > Fable took 18 minutes to complete this test and cost $12.60 > Kimi K3 took 10 minutes and cost around $6.95 Which one wins here?show more

Sakana Fugu surprisingly performed near GLM 5.2 level but... 17× more expensive! We gave the same prompt to 4 models: build a complete live Trader Desk with both frontend and backend components, real-time market data fetched from external APIs for 8 symbols, and a custom dark-theme UI. Outputs: Fugu Ultra — 22,225 t, $0.51 Opus 4.8 — 15,802 t, $0.31 GPT-5.5 — 11,474 t, $0.26 GLM 5.2 — 13,677 t, $0.03 Fugu created the most polished and feature-rich trading desk in the run. GLM 5.2 was very close behind, with a similarly complete multi-panel interface and live data, but at a much lower cost. Opus and GPT also performed well, delivering solid results with a better balance between quality and costshow more

You can now use GPT 5.5, Gemini 3.7 Flash,... Kimi K3 and 47 other AI models completely free😱 No subscription. No credit card. Even the API usage costs $0. AIHubMix just opened a free catalog with 50 AI models. Some of the available models: • Ox Alpha • Gemini 3.7 Flash • GLM 5.2 • Kimi K3 • MiniMax M3 • GPT 5.5 • 40+ more And you don’t need separate API keys for each model. Setup takes 2 minutes: > Step 1: Go to > Create an account using your email or OAuth. No card needed. Step 2: Create one API key > The same key works with every free and paid model. Step 3: Add it to any OpenAI-compatible tool Base URL: Then choose any model ending in -free, such as: coding-glm-5.2-free gpt-5.5-free That’s it. One API key. 50 AI models. $0 for both input and output. Save this. You might need a free multi-model setup later.show more

Grok 4.5 performed GPT Sol level for free! We... gave 4 models the same prompt: build three self-contained HTML5 canvas scenes with real physics demos Prompts: -robot deathmatch, Tombstone vs Minotaur -a hydraulic press flattening stuff on a conveyor -a semi truck jumping a canyon Outputs: GPT-5.6 Sol: 12.9K tokens, $0.51 (~7 min) Grok 4.5: 10.8K tokens, $0 (~5 min) Muse Spark 1.1: 26.8K tokens, $0.12 (~7.5 min) GLM 5.2: 10.9K tokens, $0.02 (~12 min) Grok 4.5 handled all three scenes genuinely well and got surprisingly close to GPT-5.6 this round. On top of that, it ran on the free tier. GPT-5.6 Sol, the frontier model, put out solid but not standout work. GLM 5.2 rendered all three scenes for pennies, but it came out the roughest of the four. Meta's new Muse Spark burned the most tokens yet still stayed cheap, delivering an average result.show more

Fable 5 totally crushed our new contest, but it... cost 6x more than Opus 4.8! We gave 4 models the same prompt: build three self-contained HTML5 canvas scenes with real physics demos Prompts: — A train derailing off a broken bridge into the water — Two cars jumping off ramps and colliding mid-air over a canyon — A monster truck crushing a row of parked cars Outputs: Fable 5: 62,158 tokens, $3.12 GPT 5.5: 37,753 tokens, $1.14 Opus 4.8: 22,280 tokens, $0.56 GLM 5.2: 36,246 tokens, $0.08 Fable 5 did all three scenes at A+. The crashes looked real, things fell and broke the right way, and nothing went through the ground or floated. GPT 5.5 was the closest to Fable. In the Bigfoot show, we think GPT was even a little better. GLM 5.2 did not win any scene, but it was the cheapest by far. Fable is the best pick for quality, but you pay more for it.show more

KIMI K3 + OBSIDIAN + LOOP ENGINEERING = A... VAULT THAT RUNS ITSELF the core idea: the vault is the loop's state, not the chat window everything Kimi K3 knows lives in a .md file the loop: > capture - a thought lands in 00-inbox > context - K3 pulls links, tags, and neighbouring notes > draft - edits happen inside a git worktree, never the live vault > review - a critic agent checks the diff before anything ships > commit - appended to the vault, nothing gets rewritten the key insight: frontmatter fields like supports, contradicts, and supersedes are graph edges, not metadata - the note format is the write API start with a plain loop, it runs about 2-4x the cost of one direct call > only move to a full graph once state has to outlive the session, several agents need to coordinate, or you have to explain what changed - that jump can run 10-50x one review assistant climbed from 55% to 72% to 84% just by moving through these shapes in ordershow more

$180,000 VOLVO SEMI TRUCK. EVERY WIRE AND PIPE IN... THE CAB AND ENGINE BAY MAPPED IN ADVANCE. BUILT BY TYPING SENTENCES INSTEAD OF PLACING THEM BY HAND. A cab and engine bay this dense needed its wiring, cooling lines, and air and brake lines modeled and checked for clashes before a physical prototype could be built. The design team connected Kimi K3 to their CAD software through MCP, then had it read photos of the cab and engine bay alongside plain English descriptions of each system and build the model directly instead of placing every line by hand. Old process: 2 engineers, 3 weeks, around $19,000. New process: 1 engineer reviewing and refining the AI's output, 5 days, roughly $3,500 combined, engineer's time plus AI compute. Routing clashes caught this way cost a few hundred dollars to fix. The same clash caught after a physical prototype is already built costs 10 to 100 times more. See the article below for a smaller example of the Kimi K3 + Blender MCP workflow.show more

New open-source agent harness just landed! I got early... access to TrueForge by TrueFoundry and have been running it locally for the past few days. The harness layer deserves as much attention as the model, and open source matters here because you can inspect the loop, run it on your own infrastructure, and swap to the latest or cheaper models. TrueForge handles the runtime work that makes an agent reliable. It drives the tool-calling loop, manages context, coordinates subagents, and executes code in a sandbox, with any model you choose. Every tool call re-sends the growing context to the model, so in practice the harness controls most of what an agent costs to run. A few things stood out from my testing and their published benchmarks. Vendor-Neutral by design. It runs OpenAI, Anthropic, and Google models alongside open-weight models like Kimi, GLM, and DeepSeek. Model routing is a setting, and you can send each task to the model that fits it. On a 14-task enterprise agent benchmark, it matched the accuracy of Claude Managed Agents running the same Opus 4.8 model at roughly 30% lower cost per run (3.8M tokens vs 10M for the same answers). Routing the same tasks to GLM-5.2 held accuracy and brought cost down by about 75%, around $3 per run instead of $12. Fully self-hosted and Open Source (MIT License). I had it running locally with one command, with sandboxed code execution working out of the box. It's time to own your agent harness. Thanks to TrueFoundry for partnering on this post.show more

Qwen 3.8 27B Q4_K_M - 90 tokens/sec on a... single NVIDIA RTX 4090 (24 GB VRAM) with Dflash2! (MTP 60 tps -> 90 tps Dflash2!!!!) Local AI moves so fast (literally!) it’s terrifying. Z lab just dropped DFlash 2 for Qwen 3.8 27b and Muse Glimmer. I patched llama.cpp (PR #27342) and paired it with Unsloth’s Qwen 3.8 27B UD-Q4_K_XL quant. The result? Lossless 90 tokens/s decode. My last post highlighted native MTP hitting 60 t/s at 130,000 context. But DFlash 2 just completely shattered that ceiling. By using parallel block diffusion drafting (predicting whole blocks of tokens in a single pass using dynamic convolutions), DFlash achieves a massive 5.39 token acceptance rate. THE ALPHA TWEAK: `n-max 7` eats too much VRAM for draft states. But if you drop the draft limit to `--spec-draft-n-max 4`, you slash the VRAM overhead and actually increase the throughput. Here is the new 24GB VRAM Physics Matrix (DFlash 2 @ n-max 4): - 30k Context: 1,725 t/s prefill | 87.05 t/s decode | 22.2 GB VRAM - 80k Context: 1,789 t/s prefill | 84.20 t/s decode | 23.3 GB VRAM - 110k Context: 1,767 t/s prefill | 83.35 t/s decode | 23.96 GB VRAM (110k context at 83+ tokens a second sitting exactly on the 24GB hardware limit is absolute wizardry). How to compile the PR today: git clone cd llama.cpp git fetch origin pull/27342/head:pr-27342 git switch pr-27342 cmake -B build -DGGML_CUDA=ON && cmake --build build -j Llama.cpp flags for Dflash (110k Context Ceiling): ./build/bin/llama-server -m Qwen3.8-27B-UD-Q4_K_XL.gguf -md Qwen3.8-27B-DFlash2-Q4_K_M.gguf --spec-type draft-dflash --spec-draft-n-max 4 -c 110000 -ngl 99 --port 8080 -ctv q4_0 -ctk q4_0 The fact that the open source community is shipping block diffusion drafters so quickly that run entirely locally on a gaming GPU is unbelievable. If you own a single RTX 3090 or 4090, it is officially time to upgrade to qwen 3.8 27b with dflash 2 and cancel your API subscriptions and let local silicon eat the cloud. This model beats GPT 5.6 Terra, GLM 5.2 DeepSeek V4 Pro, Muse Spark 1.2 and Claude Opus 4.8 on the artificial analysis agentic index (details in the replies) Hugging Face GGUF links (Base + DFlash2) and the full visual VRAM scaling and Dflash2 vs MTP graphs are also in the replies below. are you sticking to native MTP for the 130k context, or sacrificing 20k context to redline your decode speed? How many tokens/sec are you pushing on your current local rig?show more

it took us 3 weeks to get 500k+ views... with a fully AI UGC on instagram and it taught me something huge if you create good content for long enough the algorithm is very simple it works to keep finding the type of user that engaged with your post similar to paid ads but this only works if you stay within niche and post consistently great content so stop seeing 1,000-4000 views as bad it’s simply taking the time to learn we’re being extremely aggressive with our marketing and created 40+ content variations each A/B testing a different part of the video - hook, music, transitions, wallpapers, workout, location, CTA, captions every single component is being tested we had a huge set back with Meta randomly banning 8 of our accounts last week we got the content ready and aim to post 20x per day minimum by 6/10 each day breaking down the data and creating more of what works then running it on paid ads both tiktok spark ads and meta ads marketing is definitely takes more time but at least shipping took just 2 weeks using Rork Max + Opus 4.7 at the time i launched, there was no other possible way to natively code in SwiftUIshow more

I experimented a lot with the prompt until I... landed on this version😎 Only face reference + exact 15-second audio. Generated in Higgsfield AI 🧩 Seedance 2.0. Small tip: when uploading reference audio, cut it precisely to whole seconds (6, 8, 10, 15…) with no leftover tails 🎵 Prompt: A cinematic 15-second music video shot in a dark modern dance studio with smooth grey reflective floor, black walls and horizontal neon light tubes. Continuous dynamic camera movement, mostly medium shots and close-ups, never too wide. The woman is constantly moving, no freezes or static poses. Main subject: young woman with messy medium-length wavy brown hair with bangs partially covering her face, freckles on nose and cheeks, blue-grey eyes, full glossy lips. She wears a tight black fishnet bodysuit. She is singing the entire time with clear, precise lip-sync, mouth actively moving, intense emotional expression. Lyrics: "Ye! Ye! You keep a box of names / In the drawer by your bed / Polaroids and ticket stubs / Stuffed under the red / You never take one thing / You take the whole last spark / Leave a little thumbprint / On every private heart". 0-2s: Tight medium close-up. She leans her upper body back, head tilting, singing passionately with strong lip-sync, hair falling over her face, body arching, one hand sliding across her chest. 2-4s: Camera slowly pushes in and circles. She comes out of the deep arch, torso still bent forward, hands on her thighs, lifting her head and looking straight into camera while singing with aggressive lip-sync. Three male dancers of different appearances (different ethnicities, hair styles and builds) wearing black tank tops and black wide pants are already close around her, moving with her in low tense postures. The other two men are visible at the edges of the frame, approaching. 4-7s: Medium close-up. She drops lower, body still in constant motion, hair swinging, singing intensely with clear mouth movements, sharp head turns, eyes locked on camera. Male dancers stay close, their hands lightly touching her as they move together. 8-12s: Medium shot with slow camera drift. Exactly five male dancers of completely different appearances, all wearing black tank tops and black wide pants, surround her tightly on the floor in a dense, intertwined formation. She is in the center, body still moving, upper body rising and shifting, singing with strong lip-sync. All five men move subtly with her, never static. The men never fully obscure her body. 13-15s: Dynamic medium shot. The five diverse male dancers lift her into the air in a powerful deep backbend. Her body is fully extended and arched, head thrown back, still singing with clear lip-sync. While holding her they gently rock her up and down in time with the beat. The camera starts from a clear side view of her arched body and smoothly transitions to a frontal view of her face. At the end they lower her smoothly onto her feet; she lands and immediately continues singing as the five men stay low on the floor around her. The men never fully obscure her. High fashion dance energy, sweaty skin, sharp timing, continuous fluid motion of the woman, priority on accurate lip-sync in every frame. Rules: no hand morphing, no body distortions, clean stable anatomy, fingers and hands remain consistent and natural throughout the entire video. #AIvideo #AIMusicVideo #AIFilmmakingshow more

AUKUS, Mistakes and Opportunities In 2016, Japan offered Australia... state-of-the-art, diesel-electric, ultra-quiet submarines with the option of local production at the Henderson shipyard. The Australian government rejected the proposal, claiming its goal was always nuclear-powered submarines. Instead, Australia decided to spend roughly A$4-5 billion extending the life of its ageing Collins-class fleet until the 2040s . enough money to have bought seven-eight Japanese Taigei-class submarines outright. If that’s really what the government wanted, the Americans and British certainly sent them the bill for AUKUS. Australia is footing almost the entire cost: A$368 billion over three decades. - The United States receives US$3 billion from Australia to expand its industrial base, build more Virginia-class submarines, and then sells 3–5 second-hand boats back to Canberra. - The United Kingdom receives around £2.4 billion from Australia for design and infrastructure work, shares some development costs, and ends up using the exact same SSN-AUKUS design for its own future fleet at essentially no extra cost. I’m genuinely intrigued by how they managed to sell the Australians on this deal. I’d love to meet and congratulate the American and British negotiators – true sales geniuses. Nuclear submarines must have been a childhood dream of that Australian government; there’s no other explanation. But the problems don’t end there. Just as the Americans have cancelled over 300 programmes and thrown away more than US$200 billion in the last 20 years, the British have serious and very recent issues with their own naval projects. It feels like a structural disease in the Western defence industry. - The Astute programme is more than a decade late, costs have tripled, only 5 of the planned 7 boats have been delivered, and engineering problems keep cropping up. - The Dreadnought class (replacement for the Vanguard ballistic-missile submarines) has ballooned by billions and is now delayed well beyond 2030 because of failures integrating propulsion systems and Trident missiles. - And the crown jewels – the Queen Elizabeth and Prince of Wales aircraft carriers – are operational but chronically short of compatible F-35s and cost a staggering £10 billion in overruns. - The Type 45 destroyers suffered catastrophic electrical failures that left them inoperable for years, and the Type 26 frigate programme has been repeatedly cut back, reflecting completely misplaced priorities. And a programme that is supposed to deliver eight submarines to Australia sometime around 2050–2060 is extremely unlikely to proceed as planned, not only because of budgets and operational complications, but because underwater drones are evolving fast and China is leading that race. The Americans and British have a long naval history, but they are also visionaries who understand perfectly well that the future lies in decentralisation: swarms of UUVs, lithium or solid battery submarines, or even small nuclear-powered ones using micro-reactors. These platforms cost 10–20 % of today’s conventional SSNs to maintain, are lighter, and leave far more internal volume for weapons – meaning smaller, cheaper, and more heavily armed submarines. And what does Australia get left with? Far more than just a submarine partnership with Japan – an entire security ecosystem. By 2026-2028, Japan plans to have the HVPG hypersonic glide vehicles fully operational with ranges up to 2,000 km. Their upgraded Type 12 missiles will reach 1,000–1,500 km and can be launched from ships, aircraft, and land batteries. This is enough to cover and protect the entire Australian coastline for thousands of kilometers. And finally, a 3,000 km-range hypersonic missile is being integrated into the Taigei-class and its successor. That arsenal is far beyond anything currently fielded by any Western nation and only Russia and China have comparable systems.show more

Just finished a one-week trip to China. I've now... "survived" all the major (~20) L2 self-driving and robotaxi vehicles in both the US and China. Some thoughts & observations: ▶️L2 self-driving I tested major brands like $Huawei, $Li, $NIO, $Xpeng, and $Xiaomi. Overall, they exceeded my expectations. The rides were not overly cautious and handled complex situations (yes, road conditions in China are very challenging!) quite well. Nothing compares to $Tsla's approach. I see imitation learning/end-to-end as the only effective approach for self-driving. While Chinese peers perform well on main roads, they struggle on frontage roads due to reliance on high-precision maps and rule-based methods (e.g. cars stopped in the middle of the road where there was no clear white lining). Chinese EVs' self-driving capabilities are far ahead of those from US and EU brands. I doubt any Chinese players can profit from L2 self-driving, not because it’s not useful, but because it’s hard to differentiate, and price wars dominate the market in China. Chinese consumers and regulators seem much more receptive to self-driving. Even with a 5/10 self-driving capability, cars are practically *hands-free(!)* Insurance-wise, for L3+ cars, OEMs bear responsibility for incidents, so OEMs avoid labeling cars as L3+. ▶️Robotaxi I tested major brands like $Didi, and $Bidu. I'd rate equal to $Waymo, and it's ahead of other peers. However, the same issue applies here: user experience is nearly perfect (in Yizhuang, Beijing), but expansion is the real question. Chinese robotaxi companies are very sophisticated. While the rest of the world focuses on technology, Chinese peers treat it as a product, considering unit economics, operations, mass production, etc. Interestingly, most companies expressed a preference NOT to operate fleets themselves. They aim to be asset-light and let fleet managers handle operations. Policy Support: China has a very clear approval process, driven by data (autonomous driving distance, fully driverless distance, intervention rate, passenger ratings, etc.). ▶️Chinese EVs In major cities like Beijing or Shanghai, EV adoption (green license plates vs. gas cars with blue license plates) seems to be 40%+. If 40% of cars on the road are EVs, then EV penetration (defined as the % of new car sales) must already be over 50%. In shopping malls, the ground floor is filled with EV showrooms—easily 10+ brands, many of which are unfamiliar Chinese brands. It appears almost too easy to make an electric car, which is a stark contrast to the US. $Xiaomi, for example, can achieve a 10% gross profit margin in its first year of operation, compared to $RIVN's -45%. Additionally, $Xiaomi cars are priced at 30% of $RIVN's price. It's fascinating to see how China transitioned from "couldn't make their own gas cars at all (only JVs)" to "dominating EVs globally." The government deserves credit for setting the direction and executing effectively. China now controls the entire supply chain, with $CATL holding 40% of the global market share. 🔹How did it happen? The success of the industry Incentives were set just right: the government provided incentives early on to make EVs and gas cars have comparable MSRPs, allowing consumers to choose based on functionality. This approach differs from how the IRA offers incentives... Perfectly competitive market: $TSLA was brought in, and competition was welcomed, unlike the US, which has a 100% import tax on Chinese EVs. Strategic regulations: License plate restrictions were used effectively; for example, taxis and minivans are required to be EVs. 🔹The challenges Despite the success, the industry faces challenges with low-margin companies and struggling stocks. The intense competition shows no sign of ending. Well-funded global OEMs and Chinese state-owned car companies continue to subsidize, leading to new EV brands emerging annually. The natural tendency in China is to race to the bottom. I think this ties back to China's history as the "world’s factory," where manufacturers price products at "cost plus" versus the US and developing countries, which price based on "affordability/value creation." 🔹The wow EV feature >Software features that surprised me the most: - Everything in the car can be voice-controlled. Not just simple tasks like playing music; users can adjust the height of the steering wheel and set the temperature easily. - Self-parking, which $Tsla has yet to release to all FSD users, is already a table stake in China (I'd rate the quality as 10/10). >Other fun hardware features: - Mini fridges in the car - Infotainment systems - IoT: remote access the car/home via cellphone - all connected together - Heads-up displays - UV-protected glass roofs: $Xiaomi took $Tsla's design, but the glass roof of the $Xiaomi car is made of double layers with silver, blocking 99.9% of UV and infrared rays...as a result, heat is no longer a problem inside the carshow more

Not every hero wears a cape. Some just share... their bread. Created with Seedance 2.0 + GPT Image 2 on OpenArt Prompt: TITLE The Piece of Bread REFERENCE Use the provided combined character sheet and storyboard board as the main visual reference. Follow the same woman design, stray cat design, bread, sidewalk, low wall, cloth shoulder bag, water bottle, warm sunset lighting, and emotional story beats. Keep the woman and cat visually consistent in every shot. Do not add extra characters. Do not change the core story. SUBJECTS Woman: A young woman in her early 20s with shoulder-length slightly messy dark brown hair loosely tied back, a soft oval face, gentle expressive anime eyes, and a tired but kind expression. She wears a faded oversized hoodie, loose trousers, worn sneakers, and carries a simple cloth shoulder bag. She appears hungry, humble, compassionate, and resilient. Her acting should remain subtle, natural, and emotional. Cat: A small original stray cat with short charcoal-gray fur, a cream-colored chest and paws, one ear with a tiny notch, a long curved tail, and expressive anime-style eyes. The cat feels timid, hungry, hopeful, innocent, and lovable. It begins cautious and hungry, then gradually becomes trusting and comforted. Bread: One small round bread bun. This is the central story object. The woman breaks it into two pieces and shares one half with the cat. ENVIRONMENT Quiet city sidewalk at sunset. Low concrete wall behind them. Soft blurred road and distant buildings in the background. A simple cloth shoulder bag and a small water bottle placed beside the woman. Warm golden-hour light with long shadows. Peaceful, lonely, emotional atmosphere. STYLE 2D Japanese anime short film with a hand-drawn aesthetic. Clean inked outlines, flat-to-soft cel shading, simplified shadow blocks, no CG or 3D rendering. Soft emotional storytelling. Warm golden-hour lighting with painterly anime backgrounds. Expressive anime eyes with subtle, believable facial acting. Gentle cinematic movement with a traditional anime animation feel. No 3D render. No CGI. No Pixar-style shading. No photorealism. No comedy. No chaos. No copyrighted characters. No text. No subtitles. No logos. No social media UI. No background music—only natural ambient sound effects. CAMERA 16:9 cinematic framing. Use close-ups and medium shots to emphasize emotion. End with one wide cinematic shot. Slow push-ins and gentle cuts. Shallow depth of field. Keep both characters clear and expressive. Avoid fast movement or exaggerated actions. TIMELINE 0:00–0:02 Extreme close-up. The woman slowly lifts a small bread bun toward her mouth. She is about to take a bite. Warm sunset light softly illuminates her face and hands. Her expression shows hunger, exhaustion, and quiet resilience. She pauses just before eating. SFX: quiet street ambience, soft breathing, gentle hand movement. --- 0:02–0:04 Medium shot from the woman's side. A small stray cat sits a few feet away on the sidewalk. The cat gazes at the bread with sad, hopeful eyes. It remains still, timid, and cautious. The woman notices the cat and slowly lowers the bread. SFX: soft cat meow, light breeze, distant city ambience. --- 0:04–0:06 Close-up of the woman's hands. She slowly breaks the bread into two pieces. Tiny crumbs fall gently. The moment feels like an important emotional decision. Her hands pause briefly after splitting the bread. SFX: soft bread tearing, tiny crumbs falling. --- 0:06–0:08 Medium side shot. The woman gently extends one half of the bread toward the cat. The cat looks at the bread, then into the woman's eyes. It is nervous but curious. The woman gives a soft, reassuring smile and keeps her hand perfectly still. SFX: gentle hand movement, cat sniffing, quiet breeze. --- 0:08–0:10 Low close shot near the cat. The cat slowly steps forward. It carefully takes the bread from the woman's hand. The woman remains calm and gentle. The cat begins eating, and its expression gradually softens into trust. SFX: tiny paw steps, soft bite, gentle chewing. --- 0:10–0:12 Medium shot. The woman sits comfortably on the sidewalk. The cat comes closer and sits beside her. She gently strokes the cat's head. The cat leans into her hand and relaxes. The moment feels warm, peaceful, and safe. SFX: soft fur brushing, content cat purring, distant street ambience. --- 0:12–0:14 Close emotional shot. The cat rests its head on the woman's lap. She looks down with a warm, slightly bittersweet smile. She still holds her own half of the bread in her other hand. Both appear comforted, no longer feeling completely alone. SFX: quiet breathing, soft breeze, distant city sounds. --- 0:14–0:15 Wide sunset shot from behind. The woman and the cat sit side by side facing the glowing sunset. Their long shadows stretch across the sidewalk. The cloth shoulder bag and water bottle rest nearby. The final frame feels peaceful, hopeful, and heartwarming. SFX: soft wind, distant street ambience, gentle satisfied cat purr. Storyboard Prompt: Create a single horizontal animation pre-production board for an original emotional 2D anime-style short film titled "The Piece of Bread." The output must be one image only and combine: 1. a character design sheet 2. a hand-drawn storyboard page. IMPORTANT Do not make the woman or cat resemble any reference screenshots or existing characters. Keep the same emotional story concept, but create completely original character designs, unique silhouettes, and distinct facial features. No copyrighted characters or close resemblance to any existing animated films or anime. STYLE Professional anime production board. Hand-drawn storyboard style with loose pencil sketch lines, light gray shading, red panel borders, blue motion arrows, and short handwritten production notes. Rendered with anime-inspired linework featuring clean inked outlines, expressive eyes, and simplified shading blocks. It should look like a real animation studio planning sheet, not a polished final illustration. LAYOUT Clean horizontal 16:9 board divided into two sections. SECTION A: CHARACTER SHEET Show both characters consistently. Woman A young woman in her early 20s with shoulder-length slightly messy dark brown hair tied loosely at the back, a soft oval face, gentle expressive anime eyes, and a tired but kind expression. She wears a faded oversized hoodie, loose trousers, worn sneakers, and carries a simple cloth shoulder bag. She should appear humble, exhausted, compassionate, and resilient. Show: front view side view 3/4 view expressions: hungry, thoughtful, soft smile, emotional pose holding a small bread bun pose offering bread Cat A small original stray cat with short charcoal-gray fur, a cream-colored chest and paws, slightly oversized anime-style eyes, one ear with a tiny notch, and a long curved tail. The cat should feel timid, hungry, hopeful, innocent, and lovable. Show: front view side view 3/4 view expressions: sad, shy, hopeful, happy, trusting sitting pose taking bread pose cuddling beside the woman Include tiny handwritten notes and a few small color swatches. SECTION B: STORYBOARD Create 8 cinematic storyboard panels arranged neatly in a grid. Keep character designs consistent throughout. Each panel should include simple handwritten shot notes and blue arrows indicating motion. STORY BEATS 1. Close-up of the woman about to take a bite from a small bread bun during sunset. 2. Medium shot of the hungry stray cat sitting nearby, staring at the bread with sad, hopeful eyes. 3. Close-up of the woman breaking the bread into two pieces as crumbs fall. 4. Medium shot of the woman offering one piece to the cat. 5. Close shot of the cat cautiously stepping forward and taking the bread. 6. Medium shot of the cat sitting beside the woman while she gently pets it. 7. Emotional close-up of the cat resting its head on the woman's lap. 8. Wide sunset shot from behind, showing the woman and cat sitting together with a cloth bag and water bottle beside them. ENVIRONMENT Quiet city sidewalk beside a low wall, soft urban background, cloth shoulder bag, small water bottle, warm golden sunset lighting, peaceful atmosphere, and an emotional mood. FINAL GOAL Create a dense, clean, production-ready anime pre-production board combining a character design sheet and storyboard in a single 16:9 horizontal image, conveying a heartfelt story of compassion and kindness through completely original character designs.show more

i spent $26,600 on cloud GPU rentals over 14... months before i found a NVIDIA DGX Spark at $2,999 (founder's edition) or $3,999 (shipping price) it paid for itself in 6 weeks i run 200B parameter models locally now and my old cloud provider keeps sending me loyalty discount emails the math on that $26,600 is embarrassing to type out loud $1,900/month for 14 months, H100 instances on a specialist cloud provider, because anything bigger than a 70B model simply would not fit anywhere else i paid the invoices like they were a utility bill and told myself it was just the cost of doing serious AI work it took me over a year to find out it wasn't 14 months, broken down: → months 1-4: $1,400-1,600/month - felt like manageable infrastructure overhead → months 5-9: crept to $1,900-2,100 as i started running DeepSeek-class experiments, costs tracking directly with model size → months 10-12: one agent loop ran for 36 hours against a 130B model while i slept, that month hit $2,400 → month 13: ran the cumulative total for the first time, saw $23,800, felt physically sick → month 14: another $2,800 month while i waited for the hardware to ship the box is the NVIDIA DGX Spark - roughly the footprint of a large mac mini, powered by a GB10 Grace Blackwell chip with 128GB of unified LPDDR5X memory that unified memory is the whole thing an RTX 4090 has 24GB of VRAM, which means a 70B model in full BF16 precision physically does not fit, you're quantizing down or you're renting cloud, those are your options this box loads a 200B parameter model quantized and serves it through vLLM over localhost, same API interface the cloud endpoint used the migration took one line of code - i changed the base URL from the provider's endpoint to 127.0.0.1:8000 and everything just worked electricity to run continuous 200B inference locally comes out to about $12/month the payback arithmetic is almost too clean: $2,999 hardware cost against $1,900/month saved, the box paid for itself before i'd owned it two months what i didn't account for was how completely the cost model changes your behavior when there's no hourly meter running, you greenlight experiments you'd never approve on cloud - agent loops that churn for hours, running 10,000 documents through a reasoning pass at 3am, speculative fine-tuning jobs you'd normally skip because the cost felt unjustifiable i ran more experiments in the first 30 days after the box arrived than in the four months before it the loyalty discount email landed about 8 weeks after i cancelled the cloud subscription 15% off my next three months, valued customer, we'd love to have you back i didn't reply the box was already runningshow more

more seedance 2.5 realism tests: > it always makes... the arms a bit too long > shot selection is pretty insane > audio is a bit sub par, needs more work > in aggregate, can 100% pass as real (read the article below for full guide) prompt: Create an exactly 15-second photorealistic shopping vlog in 16:9: extremely sharp 4K compact-camera detail, natural HDR, accurate skin, realistic motion and perfect identity continuity. No VHS, grain, subtitles, logos, watermarks or graphics. SUBJECT: A 21-year-old American woman with long wavy chestnut hair, light everyday makeup and an expressive, approachable face. Fitted black long-sleeve top, light-blue straight jeans, sneakers and a small black shoulder bag. Keep her face, hair, outfit, bag and jewellery identical through every cut. She talks like a normal friend, never a presenter. STORE: A large modern high-street womenswear shop resembling an upscale European fast-fashion retailer without real branding. Cream walls, black fixtures, mirrors, warm-neutral spotlights, polished floor, fitting rooms, folded denim and dense racks of blazers, tops, dresses, knitwear and trousers. Adult shoppers browse independently. No readable brands, signs, price tags or labels. CAMERA: The same compact camera stays in her hand throughout. Every frame is one-handed arm’s-length selfie, her mirror reflection holding it, or her rotating it outward to show clothes. Never place it down, prop it up, pass it to anyone or use an external angle. Sharp 20–24mm-equivalent wide lens, mild barrel distortion, deep focus, reliable eye autofocus and true 24fps with 1/48-second blur. Stabilization controls impacts but retains footstep bob, wrist sway, breathing, grip shifts, late pans and imperfect reframing. Preserve pores, flyaways, teeth, fabric and accurate reflections without beauty filtering or fake bokeh. Exposure stays face-priority but shifts subtly between mirrored aisles and the fitting-room corridor. Fast wrist turns show slight rolling-shutter skew and natural background blur while her face and garments remain sharply resolved. EDITING IS CRITICAL: Use 12 clips and 11 hard jump cuts. Most cuts remain in the same aisle or beside the same rack, with only a small jump in pose, expression, camera distance, garment position or sentence. Remove mundane pauses and bits of walking. Cut mid-gesture, during camera turns and inside sentences for the fast rhythm of an edited social vlog. Geography and identity stay coherent despite time compression. No fades, transitions, whip effects or cinematic montage. Cuts are not beat-synced; some happen for no narrative reason and simply skip a fraction of an ordinary action. CLIP 1 — 0.0–1.2: Walking into the store in loose selfie framing: “Okay, I’m literally just getting—” CLIP 2 — 1.2–2.3: Jump cut a few steps inside, nearly the same angle but her hair and bag have shifted: “—one top. That’s it.” CLIP 3 — 2.3–3.4: She rotates the camera toward a crowded rack and slides hangers with her free hand. Hangers clack; no dialogue. CLIP 4 — 3.4–4.6: Jump cut beside the same rack. In selfie view she holds a dark-red top against herself: “Wait, this is cute.” CLIP 5 — 4.6–5.5: Cut to a nearby mirror, camera visible in her hand. She checks the top. Her female friend, mostly off-camera, says, “No.” She turns: “Really?” CLIP 6 — 5.5–6.4: Close selfie micro-cut in the same spot. She laughs: “Okay, rude.” CLIP 7 — 6.4–7.6: She rotates toward a blazer rail and pulls out a charcoal jacket. Focus briefly transfers from mirror to fabric. CLIP 8 — 7.6–8.8: Jump cut in the same aisle. The top and jacket now rest over her free arm. Looking down: “Why do I already have four things?” CLIP 9 — 8.8–9.8: Near-identical angle after another micro-cut. Friend: “Because you picked them up.” She smiles: “Fair.” CLIP 10 — 9.8–11.2: Cut to her walking past another rack. She pans across dresses, passes the item she wants, then corrects back. No dialogue. CLIP 11 — 11.2–13.2: Mirror near the fitting rooms. Holding the jacket against herself: “Do I need this?” Friend: “No.” She smiles, ignoring the answer. CLIP 12 — 13.2–15.0: Close walking selfie approaching the fitting rooms, garments over her free arm: “Okay, I’m trying it on.”show more

Everyone's sleeping on image-to-3D AI models. They can make... your app look incredibly unique, with just a little effort. Here's how. This is my calorie tracker, built in a week with nothing but prompting. Just Claude Code + a couple APIs. The visuals are all AI-generated. I'll be sharing the full workflow + all the crazy technical stuff Claude and I did to make this work, so nobody has to struggle through it like me. Deep dive coming soon! Till then, this is the high-level idea: 1. Get a clean image of the food (or whatever your asset is) - In my app, the user describes foods via text, or attaches images (or both) - If text, an LLM extracts the food description and formats it into a specific prompt I tuned for this design, and we generate an image using Z-Image Turbo through fal - If image, we do the same thing but with FLUX.2 [dev] to edit the user image into our reference design - Originally, both used Google Nano Banana, but switching to open models cut costs and latency a ton 2. Gaussian splatting (2D image → 3D model) - I tried various 2D-to-3D options on fal and ended up with TripoSplat as my preferred balance of speed, cost, latency; this turns an image into a 3D model that looks super high quality (link below) - The app displays the 2D image while our backend generates the 3D splat - We "groom" the splat to reduce size and load time by culling low-opacity/scale points 3. Render efficiently on device Originally, it looked great but ran at 10 FPS. Getting to 120 FPS was a crazy journey. TL;DR: - SwiftUI had to go; it forced us to render each asset in independent MTKViews, which wasn't workable - Instead, we composite every dish into one full-bleed CAMetalLayer using MetalSplatter (link below) - We had to make some optimizations within MetalSplatter's code too, to reduce the overhead of sorting points per render Then I added some finishing touches like the subtle rotation and parallax as they move around. I think it turned out pretty cool :) Overall, this took some effort, but we still got it done in less than a day. Hopefully your agent can follow in the footsteps of mine and do it much faster. Keep an eye out for the bigger writeup, which'll give your agent everything it needs. If you have any questions, drop em below!show more

Do you want to own part of a AAA... game? I know, you hear it all the time. “Triple A game”, you go to play it, it’s crap. This is different, and it’s only possible with Sonic (Sonic) speed, transaction cost, and of-course FeeM. A game that includes talent from Kojima, Ubisoft, EA Sports, Gameloft & more with advisors from NVIDIA. A game that you’ll be able to play on mobile, desktop, and then Xbox and PlayStation (yes really)! YES! A PRETTY BIG DEAL! Before I tell you about the sale, let me at least tell you about this game (being a massive gamer nerd, this excited me), so…. Introducing Animera (Search for Animera): • Fast-paced skill-based PvP in the Nubera galaxy • Compete in real-time space battles for real rewards It will be powered with $STRIKE: • Compete2Earn: win matches, earn tokens • Play2Burn: 5% of $STRIKE used in matches gets burned Oh, and with 8.75% of all game revenue will be used to buy & burn $SWPx, so the SwapX (SwapX) community owns a real stake in this AAA title. Absolutely insane. > Now let me tell you about its beta run quickly: • 16K+ beta signups • 500+ players added weekly • 7.5K+ matches already played • Launching to 500K+ mobile users via Nomina Games > How can you own a piece of Animera? June 5th at 2pm EDT the sale will go live on SwapX, it will go in three phases each lasting 12 hours or until sold out: PHASE 1️⃣: xNFT Holders Early access with exclusive perks and bonuses. These are for xNFT holders only you can get these here on paintswap PHASE 2️⃣ Whitelisted Communities These will be whitelisted from Creo Engine, SFA AGC, derp, and GOGLZ | SONIC 🥽💥. PHASE 3️⃣ Public Round Any remaining allocation will open to the public - only if Phases 1 & 2 don’t sell out. > What is the raise? Token Price & Allocation: • Token: $STRIKE • Currency: USDC • Total tokens for sale: 101.75M Unlock structure: • 50% unlocked at TGE • Remaining 50% claimable in 30 days • Raise cap: Max $100,000 per user, capped at $10,000 per xNFT • Purchase window priority: xNFT holders get early access (see above)! Transparency is key: Why I love working with the team is because transparency is crucial, so I’m going to tell you about its tokenomics, seed, and fully diluted valuation here: Token Symbol: STRIKE Total Supply: 370,000,000 Initial FDV: $1.48M Total Raise: $950,160 Total Initial Unlock: 112,947,501 STRIKE Initial Market Cap (excluding liquidity): $303,790 Token Allocation: • Seed Round: 59.2M tokens (16% allocation), with a 1-month cliff and linear vesting over 9 months. • Private Round: 94.35M tokens (25.5% allocation), with a 1-month cliff and 6-month vesting period. • Crowdsale: 10.75M tokens (2.91% allocation), unlocked 50% at TGE. • xNFT Holders: 10M tokens (2.7% allocation), with a 1-month cliff. • Liquidity: 37M tokens (10% allocation), with no lock or vesting. • Team: 18.5M tokens (5% allocation), with a 6-month cliff and 12-month vesting. • Rewards: 28.6M tokens (8% allocation), vested over 18 months. • Product Growth: 19.6M tokens (5.3% allocation), vested over 24 months. Token Offering: • Seed Round: Priced at $0.0033 per token, raising $195,360 by selling 59.2M tokens. 10% unlocks at TGE, with a 1-month cliff and 9-month vesting. The initial market cap from seed unlock is $234,127. • Private Round: Priced at $0.0037 per token, raising $349,095 for 94.35M tokens. 15% unlocks at TGE, with a 1-month cliff and 6-month vesting. Initial market cap contribution is $262,508. • Crowdsale: Priced at $0.0040 per token, raising $407,000 by selling 10.75M tokens. 50% unlocks at TGE, with no cliff or vesting. Adds $283,790 to the initial market cap. It’s important you had the full information at hand so you can decide whether or not you’d like to participate. I will be, because it’s a low FDV and it looks great. This is not financial advice, I’m helping the team out. Below is real gameplay: Further details: 👇show more

She doesn't just ink skin, she ink's constellations into... flesh. Watch a phoenix shatter into stardust across his back, one needle stroke at a time. From stencil to soul, this is how myths get made. Made with Seedance 2.0 on BudgetPixel AI Prompt: VIDEO PROMPT — 15 SECONDS (1-second cut scenes) Style: Dark painterly cinematic animation, Arcane-inspired, rich shadows, expressive faces, sharp directional lighting, moody tattoo-studio atmosphere, textured brushwork, high-contrast color grading, realistic anatomy, professional concept-art quality. Character note: The tattoo artist shares only the facial likeness of the reference image "Jessy" — same face shape, eyes, and features. Her hairstyle is changed to a sleek, short asymmetric black bob with a shaved undercut side, and her outfit is changed to a fitted black technical vest over a long-sleeve compression top, with fingerless tactical gloves swapped for tight black latex gloves once the session begins. New tattoo subject: Instead of a samurai, the design is a cosmic phoenix made of shattering constellations — black linework forming a phoenix mid-transformation, its wingtips dissolving into scattered stars and fine red ember-like accents, giving it a mythic, otherworldly quality rather than a standard warrior motif. Second 1: Wide establishing shot — neon-lit tattoo studio at night, rain streaking the window; the artist (Jessy-faced, black bob, black vest) steps into frame as the huge bald muscular man lies face down on the table. Second 2: Close-up — she snaps on black latex gloves, quick confident motion, light catching the vest's fabric. Second 3: Close-up — ink caps in a row, deep black and a single vivid ember-red poured with precision. Second 4: Medium shot — she holds up a glowing tablet/reference showing the cosmic phoenix design, studying it against his back. Second 5: Over-the-shoulder — she lays the stencil across his broad back, aligning the phoenix wingspan to his shoulder blades. Second 6: Close-up — gloved hands smooth the stencil down, faint static crackle effect where it presses to skin. Second 7: Medium shot — stencil peels away, revealing the full black-line phoenix-and-shattering-stars outline across his back. Second 8: Close-up — the tattoo machine powers on, needle glinting, her focused eyes reflected in the metal. Second 9: Extreme close-up — first bold black lines forming the phoenix's wing feathers. Second 10: Side angle — the man stays still, muscles tense under lamplight, she works with total concentration. Second 11: Close-up — ember-red ink added to the phoenix's core, glowing against the black linework like embers catching fire. Second 12: Overhead shot — sparks of red bleed outward into the "shattering stars," visual rhythm of the needle in motion. Second 13: Close-up on her face — a flicker of satisfaction, sweat and studio light on her skin. Second 14: Wide dramatic shot — she leans back, revealing the half-finished phoenix glowing under directional light, black smoke/mist styling in the air. Second 15: Final hero shot — slow push-in on the tattoo, black and ember-red phoenix mid-transformation across his back, studio neon reflected in the wet floor, freeze on a cinematic beat.show more

The upcoming weeks could be a period that requires... close attention in terms of the geopolitical direction of East Africa. The overall picture in Addis Ababa indicates that the region is passing through a critical threshold. We are in Addis Ababa as part of President Erdoğan’s visit to Ethiopia. 1. This year marks the 100th anniversary of Ethiopia–Türkiye relations. Türkiye’s first embassy in Sub-Saharan Africa was opened in Addis Ababa in 1926. The main theme of the visit, consistent with this historical continuity, is a “Permanent Partnership.” 2. Ethiopia occupies an extremely sensitive position within the conflict-prone and competitive environment of the Horn of Africa. The leadership seeks to consolidate its position on issues it considers vital. At the same time, it tends to turn any emerging opportunity into a strategic gain. 3. However, growing demographic pressure, economic challenges, the fatigue of internal conflict, and other domestic vulnerabilities-combined with the structural disadvantages of being a landlocked country-are pushing Addis Ababa toward alternative strategic options. 4. In this context, the strategy of “access to the sea” is seen as a form of strategic relief-both for managing internal pressures and deepening regional influence. Creating economic breathing space, reducing logistical dependence, and increasing strategic maneuverability are the core motivations behind this approach. 5. This places Ethiopia in a difficult equation: it must produce policy within a competitive-and at times confrontational-regional environment where its ambitions may clash with the sovereignty sensitivities of its neighbors. The search for maritime access therefore stands out as an issue that could easily harden the region’s fragile balance. 6. Its closer alignment with the UAE in the Sudan civil war reflects the cumulative pressures of multiple security and economic challenges. Recent developments suggest that Ethiopia may seek a more active and visible role in the Sudan file in the near term. 7. At the same time, military movements along the Eritrean border carry risks that could trigger a new confrontation. For Ethiopia, the Eritrea issue is directly linked to its most critical maritime option. Access to the sea through Assab remains central to its strategic thinking. Historical grievances and current security concerns are deeply intertwined in this file. 8. Another critical relationship is with Egypt. The core issue here is the Nile Basin and water security. While the GERD dam represents energy production, development, and strategic leverage for Ethiopia, it is perceived by Egypt as a high-risk scenario. What began as a technical water management issue has evolved into a geopolitical matter directly tied to regional power balances. 9. When regional dynamics and the engagement of external actors are considered together, it becomes clear that Ethiopia is navigating a multilayered but high-risk foreign policy landscape-one that affects fault lines across North Africa, the Middle East, the Gulf, and the Horn of Africa. 10. Within this broader tension environment, Türkiye-Ethiopia relations remain in a positive trajectory. Despite differences in certain areas, the potential for deeper cooperation is high. Today, both leaders delivered strong messages in this direction. 11. President Erdoğan’s visit comes at a time when regional conflict dynamics are becoming increasingly visible. Ankara is seeking to develop a balancing approach aimed at preventing potential conflicts and lowering rising regional tensions. Given that Türkiye has both direct and indirect interests tied to the nature of its relationship with Ethiopia, the process is being managed with careful strategic consideration. 12. From arrival to departure, President Erdoğan was warmly received in Addis Ababa. The key areas of focus included economic cooperation, defense industry partnerships, and critical regional issues. Murat Yeşiltaş SETAshow more