Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

We raced GLM-5.3 Flash on Wally against Z.ai 's own paid fast tier, FlashX. Same prompt, live. Wally: 363 tok/s. FlashX: 143 tok/s. 2.5x faster, 3.5x cheaper. Why are you still using slow AI? Try Wally, $5 free.

RunAnywhere

2,247 subscribers

16,410 views • 16 days ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

10 Comments

RunAnywhere16 days ago

Feel the speed yourself, Drop this into your favorite coding agent to get started: "set up

Chon Tang16 days ago

@Zai_org What the heck!? How is this possible? Will have my agent check it for reasoning levels but if true - you got my business.

RunAnywhere16 days ago

@Zai_org Reasoning is max by default, DMs are open for any feedback!

avhq16 days ago

@Zai_org flashX still came up with the better result

luis16 days ago

@Zai_org @Presidentlin

Phuc Nguyen15 days ago

@Zai_org How many users are you guys serving?

Bali as a Colony of Jakarta16 days ago

@Zai_org lol only for stupid people. it's fast because you serve text-model only without even native tools & sandbox

Fredo16 days ago

@Zai_org wow it would be great to see that same fight against @isoquantai

AlexK15 days ago

@Zai_org It is very fast, but it doesn't work :D

AlexK15 days ago

@Zai_org really??? It was only one agent session and literally few turns.

Related Videos

GLM-5.3 weights from Z.ai are live, with SGLang powering day-0 serving support! GLM-5.3 inherits every optimization and feature we battle-tested for GLM-5.2 over the past months. On real-world multi-turn agentic workloads, we measured 537.6 tok/s/user on NVFP4 and 413 tok/s/user on FP8, at BS=1 with TP8 on 8x B300. It's fast, efficient, and production-ready today on NVIDIA AI Blackwell and Hopper, and AI at AMD MI300X/325X/355X. We believe this is a big step forward for GLM-Series in agentic tasks, with a path toward mythos-class cyber capability. Can't wait to see what people build with GLM-5.3 and SGLang🚀 Cookbook👇

GLM-5.3 weights from Z.ai are live, with SGLang powering day-0 serving support! GLM-5.3 inherits every optimization and feature we battle-tested for GLM-5.2 over the past months. On real-world multi-turn agentic workloads, we measured 537.6 tok/s/user on NVFP4 and 413 tok/s/user on FP8, at BS=1 with TP8 on 8x B300. It's fast, efficient, and production-ready today on NVIDIA AI Blackwell and Hopper, and AI at AMD MI300X/325X/355X. We believe this is a big step forward for GLM-Series in agentic tasks, with a path toward mythos-class cyber capability. Can't wait to see what people build with GLM-5.3 and SGLang🚀 Cookbook👇

SGLang

99,455 views • 1 month ago

GLM-5.3-Flash (320B MoE) serving on Kaggle's free TPU: ~64 tok/s, 262k context. No stack could run this model on a TPU, so I wrote a JAX engine for it: 3-bit experts (from Unsloth AI) in HBM; Pallas kernels decode them inside the matmul. Z.ai Kaggle

GLM-5.3-Flash (320B MoE) serving on Kaggle's free TPU: ~64 tok/s, 262k context. No stack could run this model on a TPU, so I wrote a JAX engine for it: 3-bit experts (from Unsloth AI) in HBM; Pallas kernels decode them inside the matmul. Z.ai Kaggle

Abdur Rahim

109,018 views • 26 days ago

Qwen3.8-Flash-Next just got a massive speed boost on my old hardware. This guy started with roughly 11 tok/s. After tuning the inference stack, I’m now getting around 24–28 tok/s. That’s roughly a 2× speedup without changing the GPU. The setup: → RTX 3090 24GB → 64GB DDR4 system RAM → Qwen3.8-Flash-Next → Cafe-llama.cpp The interesting part is that Cafe-llama.cpp isn’t just another llama.cpp fork. The optimizations are making a real difference for this particular architecture, especially when you’re dealing with a model that relies heavily on system memory and offloading. Going from 11 tok/s to nearly 30 tok/s completely changes how practical the model feels for local use. And now I’m curious what happens with DDR5. If memory bandwidth becomes another bottleneck, a faster RAM subsystem could squeeze even more performance out of the same 3090. This is why local AI optimization is getting so interesting. Sometimes you don’t need a new GPU. You just need a better inference stack. 11 tok/s → 24–28 tok/s.

Qwen3.8-Flash-Next just got a massive speed boost on my old hardware. This guy started with roughly 11 tok/s. After tuning the inference stack, I’m now getting around 24–28 tok/s. That’s roughly a 2× speedup without changing the GPU. The setup: → RTX 3090 24GB → 64GB DDR4 system RAM → Qwen3.8-Flash-Next → Cafe-llama.cpp The interesting part is that Cafe-llama.cpp isn’t just another llama.cpp fork. The optimizations are making a real difference for this particular architecture, especially when you’re dealing with a model that relies heavily on system memory and offloading. Going from 11 tok/s to nearly 30 tok/s completely changes how practical the model feels for local use. And now I’m curious what happens with DDR5. If memory bandwidth becomes another bottleneck, a faster RAM subsystem could squeeze even more performance out of the same 3090. This is why local AI optimization is getting so interesting. Sometimes you don’t need a new GPU. You just need a better inference stack. 11 tok/s → 24–28 tok/s.

FHILY👑

23,243 views • 1 month ago

🤯 A localmaxxer hit ~381 tok/s on a SINGLE RTX 3090 with Qwen3.8-27B. This developer has turned a 24GB RTX 3090 into a monster Qwen inference box - w/some creativity. Four days ago 👉 ⚡ ~82 tok/s single-user Then 👉 ⚡ ~114 tok/s with optimized MTP ⚡ ~138 tok/s with DFlash2 + lookup drafting Now 👉 🔥 ~381 tok/s on ONE request How? The recipe combines ... 🧠 Qwen3.8-27B 🎮 1× RTX 3090 24GB @ 250W ⚙️ heavily optimized vLLM ⚡ DFlash2 speculative decoding 🔎 lookup-augmented drafting 📚 prefix caching 🧮 16-token verification blocks 💾 quantized KV / heads / activations DFlash2 normally proposes 7 tokens. The developer realized the verification block doesn't have to stop there. If Qwen is answering from a document already sitting in the prompt, the system can fill the remaining draft positions using tokens found directly in that context. 🎯 So the target model can verify 16 tokens at once. On a ~25K-token document reproduction task: Previous DFlash2 👉 ~260 tok/s Longer verification + context lookup 👉 🔥 ~382 tok/s Acceptance: 🤯 15 of 16 tokens per verification step That is where the crazy number comes from. ⚠️ On ordinary real-world chat prompts, the same setup is around ~133 tok/s Still extremely fast for a dense 27B model on an RTX 3090. The 381 tok/s mode shines when the answer largely comes from material already in context so these are best use cases 📚 RAG / document Q&A 💻 Coding assistants applying edits 📝 Quoting or rewriting documents 🔎 Extracting information from long prompts And another optimization 👉 With prefix caching, a second question against the same 25K-token document reportedly goes from: 🐌 22.4 sec TTFT → ⚡ 0.56 sec TTFT Because the model doesn't need to process the whole document again. 🎯 It's specifically a mode for RAG front ends and coding agents. Follow iamMess on Reddit or syv-ai on GitHub 🔗 Reddit: r/LocalLLaMA/comments/1vtup5s/ 🔗 GitHub: /syv-ai/qwen38-27b-rtx3090

🤯 A localmaxxer hit ~381 tok/s on a SINGLE RTX 3090 with Qwen3.8-27B. This developer has turned a 24GB RTX 3090 into a monster Qwen inference box - w/some creativity. Four days ago 👉 ⚡ ~82 tok/s single-user Then 👉 ⚡ ~114 tok/s with optimized MTP ⚡ ~138 tok/s with DFlash2 + lookup drafting Now 👉 🔥 ~381 tok/s on ONE request How? The recipe combines ... 🧠 Qwen3.8-27B 🎮 1× RTX 3090 24GB @ 250W ⚙️ heavily optimized vLLM ⚡ DFlash2 speculative decoding 🔎 lookup-augmented drafting 📚 prefix caching 🧮 16-token verification blocks 💾 quantized KV / heads / activations DFlash2 normally proposes 7 tokens. The developer realized the verification block doesn't have to stop there. If Qwen is answering from a document already sitting in the prompt, the system can fill the remaining draft positions using tokens found directly in that context. 🎯 So the target model can verify 16 tokens at once. On a ~25K-token document reproduction task: Previous DFlash2 👉 ~260 tok/s Longer verification + context lookup 👉 🔥 ~382 tok/s Acceptance: 🤯 15 of 16 tokens per verification step That is where the crazy number comes from. ⚠️ On ordinary real-world chat prompts, the same setup is around ~133 tok/s Still extremely fast for a dense 27B model on an RTX 3090. The 381 tok/s mode shines when the answer largely comes from material already in context so these are best use cases 📚 RAG / document Q&A 💻 Coding assistants applying edits 📝 Quoting or rewriting documents 🔎 Extracting information from long prompts And another optimization 👉 With prefix caching, a second question against the same 25K-token document reportedly goes from: 🐌 22.4 sec TTFT → ⚡ 0.56 sec TTFT Because the model doesn't need to process the whole document again. 🎯 It's specifically a mode for RAG front ends and coding agents. Follow iamMess on Reddit or syv-ai on GitHub 🔗 Reddit: r/LocalLLaMA/comments/1vtup5s/ 🔗 GitHub: /syv-ai/qwen38-27b-rtx3090

David Hendrickson

270,913 views • 1 month ago

New LFM2.5-2.6B hits DeepSeek-V4 level on tool calling and runs 3.7x faster! We ran Liquid AI 's new LFM2.5-2.6B against DeepSeek-V4-Flash on one box with 4x RTX 5090. Both got the same three jobs, and each one only completes if the model fires every tool call Topics: -weather and local time in six cities -one budget into six currencies -four hotels checked and booked for one date Outputs: LFM2.5-2.6B: 35/35 tool-calls, 366 tok/s, 19s DeepSeek-V4-Flash: 35/35 tool-calls, 77 tok/s, 70s Both models made all 35 calls, and both fired twelve of them in one turn. LFM was trained for agents, and tool calling is where that shows. Strong result for a model you can run right on your phone

New LFM2.5-2.6B hits DeepSeek-V4 level on tool calling and runs 3.7x faster! We ran Liquid AI 's new LFM2.5-2.6B against DeepSeek-V4-Flash on one box with 4x RTX 5090. Both got the same three jobs, and each one only completes if the model fires every tool call Topics: -weather and local time in six cities -one budget into six currencies -four hotels checked and booked for one date Outputs: LFM2.5-2.6B: 35/35 tool-calls, 366 tok/s, 19s DeepSeek-V4-Flash: 35/35 tool-calls, 77 tok/s, 70s Both models made all 35 calls, and both fired twelve of them in one turn. LFM was trained for agents, and tool calling is where that shows. Strong result for a model you can run right on your phone

atomic.chat

113,185 views • 2 months ago

Qwen3.8 Flash Next is starting to look ridiculous on Apple Silicon. I’m running the 4-bit MTP build locally on an M3 Ultra Studio, and the latest OMP run hit: 97.1 tok/s decode 23K context ~1,131 tok/s uncached prompt processing That first number is the one that caught my attention. Nearly 100 tokens per second from a local Qwen3.8 Flash Next setup is already fast enough that the usual “local models are slow” argument starts feeling pretty outdated. And the prompt processing speed is even crazier. Over 1,100 tok/s on an uncached prompt means the model can chew through a large amount of context before generation even starts. The setup matters here. This isn’t just Qwen3.8 Flash Next running untouched. It’s a 4-bit quantized build with MTP, and the inference stack is clearly doing a lot of work behind the scenes to make the hardware perform like this. There’s already a PR open for the implementation on oMLX, so this isn’t just a one-off local experiment either. If these optimizations make their way into the broader MLX ecosystem, running large models locally on Apple Silicon gets even more interesting. The other thing I like about numbers like this is that they put the focus back on the entire inference stack. Model size is one variable. Quantization is another. Then you have MTP, KV cache configuration, runtime optimizations and the hardware itself. Change the recipe and the same model can feel completely different. Qwen3.8 Flash Next at ~97 tok/s on an M3 Ultra is a pretty good demonstration of that.

Qwen3.8 Flash Next is starting to look ridiculous on Apple Silicon. I’m running the 4-bit MTP build locally on an M3 Ultra Studio, and the latest OMP run hit: 97.1 tok/s decode 23K context ~1,131 tok/s uncached prompt processing That first number is the one that caught my attention. Nearly 100 tokens per second from a local Qwen3.8 Flash Next setup is already fast enough that the usual “local models are slow” argument starts feeling pretty outdated. And the prompt processing speed is even crazier. Over 1,100 tok/s on an uncached prompt means the model can chew through a large amount of context before generation even starts. The setup matters here. This isn’t just Qwen3.8 Flash Next running untouched. It’s a 4-bit quantized build with MTP, and the inference stack is clearly doing a lot of work behind the scenes to make the hardware perform like this. There’s already a PR open for the implementation on oMLX, so this isn’t just a one-off local experiment either. If these optimizations make their way into the broader MLX ecosystem, running large models locally on Apple Silicon gets even more interesting. The other thing I like about numbers like this is that they put the focus back on the entire inference stack. Model size is one variable. Quantization is another. Then you have MTP, KV cache configuration, runtime optimizations and the hardware itself. Change the recipe and the same model can feel completely different. Qwen3.8 Flash Next at ~97 tok/s on an M3 Ultra is a pretty good demonstration of that.

FHILY👑

15,378 views • 28 days ago

Google's Gemma 4 26B A4B QAT hits 25+ tokens/sec and 320+ tokens/sec prefill on 8 GB VRAM (RTX 4060) + 16 GB RAM using TurboQuant Prefill just went from 200 → 320+ tok/s on the same 8GB card. 1.6x, no new hardware, no new quant, just a KV cache trick stacked on top of the Gemma 4 26B MoE setup from a few days ago. A few days ago I posted Gemma 4 26B A4B hitting 28 tok/s decode on 8GB VRAM using native MTP. prefill was stuck around 200 tok/s. fair callout by the community. So today I tested something I'd already been meaning to try: TheTom/llama-cpp-turboquant, the TurboQuant KV cache fork by Tom Turney (Tom Turney). (github link in the comments) thanks to him, the fork just got resynced to mainline, so MTP + TurboQuant now run together cleanly (I didnt see any meaningful gains by using MTP with this setup though but you can try). The flags (No MTP): -m gemma-4-26B-A4B-it-qat-UD-Q4_K_XL.gguf -cnv -c 64000 --cache-type-k q8_0 --cache-type-v turbo3 Results on the same RTX 4060 8GB, tested with a 27k token prompt at 64k context loaded: Prefill: 200 tok/s → 320+ tok/s Decode: stayed above 25 tok/s (without MTP) Why it works: TurboQuant uses walsh hadamard rotation + polar quantization on the KV cache. keys are sensitive to compression, values aren't much, so it splits the difference: K stays at q8_0, V drops to turbo3 (~3 bits). bonus from the memory savings: same 8GB card can now stretch to 100-120k context with minimal decode penalty. It should now be snappier with any agent harness such as hermes agent without compromise on intelligence. If you're already running Gemma 4 on a small card, this stacks on top for free. Try --cache-type-k q8_0 --cache-type-v turbo3 on your setup and report back what your prefill/decode split looks like. unsloth model gguf and llama.cpp turboquant fork links in the comments. what's your prefill number before vs after?

Google's Gemma 4 26B A4B QAT hits 25+ tokens/sec and 320+ tokens/sec prefill on 8 GB VRAM (RTX 4060) + 16 GB RAM using TurboQuant Prefill just went from 200 → 320+ tok/s on the same 8GB card. 1.6x, no new hardware, no new quant, just a KV cache trick stacked on top of the Gemma 4 26B MoE setup from a few days ago. A few days ago I posted Gemma 4 26B A4B hitting 28 tok/s decode on 8GB VRAM using native MTP. prefill was stuck around 200 tok/s. fair callout by the community. So today I tested something I'd already been meaning to try: TheTom/llama-cpp-turboquant, the TurboQuant KV cache fork by Tom Turney (Tom Turney). (github link in the comments) thanks to him, the fork just got resynced to mainline, so MTP + TurboQuant now run together cleanly (I didnt see any meaningful gains by using MTP with this setup though but you can try). The flags (No MTP): -m gemma-4-26B-A4B-it-qat-UD-Q4_K_XL.gguf -cnv -c 64000 --cache-type-k q8_0 --cache-type-v turbo3 Results on the same RTX 4060 8GB, tested with a 27k token prompt at 64k context loaded: Prefill: 200 tok/s → 320+ tok/s Decode: stayed above 25 tok/s (without MTP) Why it works: TurboQuant uses walsh hadamard rotation + polar quantization on the KV cache. keys are sensitive to compression, values aren't much, so it splits the difference: K stays at q8_0, V drops to turbo3 (~3 bits). bonus from the memory savings: same 8GB card can now stretch to 100-120k context with minimal decode penalty. It should now be snappier with any agent harness such as hermes agent without compromise on intelligence. If you're already running Gemma 4 on a small card, this stacks on top for free. Try --cache-type-k q8_0 --cache-type-v turbo3 on your setup and report back what your prefill/decode split looks like. unsloth model gguf and llama.cpp turboquant fork links in the comments. what's your prefill number before vs after?

Alok

119,821 views • 3 months ago

Darkbloom is the first model provider to support Ternary Bonsai 2 27B -- concentrated intelligence that fits on your phone. Try it now: First 250 users get 100 million free tokens; PrismML's new flagship model: - a ternary compression of Qwen3.8 27B at 2bits per parameter. - 8.5 GB total, 5x smaller than original - keeps 98.2% of the Qwen's FP16 benchmark performance. - 75% cheaper than Qwen 27B. Qwen3.8 27B already operates comparably with Opus 4.6 and 5.6 Luna on certain tasks. This one does it in the memory of a phone. 262K context, image input, Apache 2.0. From our first run on the network, on a single M5 Max with no caching: - 35 tok/s decode at 1K context, - 31 tok/s at 10K, - 19 tok/s at 50K. But we expect more performance gain coming in a few weeks! That's a full 27B reasoning model running comfortably on any Mac. 1,000+ Macs are serving on Darkbloom right now. Go try it out!! Thank you to Babak Hassibi Sahin Lale Omead Pooladzandi Reza Sadri Tushar Bansal Evan Walters and the whole PrismML team. This is exactly the kind of model Darkbloom was built for. You can read the essay by Bonsai on Why Local AI Matters:

Darkbloom is the first model provider to support Ternary Bonsai 2 27B -- concentrated intelligence that fits on your phone. Try it now: First 250 users get 100 million free tokens; PrismML's new flagship model: - a ternary compression of Qwen3.8 27B at 2bits per parameter. - 8.5 GB total, 5x smaller than original - keeps 98.2% of the Qwen's FP16 benchmark performance. - 75% cheaper than Qwen 27B. Qwen3.8 27B already operates comparably with Opus 4.6 and 5.6 Luna on certain tasks. This one does it in the memory of a phone. 262K context, image input, Apache 2.0. From our first run on the network, on a single M5 Max with no caching: - 35 tok/s decode at 1K context, - 31 tok/s at 10K, - 19 tok/s at 50K. But we expect more performance gain coming in a few weeks! That's a full 27B reasoning model running comfortably on any Mac. 1,000+ Macs are serving on Darkbloom right now. Go try it out!! Thank you to Babak Hassibi Sahin Lale Omead Pooladzandi Reza Sadri Tushar Bansal Evan Walters and the whole PrismML team. This is exactly the kind of model Darkbloom was built for. You can read the essay by Bonsai on Why Local AI Matters:

Gajesh

48,051 views • 23 days ago

"which quant should I download?" is a question you may never have to answer again the team Hamster Labs has figured out how to kill it with pMLX. download once at full precision (bf16) and the engine re-fits it to your machine on the fly, based on the job you give it tell it two things: how much context you need, and the slowest speed you'll accept. it reads your mac and picks the quantization plus how many experts stay in ram vs. stream from disk. no need to download a smaller quant or figure out which quant fits your hardware. same model for every use case and the config changes based on what you need I ran this on my own M4 Max and Qwen3.8-Flash-Next splits into two configs. short context, under 32k: full bf16, ~85% of experts in ram, fast at full precision since it fits my ram long context, 64k to 256k: keep ~70% bf16 and stream the rest from SSD at 22 tok/s, or drop to q8 and get 40 tok/s. I pick per task and the model itself never changes. there are many possibilities since I have the ram to spar - if i need speed, Q4 100% resident (74gb ram @ 62 tok/s) - if i need balance, Q8 95% resident (76gb @ 38 tok/s) - if i need accuracy, bf16 70% resident (96gb @ 16 tok/s) given whatever RAM you have (16/32/64/128/256 GB) + your context + your min speed, the engine picks the precision (bf16→q8→q4) and the expert-residency/paging split that fits your box and maximizes quality & speed last thing to optimize is speed. there are so many things we want to power with open models at Hamster and these 180b-300b class models have the potential to play a big role in that

"which quant should I download?" is a question you may never have to answer again the team Hamster Labs has figured out how to kill it with pMLX. download once at full precision (bf16) and the engine re-fits it to your machine on the fly, based on the job you give it tell it two things: how much context you need, and the slowest speed you'll accept. it reads your mac and picks the quantization plus how many experts stay in ram vs. stream from disk. no need to download a smaller quant or figure out which quant fits your hardware. same model for every use case and the config changes based on what you need I ran this on my own M4 Max and Qwen3.8-Flash-Next splits into two configs. short context, under 32k: full bf16, ~85% of experts in ram, fast at full precision since it fits my ram long context, 64k to 256k: keep ~70% bf16 and stream the rest from SSD at 22 tok/s, or drop to q8 and get 40 tok/s. I pick per task and the model itself never changes. there are many possibilities since I have the ram to spar - if i need speed, Q4 100% resident (74gb ram @ 62 tok/s) - if i need balance, Q8 95% resident (76gb @ 38 tok/s) - if i need accuracy, bf16 70% resident (96gb @ 16 tok/s) given whatever RAM you have (16/32/64/128/256 GB) + your context + your min speed, the engine picks the precision (bf16→q8→q4) and the expert-residency/paging split that fits your box and maximizes quality & speed last thing to optimize is speed. there are so many things we want to power with open models at Hamster and these 180b-300b class models have the potential to play a big role in that

Eyal Toledano

10,743 views • 1 month ago

this is what 12gb of vram builds in 2026, absolute magic > rtx 3060 12gb, #1 gpu on steam > bonsai 2 27b + mtp, 5.95 gb of weights > hermes agent, 5 hours, 328k tokens written > 8 js files, 2,368 lines, zero hand written code > 50 tok/s fresh, 22 tok/s average, 125k context watch the full video, 5 hours in 12 minutes of pure dance of a local ai model on rtx 3060 12gb vram, and stay till the end for the full gameplay. this entire game was built by bonsai2, a qwen 3.8 27b dense compressed to ternary, and this small model is punching way above its weight. it built multi file engineering work using hermes agent, sure it's not fast but perfect for overnights and routine work and the quality is insane, and context holding is another best one, it does not lose the thread. i ran PrismML bonsai 1 made from qwen 3.6 27b dense and i built things with it, but this time with the latest base model qwen 3.8 these results are insane, and because i loved building with it a lot i thought many more of you would run it because this gpu exists in almost every home. so i packaged mtp, doubled the speed from 26 tok/s to 50 tok/s, packed the prefill fix in and released it on huggingface, almost 4,000 downloads in 3 days. i'll leave a link below.

this is what 12gb of vram builds in 2026, absolute magic > rtx 3060 12gb, #1 gpu on steam > bonsai 2 27b + mtp, 5.95 gb of weights > hermes agent, 5 hours, 328k tokens written > 8 js files, 2,368 lines, zero hand written code > 50 tok/s fresh, 22 tok/s average, 125k context watch the full video, 5 hours in 12 minutes of pure dance of a local ai model on rtx 3060 12gb vram, and stay till the end for the full gameplay. this entire game was built by bonsai2, a qwen 3.8 27b dense compressed to ternary, and this small model is punching way above its weight. it built multi file engineering work using hermes agent, sure it's not fast but perfect for overnights and routine work and the quality is insane, and context holding is another best one, it does not lose the thread. i ran PrismML bonsai 1 made from qwen 3.6 27b dense and i built things with it, but this time with the latest base model qwen 3.8 these results are insane, and because i loved building with it a lot i thought many more of you would run it because this gpu exists in almost every home. so i packaged mtp, doubled the speed from 26 tok/s to 50 tok/s, packed the prefill fix in and released it on huggingface, almost 4,000 downloads in 3 days. i'll leave a link below.

Sudo su

87,876 views • 18 days ago

overdrive update: download speed depends a lot on the source. some do 50 MB/s, some barely 3 MB/s. overdrive will launch as it is now, with these mixed sources. some providers are just slow, so i can't promise download speeds on those. for later i'm thinking about building our own CDN on Cloudflare R2. it's just an idea for now. zero egress fees, and we could target 50 to 100 MB/s on good connections. a 50 GB game: 3 MB/s → ~4h 38min 50 MB/s → ~17min 100 MB/s → ~8min the catch: we've indexed ~55 TB, and storing that on R2 costs ~$825/month. i don't want overdrive to become a paid service. so the idea is to crowdfund it. $1, $5, whatever you're comfortable with. fully transparent. every invoice public, every expense tracked, no profit. overdrive stays free and open source either way. nobody will ever have to pay. would you chip in? and can the PS5 homebrew scene pull this off? (current sources in this video)

overdrive update: download speed depends a lot on the source. some do 50 MB/s, some barely 3 MB/s. overdrive will launch as it is now, with these mixed sources. some providers are just slow, so i can't promise download speeds on those. for later i'm thinking about building our own CDN on Cloudflare R2. it's just an idea for now. zero egress fees, and we could target 50 to 100 MB/s on good connections. a 50 GB game: 3 MB/s → ~4h 38min 50 MB/s → ~17min 100 MB/s → ~8min the catch: we've indexed ~55 TB, and storing that on R2 costs ~$825/month. i don't want overdrive to become a paid service. so the idea is to crowdfund it. $1, $5, whatever you're comfortable with. fully transparent. every invoice public, every expense tracked, no profit. overdrive stays free and open source either way. nobody will ever have to pay. would you chip in? and can the PS5 homebrew scene pull this off? (current sources in this video)

rafael 🇧🇷

20,075 views • 4 days ago

$A $351 MINI PC IS RUNNING 26-BILLION-PARAMETER AI MODELS AT 20 TOKENS/SEC AND HERMES AGENT ON TOP OF IT This is the Minisforum UM790 Pro. AMD Ryzen 9 7940HS, Radeon 780M iGPU, 48GB DDR5. The BIOS reports the GPU has 4GB of VRAM Here's the part people get wrong. The 780M has no dedicated VRAM at all it borrows from system RAM. Vulkan ignores the BIOS number and reads the full 48GB pool directly That's the whole trick. 21+ GB allocated to model weights on a "4GB" GPU. Because the models are MoE, only 4 billion of the parameters activate per token a fraction of the reads a dense model needs. That's why 20 tok/s works here Gemma 4 26B MoE holds 19.5 tok/s with 196K context. Qwen3.5-35B-A3B holds 20.8. Nemotron Cascade 2 clears 24.8. A dense 31B, by contrast, drops to 4 tok/s it reads the entire model every token, no way around it On top: Hermes Agent, full agentic workflows terminal, file ops, web, 40+ tools against local models only. No API keys. The wall between you and a usable local agent used to be a GPU you couldn't afford. Now it's a BIOS setting most people never check Bookmark this & Try it yourself ↓$

A $351 MINI PC IS RUNNING 26-BILLION-PARAMETER AI MODELS AT 20 TOKENS/SEC AND HERMES AGENT ON TOP OF IT This is the Minisforum UM790 Pro. AMD Ryzen 9 7940HS, Radeon 780M iGPU, 48GB DDR5. The BIOS reports the GPU has 4GB of VRAM Here's the part people get wrong. The 780M has no dedicated VRAM at all it borrows from system RAM. Vulkan ignores the BIOS number and reads the full 48GB pool directly That's the whole trick. 21+ GB allocated to model weights on a "4GB" GPU. Because the models are MoE, only 4 billion of the parameters activate per token a fraction of the reads a dense model needs. That's why 20 tok/s works here Gemma 4 26B MoE holds 19.5 tok/s with 196K context. Qwen3.5-35B-A3B holds 20.8. Nemotron Cascade 2 clears 24.8. A dense 31B, by contrast, drops to 4 tok/s it reads the entire model every token, no way around it On top: Hermes Agent, full agentic workflows terminal, file ops, web, 40+ tools against local models only. No API keys. The wall between you and a usable local agent used to be a GPU you couldn't afford. Now it's a BIOS setting most people never check Bookmark this & Try it yourself ↓

slash1s

35,742 views • 3 months ago

The VRAM barrier is officially dead. I just ran Qwen 3.8 Flash Next (MoE) 125B A6B with a 250,000 context window on a single 24GB RTX 4090. 21 tokens/sec decode. 364 t/s prefill. no mtp. no dflash. no kv cache quantization! We are running datacenter models on consumer hardware. Tested on Ubuntu 22 | CUDA 13.0 | PCIe 4.0 x16 | 110 GB DDR4 System RAM with a continuous 28k prompt across all runs. ### The Benchmarks & Scaling # 1. Hybrid Offload (-ncmoe 40 @ 80k Context) Offloaded 40 expert layers to the GPU, pushing VRAM to the ceiling. ./build/bin/llama-server -m Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf -c 80000 --port 8080 -v --fit off -b 4096 -ub 4096 -ncmoe 40 Prefill: 383.85 t/s | Decode: 22.52 t/s Footprint: 23.85 GB VRAM | 97 GB RAM # 2. Full CPU MoE Offload (-cmoe @ 80k Context) Pinned all 512 expert layers to DDR4 RAM (-cmoe), keeping attention on the 4090. llama.cpp flags: (Same as above, replace -ncmoe 40 with -cmoe) Prefill: 355.72 t/s | Decode: 20.84 t/s Footprint: 11.66 GB VRAM (12GB+ VRAM freed up!) | 110 GB RAM # 3. The 180,000 Context Run Prefill: 357.75 t/s | Decode: 20.98 t/s | VRAM: 15.6 GB | RAM: 110 GB # 4. The 250,000 Context Absolute Ceiling ./build/bin/llama-server -m Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf -c 250000 --port 8080 -v --fit off -b 4096 -ub 4096 -cmoe Prefill: 364.29 t/s | Decode: 20.97 t/s Footprint: 18.3 GB VRAM (Still ~5.7 GB of VRAM headroom!) | 110 GB RAM ### Key Insights: -b 4096 -ub 4096: doubles the prompt ingestion from ~150 to 364+ t/s. -cmoe Free Lunch: Shifting expert layers to DDR4 RAM slashes VRAM from 24GB to 11.6GB with virtually zero decode penalty (22.5 -> 20.9 t/s), enabling the 250k context ceiling. Qwen 3.8 Flash-Next (UD-Q4_K_XL) is a massive 111.4 GB model split across 4 shards. To run this architecture, you must build from the experimental PR branch (#27742) by Daniel Han: git clone && cd llama.cpp git fetch origin pull/27742/head:qwen-next && git checkout qwen-next cmake -B build -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=native -DBUILD_SHARED_LIBS=OFF cmake --build build --config Release -j $(nproc) --target llama-server A single 4090 paired with 100 GB of cheap DDR4 RAM will comfortably serve production grade 125B inference. While Qwen 3.8 27B (dense) still holds the crown for single 3090/4090 rigs, Flash Next proves 125B hybrid models are officially viable on consumer hardware. Hugging Face GGUF link and complete performance telemetry graphs are dropped in the replies below. GLM 5.3 Flash VS Qwen 3.8 Flash Next, which one takes the open weights crown this week?

The VRAM barrier is officially dead. I just ran Qwen 3.8 Flash Next (MoE) 125B A6B with a 250,000 context window on a single 24GB RTX 4090. 21 tokens/sec decode. 364 t/s prefill. no mtp. no dflash. no kv cache quantization! We are running datacenter models on consumer hardware. Tested on Ubuntu 22 | CUDA 13.0 | PCIe 4.0 x16 | 110 GB DDR4 System RAM with a continuous 28k prompt across all runs. ### The Benchmarks & Scaling # 1. Hybrid Offload (-ncmoe 40 @ 80k Context) Offloaded 40 expert layers to the GPU, pushing VRAM to the ceiling. ./build/bin/llama-server -m Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf -c 80000 --port 8080 -v --fit off -b 4096 -ub 4096 -ncmoe 40 Prefill: 383.85 t/s | Decode: 22.52 t/s Footprint: 23.85 GB VRAM | 97 GB RAM # 2. Full CPU MoE Offload (-cmoe @ 80k Context) Pinned all 512 expert layers to DDR4 RAM (-cmoe), keeping attention on the 4090. llama.cpp flags: (Same as above, replace -ncmoe 40 with -cmoe) Prefill: 355.72 t/s | Decode: 20.84 t/s Footprint: 11.66 GB VRAM (12GB+ VRAM freed up!) | 110 GB RAM # 3. The 180,000 Context Run Prefill: 357.75 t/s | Decode: 20.98 t/s | VRAM: 15.6 GB | RAM: 110 GB # 4. The 250,000 Context Absolute Ceiling ./build/bin/llama-server -m Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf -c 250000 --port 8080 -v --fit off -b 4096 -ub 4096 -cmoe Prefill: 364.29 t/s | Decode: 20.97 t/s Footprint: 18.3 GB VRAM (Still ~5.7 GB of VRAM headroom!) | 110 GB RAM ### Key Insights: -b 4096 -ub 4096: doubles the prompt ingestion from ~150 to 364+ t/s. -cmoe Free Lunch: Shifting expert layers to DDR4 RAM slashes VRAM from 24GB to 11.6GB with virtually zero decode penalty (22.5 -> 20.9 t/s), enabling the 250k context ceiling. Qwen 3.8 Flash-Next (UD-Q4_K_XL) is a massive 111.4 GB model split across 4 shards. To run this architecture, you must build from the experimental PR branch (#27742) by Daniel Han: git clone && cd llama.cpp git fetch origin pull/27742/head:qwen-next && git checkout qwen-next cmake -B build -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=native -DBUILD_SHARED_LIBS=OFF cmake --build build --config Release -j $(nproc) --target llama-server A single 4090 paired with 100 GB of cheap DDR4 RAM will comfortably serve production grade 125B inference. While Qwen 3.8 27B (dense) still holds the crown for single 3090/4090 rigs, Flash Next proves 125B hybrid models are officially viable on consumer hardware. Hugging Face GGUF link and complete performance telemetry graphs are dropped in the replies below. GLM 5.3 Flash VS Qwen 3.8 Flash Next, which one takes the open weights crown this week?

Alok

1,033,083 views • 1 month ago

ELON MUSK JUST EXPLAINED WHY HE IGNORES QUARTERLY EARNINGS, AND IT LINES UP WITH TOM LEE MORE THAN IT LOOKS He manages SpaceX on a 5-to-10-year horizon, the Mars spending was disclosed in the S1, and short-term earnings pressure mostly comes from managers whose own pay depends on quick results. Tom Lee's market thesis runs on earnings too, just measured in years, not quarters. He argues profits rising faster than prices is why the S&P is cheaper now than in January, with AI margin expansion still ahead. One ignores the quarter, the other tracks the multi-year trend. Both are pointing in the same direction: the long arc is what counts. I wonder if Tom Lee Thomas (Tom) Lee (not drummer) FundstratDirect.com sees Musk's approach as the same long-term lens he applies to the market.

ELON MUSK JUST EXPLAINED WHY HE IGNORES QUARTERLY EARNINGS, AND IT LINES UP WITH TOM LEE MORE THAN IT LOOKS He manages SpaceX on a 5-to-10-year horizon, the Mars spending was disclosed in the S1, and short-term earnings pressure mostly comes from managers whose own pay depends on quick results. Tom Lee's market thesis runs on earnings too, just measured in years, not quarters. He argues profits rising faster than prices is why the S&P is cheaper now than in January, with AI margin expansion still ahead. One ignores the quarter, the other tracks the multi-year trend. Both are pointing in the same direction: the long arc is what counts. I wonder if Tom Lee Thomas (Tom) Lee (not drummer) FundstratDirect.com sees Musk's approach as the same long-term lens he applies to the market.

Tom Lee Tracker (Not actually Tom)

111,189 views • 2 months ago

the timeline spent two days saying bonsai 2 cannot build. here is 1 hour 24 minutes of it building, one shot, from one paragraph, on an rtx 3060 12gb, sped to 8x so you can watch the whole thing. what you are watching is a 5.9gb ternary compression of qwen 3.8 27b, served through the prismml llama.cpp fork with the full 262k window resident, writing a terminal gpu monitor in one html file. it thinks, it calls a tool mid build, it keeps going until finish. the clock in the corner is real. 8x, one frozen recorder stretch cut, nothing else touched. for anyone new here, the 3060 is the most owned gpu on steam, a five year old card, and this is the kind of card the model was compressed for. for anyone who followed july, bonsai 1 built the wankel engine on an 8gb 3060 ti, this is the sequel one tier up. the numbers, live server, thinking on > 26 tok/s fresh, 22 at 12k, 18 at 35k, 13 tok/s at 77k deep > 262k context resident on 11.7 of 12 gb > 76k tokens, 77 minutes, 149 of 150 w the entire run if you own a 12gb card, this runs on it tonight. serve line, model, fork and the exact prompt in the replies.

the timeline spent two days saying bonsai 2 cannot build. here is 1 hour 24 minutes of it building, one shot, from one paragraph, on an rtx 3060 12gb, sped to 8x so you can watch the whole thing. what you are watching is a 5.9gb ternary compression of qwen 3.8 27b, served through the prismml llama.cpp fork with the full 262k window resident, writing a terminal gpu monitor in one html file. it thinks, it calls a tool mid build, it keeps going until finish. the clock in the corner is real. 8x, one frozen recorder stretch cut, nothing else touched. for anyone new here, the 3060 is the most owned gpu on steam, a five year old card, and this is the kind of card the model was compressed for. for anyone who followed july, bonsai 1 built the wankel engine on an 8gb 3060 ti, this is the sequel one tier up. the numbers, live server, thinking on > 26 tok/s fresh, 22 at 12k, 18 at 35k, 13 tok/s at 77k deep > 262k context resident on 11.7 of 12 gb > 76k tokens, 77 minutes, 149 of 150 w the entire run if you own a 12gb card, this runs on it tonight. serve line, model, fork and the exact prompt in the replies.

Sudo su

41,330 views • 23 days ago

Deepseek V4 Flash 0731 (Q2) - 12 tokens/sec - Single RTX 4090 - 650+ tokens/sec prefill - 250k context - no kv cache quantization! DeepSeek just dropped the official V4 Flash 0731 two days ago with a massive agent capabilities upgrade. The official benchmarks are literally crushing their own V4-Pro-Preview on agentic tasks like Terminal Bench 2.1 and DeepSWE. Unsloth AI said they couldn't wait to bring it to local devices, and they delivered. If you thought my 118B Poolside Laguna S 2.1 MoE run last week on a single GPU was wild, hold onto your hardware. I just successfully ran Unsloth’s brand new 91GB DeepSeek-V4-Flash-0731 (UD-IQ2_M) GGUF entirely locally. And I pushed it to a mind-bending 250,000 context window. The VRAM ceiling is an illusion if you know how to optimize llama.cpp. Here are the benchmarks and the cheat codes to run a local frontier class model yourself. For the hardware and setup, I used a single NVIDIA RTX 4090 (24GB VRAM) hooked up via a PCIe 4 bus, running Ubuntu 22.04 LTS and CUDA 13.0. You don't need a massive enterprise server for this, if you have more than 80 GB of standard DDR4 RAM and a 24GB card like an RTX 3090 or 4090, you can run this exact stack yourself. All benchmarks were run using a massive 28k token prompt to truly stress test the prefill limits. no kv cache quantization THE BENCHMARKS (Scaling Context): # 80k Context (Baseline: -b 2048 -ub 2048): Prefill: 465.43 t/s | Decode: 13.00 t/s | VRAM: 22.87 GB # 80k Context (Optimized: -b 4096 -ub 4096): Prefill: 643.15 t/s | Decode: 12.20 t/s | VRAM: 23.00 GB (Notice how doubling the batch flags spiked my prefill throughput by nearly 200 t/s with almost zero VRAM penalty) # 180k Context (-b 4096 -ub 4096): Prefill: 629.18 t/s | Decode: 11.92 t/s | VRAM: 23.40 GB # 250k Context MAXIMUM (-b 4096 -ub 4096): Prefill: 619.02 t/s | Decode: 11.54 t/s | VRAM: 23.40 GB # THE SECRET SAUCE (Why this works): Unsloth’s UD-IQ2_M quant is ~91GB across 3 files. Since I only have 24GB of VRAM, the PCIe 4 bus and system RAM have to do the heavy lifting. The magic bullet is the --no-mmap flag. By completely bypassing OS disk paging, I forced llama.cpp to load the massive model weights directly into the system RAM upfront. Combined with Flash Attention (-fa on) and exactly 12 CPU threads (--threads 12), I maintained an incredibly stable 11.5+ tokens/sec decode speed even at a quarter million token context. # THE EXACT COMMAND: ./build/bin/llama-server -m /workspace/models/DeepSeek-V4-Flash-0731-UD-IQ2_M-00001-of-00003.gguf -c 250000 -fa on --port 8080 --threads 12 -b 4096 -ub 4096 --no-mmap -v Local conversational and agentic coding AI is fully here. You don’t need an API or an H100 cluster. Qwen 3.8 27b drops next week making the 24GB VRAM tier even more worthwhile. What does your current local AI rig look like, and what's the craziest model you've managed to squeeze into it? Official huggingface GGUF links from Unsloth and performance graphs are dropped in the replies below!

Alok

46,100 views • 2 months ago

Hy4 preview just dropped, so I tested it on WorkBuddy against three other models: DeepSeek V4 Flash, GPT-5.6-Luna, and GLM-5.3... the top 4 models on this week’s OpenRouter coding leaderboard. The best part? They’re all available in WorkBuddy, so you don’t need to download multiple Agent apps. For the test, I used one single prompt to build one of those games you’ve definitely seen in ads: swipe/move through number gates, make calculations, and eventually defeat the boss. The results were pretty interesting 👇 For a task of this difficulty, everyone performed well except DeepSeek. It seemed to finish coding without checking for bugs, and the game couldn’t run properly. GPT-5.6-Luna worked, but felt like an arithmetic multiple-choice quiz. A little boring. GLM-5.3 and Hy4 preview were much more impressive. GLM was slower (video sped up 2.2×) and only had three movement positions, while Hy4 offered continuous horizontal movement and a more natural speed. What surprised me most was Hy4’s game design: besides basic arithmetic, it added weapons that reduce enemy damage or increase stolen power, making the game much more fun. Also worth noting: Tencent’s Sherry quantization compresses the 770B MoE model from ~1.5TB BF16 to 214GB, while keeping performance close to the original... making local deployment much more practical. Hy4 preview is free on WorkBuddy until September 10. Give it a try 👇

Hy4 preview just dropped, so I tested it on WorkBuddy against three other models: DeepSeek V4 Flash, GPT-5.6-Luna, and GLM-5.3... the top 4 models on this week’s OpenRouter coding leaderboard. The best part? They’re all available in WorkBuddy, so you don’t need to download multiple Agent apps. For the test, I used one single prompt to build one of those games you’ve definitely seen in ads: swipe/move through number gates, make calculations, and eventually defeat the boss. The results were pretty interesting 👇 For a task of this difficulty, everyone performed well except DeepSeek. It seemed to finish coding without checking for bugs, and the game couldn’t run properly. GPT-5.6-Luna worked, but felt like an arithmetic multiple-choice quiz. A little boring. GLM-5.3 and Hy4 preview were much more impressive. GLM was slower (video sped up 2.2×) and only had three movement positions, while Hy4 offered continuous horizontal movement and a more natural speed. What surprised me most was Hy4’s game design: besides basic arithmetic, it added weapons that reduce enemy damage or increase stolen power, making the game much more fun. Also worth noting: Tencent’s Sherry quantization compresses the 770B MoE model from ~1.5TB BF16 to 214GB, while keeping performance close to the original... making local deployment much more practical. Hy4 preview is free on WorkBuddy until September 10. Give it a try 👇

Orikan

40,312 views • 1 month ago

STOP PAYING FOR TOKENS ! START GETTING PAID INSTEAD Introducing Straitly, the first AI service that PAYS YOU for using tokens Unlike > Openrouter charges 5.5% service fee > Vercel charges 2.9% payment fee > Anthropic gives you low rate limits > OpenAI only lets you use GPT models Straitly: > PAYS YOU 5% CASHBACK on all closed-source models > 10% CASHBACK on open-source models > High rate limits > 177 AI models > 99.8% reliability > ZDR This means that you can save up to 25% on your token bills compared to using services like OpenRouter or Vercel, while still getting the same reliability and model choices Since our waitlist launch 3 weeks ago we have processed ~150B tokens already, and are growing at 35% week over week We're backed by LeapYear, Sand Hill North, and other incredible angels in our mission to build the utility layer for all future token usage This is just the beginning Today Straitly is live and open for public access Start getting paid today at To celebrate, we are giving out $10k in free AI credits to 400 people Comment below what your are building and what model you use the most, then head over to Straitly and sign up We'll select the winners at the end of September !

STOP PAYING FOR TOKENS ! START GETTING PAID INSTEAD Introducing Straitly, the first AI service that PAYS YOU for using tokens Unlike > Openrouter charges 5.5% service fee > Vercel charges 2.9% payment fee > Anthropic gives you low rate limits > OpenAI only lets you use GPT models Straitly: > PAYS YOU 5% CASHBACK on all closed-source models > 10% CASHBACK on open-source models > High rate limits > 177 AI models > 99.8% reliability > ZDR This means that you can save up to 25% on your token bills compared to using services like OpenRouter or Vercel, while still getting the same reliability and model choices Since our waitlist launch 3 weeks ago we have processed ~150B tokens already, and are growing at 35% week over week We're backed by LeapYear, Sand Hill North, and other incredible angels in our mission to build the utility layer for all future token usage This is just the beginning Today Straitly is live and open for public access Start getting paid today at To celebrate, we are giving out $10k in free AI credits to 400 people Comment below what your are building and what model you use the most, then head over to Straitly and sign up We'll select the winners at the end of September !

Mujtaba

196,124 views • 1 month ago

first test results are in. qwen 3.6 27b dense just banged 10 out of 10 on a single rtx 3090 24gb tier at 40 tok/s. no quant tricks. no fused kernels. just q4_k_m straight cut on llama.cpp. i wrote a particle swarm benchmark this morning, fed it the prompt, and the model autonomously built a 500 particle boids flocking system. velocity driven hue, density based brightness, trail blend rendering, mouse attraction physics, click bursts, drag paint. then it used browser automation to test its own work, found the failing tests, iterated through the code, patched tests.js, and landed all 10 green on its own. i sat there hooked for 8 minutes playing. simple but mesmerizing. mouse trails build beautiful patterns, palette cycles with space, click sends particles flying, drag paints through the swarm. simplicity that hooks you. i'll open source this prompt and the build soon so anyone can reproduce it as their own benchmark. this is the first of 5 single file agent tests i wrote for this model. four more coming. octopus invaders flagship after as final. watch the full video below. see it autonomously build from one prompt. haven't slept well since this model dropped yesterday.

first test results are in. qwen 3.6 27b dense just banged 10 out of 10 on a single rtx 3090 24gb tier at 40 tok/s. no quant tricks. no fused kernels. just q4_k_m straight cut on llama.cpp. i wrote a particle swarm benchmark this morning, fed it the prompt, and the model autonomously built a 500 particle boids flocking system. velocity driven hue, density based brightness, trail blend rendering, mouse attraction physics, click bursts, drag paint. then it used browser automation to test its own work, found the failing tests, iterated through the code, patched tests.js, and landed all 10 green on its own. i sat there hooked for 8 minutes playing. simple but mesmerizing. mouse trails build beautiful patterns, palette cycles with space, click sends particles flying, drag paints through the swarm. simplicity that hooks you. i'll open source this prompt and the build soon so anyone can reproduce it as their own benchmark. this is the first of 5 single file agent tests i wrote for this model. four more coming. octopus invaders flagship after as final. watch the full video below. see it autonomously build from one prompt. haven't slept well since this model dropped yesterday.

Sudo su

133,732 views • 5 months ago

🚨 WARNING: TOMORROW WILL BE THE BIGGEST MARKET CRASH OF 2026!! OpenAI, Anthropic, and xAI just called for an AI slowdown. But China is NOT slowing down. AI stocks now account for 62% of the S&P 500’s market cap. When the market opens on Monday, this won’t be “just a dip.” The biggest AI stock crashes so far have all been triggered by China. In the past, whenever China released a cheaper AI model, AI stocks crashed hard. Investors immediately questioned whether U.S. companies were overspending billions on chips and infrastructure they didn't need. But this time is different. It's MUCH WORSE. The people building the technology are now the ones warning that AI development itself needs to slow down. Anthropic CEO Dario Amodei says AI is improving itself faster than anyone can safely control, and companies need to deliberately slow down. Sam Altman has agreed and committed OpenAI to the same approach. OpenAI's own chief scientist has admitted that no lab has solved AI safety well enough to keep scaling at full speed. Elon Musk is slowing down as well. Meanwhile, China doesn't care and will keep improving its models. And that's where the real risk begins: The five biggest tech companies are pouring roughly $700 BILLION into AI infrastructure this year alone. All of it is based on one massive assumption: AI capabilities will continue accelerating fast enough to justify that spending. But if the people running these AI labs are deliberately slowing development down... The payoff gets pushed years further out. While the spending keeps climbing. And regulators are moving in at the same time. Lawmakers are pushing mandatory safety audits and reporting requirements for the biggest AI labs. Multiple researchers who recently left OpenAI and Anthropic are publicly warning that AI is genuinely dangerous. This creates THREE separate threats hitting the same trade at once: → AI builders are calling for a slowdown → Regulators are closing in → AI spending keeps climbing regardless The AI boom has NEVER faced all three at once. And tomorrow, people will find out what happens when they collide. Pay attention and remember: I publicly called the 2022 stock market crash. Then I publicly called Bitcoin's $126K top in 2025. And I'm warning you now. The next market call will be posted here first. Follow and turn notifications on.

🚨 WARNING: TOMORROW WILL BE THE BIGGEST MARKET CRASH OF 2026!! OpenAI, Anthropic, and xAI just called for an AI slowdown. But China is NOT slowing down. AI stocks now account for 62% of the S&P 500’s market cap. When the market opens on Monday, this won’t be “just a dip.” The biggest AI stock crashes so far have all been triggered by China. In the past, whenever China released a cheaper AI model, AI stocks crashed hard. Investors immediately questioned whether U.S. companies were overspending billions on chips and infrastructure they didn't need. But this time is different. It's MUCH WORSE. The people building the technology are now the ones warning that AI development itself needs to slow down. Anthropic CEO Dario Amodei says AI is improving itself faster than anyone can safely control, and companies need to deliberately slow down. Sam Altman has agreed and committed OpenAI to the same approach. OpenAI's own chief scientist has admitted that no lab has solved AI safety well enough to keep scaling at full speed. Elon Musk is slowing down as well. Meanwhile, China doesn't care and will keep improving its models. And that's where the real risk begins: The five biggest tech companies are pouring roughly $700 BILLION into AI infrastructure this year alone. All of it is based on one massive assumption: AI capabilities will continue accelerating fast enough to justify that spending. But if the people running these AI labs are deliberately slowing development down... The payoff gets pushed years further out. While the spending keeps climbing. And regulators are moving in at the same time. Lawmakers are pushing mandatory safety audits and reporting requirements for the biggest AI labs. Multiple researchers who recently left OpenAI and Anthropic are publicly warning that AI is genuinely dangerous. This creates THREE separate threats hitting the same trade at once: → AI builders are calling for a slowdown → Regulators are closing in → AI spending keeps climbing regardless The AI boom has NEVER faced all three at once. And tomorrow, people will find out what happens when they collide. Pay attention and remember: I publicly called the 2022 stock market crash. Then I publicly called Bitcoin's $126K top in 2025. And I'm warning you now. The next market call will be posted here first. Follow and turn notifications on.

Chiefy

1,761,923 views • 28 days ago