This is pretty remarkable. The "AI is fake" crowd is losing, and fast.

TRUMP ON IRAN: “If they had a nuclear weapon, I'd be calling the Supreme Leader and I'd be saying, ‘Mr. Supreme Leader, how are you, sir? Is there anything we can do for you?' as opposed to bombing the crap out of him." “It's very simple. We cannot let them have a nuclear weapon."

BREAKING: Trump has announced the US will substantially scale back military exercises with South Korea, citing his “very good relationship” with North Korean leader Kim Jong Un.

Kind of insane that AI will solve the Riemann Hypothesis before desktop printers become usable by a normal person

human i/o is costly, so after listening to @mattpocockuk and @trq212 i made an /align-me modification which allows for batches of questions instead of round-by-round. same intuition as spec decoding, you speed up by looking ahead 2-10 steps works INCREDIBLY for design explorations

The magic of AFD (Attention FFN Disagg) It can handle the big boi Sol too (over 3T parameters) This is the first commercially available AFD system (using an SRAM-based accelerator for FFN) on the market Blackwells handle the Attention & Prefill, while the Cerebras chip is responsible for Feed Forward Network execution

Introducing GLM-5.3: Built to Code. Ready for Cyber Defense. - Top-tier coding and agentic capabilities, achieved through post-training on the 743B base model - A major leap in cybersecurity, setting a new standard among open models Tech Blog: z.ai/blog/glm-5.3

🚨 GPT-5.6-Sol Ultrafast mode (750 TPS) is here powered by Cerebras! Here's How kernels work on Cerebras Chips cerebras built a chip with 900,000 cores on a single silicon wafer. CSL (Cerebras Software Language) is a Zig-inspired DSL that gives you direct control over the Wafer-Scale Engine. the SDK is publicly documented and the programming model is genuinely the most alien one i've ever looked at. if NKI felt different, CSL takes it much much further. in CUDA you write one thread's perspective and launch millions. in CSL, there are no threads, no warps, no shared memory, and no kernel launches. you write code for individual Processing Elements: 900,000 independent cores arranged in a 2D mesh on a single silicon wafer. each PE has its own 48 KB of private SRAM, its own program counter, and a 5-port router connecting it to its 4 neighbors. that's actually it. no DRAM. no HBM. no cache hierarchy. 48 KB is your entire world per PE, where code and data must both fit. the programming model is dataflow, and this is where it gets a bit more alien 👽 data moves between PEs as 32-bit messages called wavelets, traveling along virtual channels called colors. when a wavelet arrives at a PE on a specific color, it activates a task (a chunk of code bound to that color at compile time). tasks run to completion, then hardware picks the next activated task. tasks cannot call each other. they can only be activated. so instead of "launch N threads," you think of it like: "place code on PEs, define routes, let data flow." the memory model is also very different from GPUs. on an H100 you get 80 GB of HBM shared across all SMs. on the WSE-3, memory is 48 KB per PE, and there are 900,000 of them! this gives you 44 GB total on-chip SRAM at 21 PB/s aggregate bandwidth (vs 3 TB/s on H100). every access is single-cycle. no coalescing needed. no bank conflicts. no cache misses. but also no way to access another PE's memory. all inter-PE communication is explicit wavelet routing through the fabric. take a distributed GEMV for example. you would write a layout file that physically routes wavelets across the mesh. two PEs sit side by side. the left PE computes a partial result and routes it eastward. the right PE receives those wavelets from the west and accumulates. routing is defined at compile time. both operations are asynchronous and activate a task when they finish. you're physically routing data across silicon at 1 clock cycle per hop. cerebras is an extremely technically interesting & alien beast. reports 95-210x speedups over H100 on stencil computations. 3,000 tokens/sec inference on gpt-oss-120B. $10B+ deal with OpenAI for 750 MW of inference infrastructure. very exciting times for alternative accelerators! @gpuemi technical deep dive

Full post: naomibashkansky.com/blog/tel…

OpenAI engineers earlier this month developed an optimization that cut inference costs in half for models it was applied to. After the optimization was applied to logged-out ChatGPT traffic, it reduced the number of GPUs needed to power that traffic to a couple hundred.