Taming the AI Hardware Cambrian Explosion — Abdul Dakkak, Modular
Abdul Dakkak, chief scientist at Modular, explains why today's AI software stack is a mess, with vLLM, SGLang, TensorRT-LLM and llama.cpp each patching around different hardware, and how Modular rebuilt the stack from the ground up.
Proof1 independent outlet
Key points
- Compute is booked out for years and new chips keep arriving.
- Mojo is a Pythonic systems language for CPUs, GPUs and accelerators that aims to match CUDA's performance; Modular has open-sourced its kernels, including matmul and attention kernels he says beat the vendor libraries and FlashAttention 4, with Mojo 1.0 and an open-source Mojo coming.
- MAX is an inference framework that rethinks PyTorch-era assumptions, supports thousands of model architectures behind OpenAI-compatible endpoints, and runs across GPUs, nodes and even mixed hardware.
- • Why the explosion of AI hardware broke the software stack
Sources (1)
- [1]Taming the AI Hardware Cambrian Explosion — Abdul Dakkak, ModularAI Engineer (YouTube) · Oct 11, 09:00 PM
Abdul Dakkak, chief scientist at Modular, explains why today's AI software stack is a mess, with vLLM, SGLang, TensorRT-LLM and llama.cpp each patching around different hardware, and how Modular rebuilt the stack from the ground up.
Compute is booked out for years and new chips keep arriving.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 11, 2026Gemma4 E4B EXL3 - Alternative to Llama.cpp on Jetson Orin
- Oct 11, 2026"Strata" for GLM5.3 Flash is here for some! Project Maya
- Oct 11, 2026openai/openai-python v3.29.0
- Oct 11, 2026Agents That Own Their Inference — Du'an Lightfoot & Khaja Omer, Akamai Technologies
- Oct 10, 2026Satya Nadella says we should assume all AI models are ‘compromised’
- Oct 9, 2026Microsoft's Decision-1 model enters the fast-growing AI decision model race
