the “chip is the model” makes zero sense except for toys and niche stuff this WILL NOT scale up to Kimi K2.5 level otherwise OpenAI/Anthropic/Nvidia/Google would have jumped on it a long time ago lol x.com/kylecompute/st…
@theahmadosman
-
Raising capital in 2024 for Llama 3, now dead on arrival
By
–
Imagine having raised in 2024 and spent the capital in doing this for llama 3 and putting that out on the market now Dead on arrival.
-
Small AI models in toys are not the next big play
By
–
They're toy models at best When I say that, I mean literal toys, like the one you buy your kids 10 years down the road, small models will be in these toys on similar cheap hardware, but IT IS NOT the next big play
-
Stop pitching hardware startups coupled with models
By
–
Please stop pitching me hardware startups that are tightly coupled with models No, printing model architectures on hardware isn't smart, it's a waste of PCBs and memory GTX 1080s from 10 years ago could run today's models, but a model on a PCB today won't be used in 10 years
-
ChatGPT images auto-labeled as AI are ngmi
By
–
People posting ChatGPT generated images that get automatically labeled as AI Generated are ngmi
-
High-Performance AI Software Won’t Be Vibecoded
By
–
The amount of software that will need to be written for AI in high performance and specialized languages far exceeds your imagination and it WILL NOT be vibecoded
-
Ultimate step-by-step projects roadmap for becoming AI researcher
By
–
The ultimate step-by-step projects roadmap for BECOMING an AI Researcher is now available online to read FOR FREE Covers building – Tokenizers / embeddings
– Positional methods
– Attention / multi-head attention
– Transformer blocks
– Training loops / objectives
– Sampling -
Educational series on getting started with local open source AI
By
–
I got an educational series on how to get started under this tweet for anyone interested in the subject of Local / Opensource AI
-
Small and specialized models will equal out
By
–
As we get closer and closer to Small & Specialized models, it will all equal out (e.g. Qwen 3.5 27B)
-
Buy GPU, run real workloads daily: top move in 2026
By
–
The single highest-leverage move an individual can still make in 2026 is this: Buy a GPU and run real workloads on them every day. – Not rent tokens.
– Not fine-tune on someone else’s cluster.
– Own the silicon + the weights + the serving stack. I’ve been doing this since 2023