"We give you 100% #opensource software and 100% open source models. You can get on Hugging Face, download the model and run it; we make it very easy for you to do that.” -Tenstorrent COO Keith Witek in Analytics India Magzine.
Justin Kinsey
3mo
Openness is the cornerstone of democratizing AI; not only in terms of the tools meant for developers and engineers, but in terms of AI reaching all the way to end users around the world, where it can impact daily life. Watch how companies that are enabling this type of open, transparent ecosystem to take shape will thrive in the years to come 👆
Founder & CEO of Dev Squad Six
3moThe reality is, tech companies that rely on high-performance & cost-effective AI solutions are facing escalating challenges. Despite their strong positions in the technology sector, CTO's of these companies have constantly battled high operational costs and outdated AI technology. If you are using NVIDIA's GPUs, you're dealing with huge energy consumption, hot running machines, and a big operational bill that becomes increasingly hard to ignore. Furthermore, the dependency on expensive High Bandwidth Memory (HBM) and proprietary interconnects like NVLink and Mellanox adds to the financial strain. Aside from these hardware challenges, the software side presents its own set of problems. NVIDIA's CUDA platform has evolved into a tool loathed by development teams. Bloated with features yet unwieldy in practice. Integration is a nightmare, and the software ecosystem is now more like a restraint than a leverage point. Anything that is loathed by development teams (that provides a bad Development experience (DX)) produces slower less quality code - studies show this time and time again. Very impressive how Tenstorrent tackles these pain points directly, solutions are needed now more than ever and it will only get worse.