Extending Sovereign AI Research to System-Level Design (Survey and Framework)radiant.co3 ポイント·投稿者 edogrider·3 か月前·0 コメント
A Template-Driven Approach to Resource Management for AI Computeori.co1 ポイント·投稿者 edogrider·7 か月前·0 コメント
Using a bare-metal primitive to avoid the hypervisor tax on GPUsori.co1 ポイント·投稿者 edogrider·10 か月前·0 コメント
Intelligent bin-packing and fractional sharing (MiG) to optimize GPU utilizationori.co1 ポイント·投稿者 edogrider·10 か月前·0 コメント
A bare-metal-first architecture to address the GPU virtualization taxori.co2 ポイント·投稿者 edogrider·10 か月前·0 コメント
LakeFS adds REST catalog support for data version control on structured datalakefs.io4 ポイント·投稿者 edogrider·昨年·0 コメント
Data version control over multiple heterogenous storage back endslakefs.io2 ポイント·投稿者 edogrider·昨年·0 コメント
edogrider·6 か月前·議論This checks out. I live in Silicon Valley but now work for a startup in London. Same vibe, less attitude.They are not the same. The energy/resources in SF/SV are 5x at a minimum - but London has it going on.
edogrider·2 年前·議論The prompt object feature is really cool and the more you think about it, the more there is to it. You can do almost anything with it. Will be interested to see how people use it.
edogrider·2 年前·議論Why not Minio? Yes, the network will still be the bottleneck but you can use whatever he works best and it is remarkably simple.
edogrider·3 年前·議論Pretty thoughtful post from someone I would have instinctively dismissed as a nutter. Keep it up - it is frankly remarkable.
edogrider·5 年前·議論When Companies Should Go to the Public Cloud + When they Should Returnhttps://blog.min.io/repatriation_wave/
They are not the same. The energy/resources in SF/SV are 5x at a minimum - but London has it going on.