Mishaal Rahman on Nostr: In the future, I'm betting a lot of these ML models will be run in VMs, further ...
In the future, I'm betting a lot of these ML models will be run in VMs, further isolating them. In fact, that's one of the things that the Android Virtualization Framework team has pitched to developers - being able to securely run proprietary ML models in a VM. And AOSP has just added flags for assigning TPUs to protected VMs, so we're getting there.
Published at
2024-08-27 16:42:08Event JSON
{
"id": "c5525a1e186310d1d526f0a6ea6bdccf1f52ffcbbe979867d3080c6ff3b4b7d8",
"pubkey": "8f3ac261dc760558cebf3ee6b891e5fe04ca8c55e1bf79ddde22207be097d53d",
"created_at": 1724776928,
"kind": 1,
"tags": [
[
"e",
"3f187a18e45ec2219d4374c390edab103a1dc2f5a111df027302d50d75458e8f",
"wss://relay.mostr.pub",
"reply"
],
[
"proxy",
"https://androiddev.social/users/MishaalRahman/statuses/113034980787933305",
"activitypub"
]
],
"content": "In the future, I'm betting a lot of these ML models will be run in VMs, further isolating them. In fact, that's one of the things that the Android Virtualization Framework team has pitched to developers - being able to securely run proprietary ML models in a VM. And AOSP has just added flags for assigning TPUs to protected VMs, so we're getting there.",
"sig": "954244deab900592f11a98c407809cff104108562bb39809fb13d200a1a0ee9f3b77c718168fc6882a7f4c8a83529aa7ac8b48847e07706950ad53afc64a7186"
}
