Can someone explain or least provide a link to how maple.ai gets around the trust requirement? How do I know that maple.ai doesn’t keep logs?
Discussion
I’m not sure, but I’m sure nostr:nprofile1qqsgafy9ye4j9p2x8vfmlq6equtpcg4m8ks7v545g0d3f7wwueeq5scprfmhxue69uhhyetvv9ujumn0wd68ycmgv43kktndv5hsz9mhwden5te0wfjkccte9ec8y6tdv9kzumn9wshssdt5n3 can answer
As someone who has multiple paid AI subscriptions for work, genuinely interested, however why only access to mid tier open source models? Why not the full 671b parameter model ? I am guessing if I am using grok 4 they are not limiting to a 70b parameter model.
On another note is there a road map for which other models they are going to add? And do they have basic tools such as web search etc? Do they also allow for api access ?
nostr:npub136jg2fnty2z5vwcnh7p4jpckrs3tk0dpueftgs7mznuuaenjpfps6tjnxf can you explain? is there a technical solution as to why we don’t have to trust maple.ai?
or is the claim just that maple.ai is more trustworthy?
We use confidential computing. Secure enclaves provide cryptographic proofs of the code running on them. A user can build the open source code and compare the fingerprint to the secure enclave.
Lots of details in our tech primer here https://blog.opensecret.cloud/opensecret-technicals/
perfect. thanks for the reply/link.