OpenSecret (@opensecretcloud) 's Twitter Profile
OpenSecret

@opensecretcloud

The confidential backend for your app. Try out the first app on our platform: Maple AI @TryMapleAI

ID: 1620555381142605825

linkhttps://opensecret.cloud calendar_today31-01-2023 22:51:42

1,1K Tweet

7,7K Followers

121 Following

Anthony Ronning (@anthonyronning_) 's Twitter Profile Photo

DeepSeek is now on Maple AI - hosted privately in the secure enclave of an H100 like our Llama 3.3 model was. End to end encryption, the most private way to use DeepSeek without hosting it yourself. We've got like 3 more models queued up as well as document features. ๐Ÿš€๐Ÿš€๐Ÿš€

Marks (@marks_ftw) 's Twitter Profile Photo

The combo of DeepSeek and Llama running full confidential is a big unlock. Iโ€™m getting increased productivity this week. The latest was writing the blog post in this thread. DeepSeek and I went back and forth writing and making revisions. Yes, Maple AI helps create its own

Maple AI (@trymapleai) 's Twitter Profile Photo

Transparency is key. Maple AI's open-source code and reproducible builds ensure you can verify our security measures. Trust through verification, not just assurance. If you're a Github user, give us a Star to support the project โญ๏ธ github.com/OpenSecretClouโ€ฆ

Maple AI (@trymapleai) 's Twitter Profile Photo

๐Ÿ“ข Maple AI just got smarter! Announcing Gemma 3 27B for Confidential Use. Our newest model available now for paid users has: - Enhanced math skills - Strong foreign language support - Newer knowledge date All within our encrypted environment. Try it today!

๐Ÿ“ข Maple AI just got smarter! Announcing Gemma 3 27B for Confidential Use.

Our newest model available now for paid users has:
 - Enhanced math skills
 - Strong foreign language support
 - Newer knowledge date

All within our encrypted environment. Try it today!
Anthony Ronning (@anthonyronning_) 's Twitter Profile Photo

This week with Maple AI we launched both DeepSeek and Gemma 3. Both models running in a secure enclave with verifiable E2EE. We're working on: - Image capabilities with Gemma - Document Upload - Qwen/Mistral Models - Audio Anything else you all want?

This week with <a href="/TryMapleAI/">Maple AI</a> we launched both DeepSeek and Gemma 3. Both models running in a secure enclave with verifiable E2EE.

We're working on: 
- Image capabilities with Gemma
- Document Upload
- Qwen/Mistral Models
- Audio

Anything else you all want?
Anthony Ronning (@anthonyronning_) 's Twitter Profile Photo

As an AI startup founder in this day and age, I constantly think of two things: 1. Is the problem I'm solving extremely hard to replicate without the experience/expertise I have? 2. Is the solution just barely functioning enough to still be useful to wider market?

Anthony Ronning (@anthonyronning_) 's Twitter Profile Photo

This is not possible with Maple AI - end to end encryption for all AI inference. In fact, if you get our models to say something particularly edgy, please share the prompts with me! Iโ€™m always on the hunt for better jailbreaks ๐Ÿ˜‚

AllisonSatchmo (@allisonmdstack) 's Twitter Profile Photo

Took me awhile to get through. Had to pause to update my to-read list, download encrypted AI chat Maple AI, and make sure my 2FA was properly backed up. These guys!

Maple AI (@trymapleai) 's Twitter Profile Photo

These two apps run DeepSeek AI. One is registered with China and obligated to share user data with the government while the other is encrypted, confidential, and private. Which would you choose?

These two apps run DeepSeek AI.

One is registered with China and obligated to share user data with the government while the other is encrypted, confidential, and private.

Which would you choose?
Anthony Ronning (@anthonyronning_) 's Twitter Profile Photo

At OpenSecret, our platform handles auth inside of secure enclaves. Not only does it feel good to not rely on other people's auth, but to ensure that nobody, not even us or other app devs can "impersonate" users or hijack/view accounts. Private data requires secure auth.