Engram is in soft launch.
A reader for articles, PDFs and notes. Select a passage and a small model on your device explains it in a note beside the text. Your library is Markdown in a private GitHub repo you own.
Free and open source: https://t.co/9gqjC41Or6
Engram is in soft launch.
A reader for articles, PDFs and notes. Select a passage and a small model on your device explains it in a note beside the text. Your library is Markdown in a private GitHub repo you own.
Free and open source: https://t.co/9gqjC41Or6
@nutlope You just created v2 of thing I am working on slow from scratch for my self after. I might pick you pointers from gloss but i like to keep my things on git so I have bit different view on storage for notes
Would love if you can give feedback
https://t.co/LlrmkJLURj
Introducing EmbeddingGemma 2! 🚀
Our lightweight, multimodal embedding model maps text, code, images, video, and audio into a single, unified embedding space. Optimized for on-device use cases, it features:
- 740M parameter form factor with modular encoders
- Flexible dimension sizes (768dim-128dim) via Matryoshka Representation Learning (MRL)
- 8K context window (4x larger than text-only EmbeddingGemma)
- A commercially permissive Apache 2.0 license
@sarthaksavvy@Google I think they are best in-place to make small models because no buddy understand android better than them. Which is biggest consumer market for ai. If you can crack partial local for sensitive data and cloud for hard thinking that's where sweet spot lies.
What you think?
@DhravyaShah I last used supermemory when it was just launched as a bookmark app. But its been long since then I think should give it a try again and learn more about agent memory, but what you think where it now stands from how it was till now compared to your "industry standards"
@SarvamAI Dubbing was never the problem Google already solved it's long back main problem was getting emotion and the non dictionary dub. Though I am now more on text side. Audio problem is also has gone beyond scope if sarvam dub has really got that feel i would love to tryout.
@dani_avila7 Cache clock is a better mod to have I use to handle it using some custom scripts but now with mods it's more easy and better.
I also have my few custom scripts out as mod would love to know if you like anything
https://t.co/xYkhPMGSXm
When i read and think together I get lost I think now I should keep my mobile away on my job if i download @askalphaxiv new app. Joke aside let's cheer for this new way of exploring papers.
If doomscrolling X is part of your research workflow, we built something for you.
Introducing Paperscrolling, now on iOS and Android 🚀
Follow the latest research with key ideas, figures, and audio explanations from alphaXiv Briefs
Check it out here!
https://t.co/aqQLsaUMvo
If doomscrolling X is part of your research workflow, we built something for you.
Introducing Paperscrolling, now on iOS and Android 🚀
Follow the latest research with key ideas, figures, and audio explanations from alphaXiv Briefs
Check it out here!
https://t.co/aqQLsaUMvo
Meet Mistral Large 4, aka Le Chonk.
• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October.