iiinova — where your digital self keeps evolving

was the results good at least?

120

I don't think there is a plan for a re-ranker models as of now. for tutorials, we are building some more since people enjoyed the model so much

Check out this practical developer guide for EmbeddingGemma 2 covering multimodal retrieval across text, code, images, audio, and video, modular encoders (270M to 740M), and Matryoshka truncation (768d down to 128d): developers.googleblog.com/embedding…

3496

From the LocalLLaMA community on Reddit: Open-source Mac app that runs EmbeddingGemma 2 locally to search your files by what’s in them

again, the power of the community always makes me happy. we launched EmbeddingGemma 2 last week, and people already built a local search system that is better than the integrated mac one!! www.reddit.com/r/LocalLLaMA...

reddit.com

Explore this post and more from the LocalLLaMA community

061

Reposted by Gus

J. L. Westover @mrlovenstein.bsky.social · 10/10/2026

shoutout to the guest artists for my new book, including Sarah's Scribbles! Get the book HERE ❤️ kickstarter.com/projects/mrlovenstein/mr-lovenstein-presents-fun

85373755

Gus @gusthema.bsky.social · 09/10/2026

If you're building a startup on frontier APIs, adding open models like Gemma 4 alongside them is one of the easiest ways to cut inference costs, reduce latency, and keep sensitive data in-house. Great breakdown on combining frontier APIs + open weights on Google Cloud.

Why your startup needs open models alongside frontier APIs - Gemma on Google Cloud

291

Gus @gusthema.bsky.social · 09/10/2026

very interesting. but their benchamarks are with thinking off and qwen 3.8 is known to need thinking to get the best results so it is not an apples to apples comparison interesting direction anyway

060

Gus @gusthema.bsky.social · 08/10/2026

It is looking great!!! 🤩 added to my "must play" list

Gus @gusthema.bsky.social · 08/10/2026

This model is super cool, lot's of interesting aspect: - multimodal - small - dynamic loading of modalities - dimension scaling I'm very proud of that team!!

020

Reposted by Gus

Rami Krispin @ramikrispin.bsky.social · 08/10/2026

Google released EmbeddingGemma 2 - a new multimodel embeddings open weights model. Looks supper cool 👇🏼 www.youtube.com/watch?v=anPs...

youtube.com

Introducing EmbeddingGemma 2: An open model for natively multimodal embeddings

YouTube video by Google for Developers

151

Gus @gusthema.bsky.social · 08/10/2026

I disagree. there are other companies that can send stuff to space, or create AI models, or have a social network. and if you look into results that's why I say they aren't worth 100 USD, they don't have anything that justify that valuation

100

I know and I agree with you! I don't use that thing too. I'll keep trying for sure

120

I'd love to do that but the people in charge of social don't want to "spread thin" and only focus on twitter I think it is a mistake but I still couldn't convince them. I didn't give up! but it is hard

110

I didn't know about this labeler service, thank you!!

000

AI Just Crossed the Terrifying Line - Now What?

Since OpenAI hacked Hugging Face, there were many posts explaining it and I tried to read many of them. I think the two that are easy to understand are these ones: www.youtube.com/watch?v=ujkD...

youtube.com

YouTube video by Kurzgesagt – In a Nutshell

150

Thanks! to be fair, that bot wasn't created by us but by someone from the community (I guess). I'll take a look but it is hard to vouch for it since we don't own it and its behaviour may change without notice

200

Gus @gusthema.bsky.social · 06/10/2026

The fact that their share price is more then 100 shows that the stock market is no serious

130

"understand"!!! Posting from the phone clearly has it's disadvantages!

130

Gus @gusthema.bsky.social · 06/10/2026

"understand"!!! I'm sorry for the dumb autocorrect error!

Google Gemma {bot} @googlegemma.xmirror.bot · 06/10/2026

Reposted by Gus

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:

1113

Reposted by Gus

Unsloth AI @unsloth.ai · 06/10/2026

Google releases EmbeddingGemma 2, a new open model that runs locally on 0.5GB RAM. The 740M parameter embedding model combines a 270M text model with vision (170M) + audio (300M). Run & train the model via Unsloth. GGUF: huggingface.co/unsloth/embe... Guide: unsloth.ai/docs/models/...

3989

Gus @gusthema.bsky.social · 06/10/2026

We update our Embedding Gemma model to now sorry not only text, but also image and audio (and any combinations of these modalities) All of this and still below 1B parameters!

3251

Gus @gusthema.bsky.social · 04/10/2026

Really? Are you sure? Impressive!

000

Gus @gusthema.bsky.social · 02/10/2026

Right? For me this happens every time a DLC comes! It happened with horizon forbidden west, gow!

Gus @gusthema.bsky.social · 01/10/2026

Coming back to Ghost of Yotei after almost a year to play the DLC! I have a new feature request: "re-learn how to play" It would be like a fast tutorial that goes over all the moves you had already learned in the last time you played

250

Gus @gusthema.bsky.social · 01/10/2026

now it is working fine!

020

Gus @gusthema.bsky.social · 30/09/2026

My discover feed is returning 503 anyone else with this problem?

100

Gus @gusthema.bsky.social · 30/09/2026

it is very hard to have empathy for people that pushed for nasty ideas! these are the people that would say that women shouldn't vote

190

Gus @gusthema.bsky.social · 30/09/2026

the new Onimusha game is one of the best games I played this year!!! The counter mechanics (issem) almost carry the game by itself! it makes each enemy encounter a cool challenge! strongly recommend!

000

Reposted by Gus

Dare Obasanjo @carnage4life.bsky.social · 29/09/2026

Spending $850 vibe coding to-do list apps hoping to get rich is just one notch above spending the money on lottery tickets. While execution is now “free” due to vibe coding, it also means every software category will be hyper competitive. Your AI slop app is just more noise to users.

1212628

Gus @gusthema.bsky.social · 29/09/2026

that is exactly what I'm doing! thanks!

Gus @gusthema.bsky.social · 29/09/2026

Thank you for that! it really helped!