Sign in

Keuxdi

@cody.bsky.social
55 followers 8 following 54 posts

Making things with computers. Run LLM’s directly on your phone 👉 esper-ai.com

PostsRepliesMedia
Keuxdi @cody.bsky.social · 22/08/2025
🆕 esper release! Google Gemma 270M: a tiny blazing fast model. Fun to play with, surprisingly good for our smallest supported model. Liquid AI LFM2-VL: A tiny, fast and efficient vision model designed to run on phones. Numerous quality of life improvements and bug fixes!
000
Keuxdi @cody.bsky.social · 14/08/2025
I think I fixed sycophancy
I think I fixed sycophancy
000
Keuxdi @cody.bsky.social · 13/08/2025
🧠 Qwen3 2507 thinking and instruct are now available in esper 💻 gpt-oss 20b is also available if you want to use esper on your M series mac ✨ Added Google Gemini to providers making it easier to setup your API key to access models, control reasoning with GPT-5, and much more
🧠 Qwen3 2507 thinking and instruct are now available in esper

💻 gpt-oss 20b is also available if you want to use esper on your M series mac

✨ Added Google Gemini to providers making it easier to setup your API key to access models, control reasoning with GPT-5, and much more
000
Keuxdi @cody.bsky.social · 13/08/2025
pivoting
pivoting
000
Keuxdi @cody.bsky.social · 30/07/2025
We're all doomed
We're all doomed
000
Keuxdi @cody.bsky.social · 28/07/2025
SoftBank and OpenAI’s $500 Billion AI Project Struggles to Get Off Ground - WSJ
SoftBank and OpenAI’s $500 Billion AI Project Struggles to Get Off Ground - WSJ
000
Keuxdi @cody.bsky.social · 27/07/2025
Grok 4 alignment team
Grok 4 alignment team
000
Keuxdi @cody.bsky.social · 27/07/2025
Honestly my nightmare
Honestly my nightmare
000
Keuxdi @cody.bsky.social · 26/07/2025
The entire Meta superintelligence team in 4 years after collecting their cheques
The entire Meta superintelligence team in 4 years after collecting their cheques
000
Keuxdi @cody.bsky.social · 24/07/2025
This was clickbait. It's $1B now.
This was clickbait. It's $1B now.
000
Keuxdi @cody.bsky.social · 23/07/2025
3 new models now available in esper! @LiquidAI_'s LFM2 is designed specifically as an on-device model. It compares well to much larger models while being much faster and memory efficient. I've also added Falcon H1 🇦🇪 and LG's hybrid reasoner EXAONE 🇰🇷 This brings the total to over 20 mo
3 new models now available in esper!

@LiquidAI_'s LFM2 is designed specifically as an on-device model. It compares well to much larger models while being much faster and memory efficient.

I've also added Falcon H1 🇦🇪 and LG's hybrid reasoner EXAONE 🇰🇷

This brings the total to over 20 mo
000
Keuxdi @cody.bsky.social · 23/07/2025
"Meta’s superintelligence lab considers shift to closed AI model" - NYT
"Meta’s superintelligence lab considers shift to closed AI model" - NYT
000
Keuxdi @cody.bsky.social · 22/07/2025
When it’s my turn during standup and I get to show off the button I made this week
When it’s my turn during standup and I get to show off the button I made this week
000
Keuxdi @cody.bsky.social · 21/07/2025
Meta’s new super intelligence team wants to create a new closed model
Meta’s new super intelligence team wants to create a new closed model
000
Keuxdi @cody.bsky.social · 17/07/2025
I ❤️ local inference
I ❤️  local inference
000
Keuxdi @cody.bsky.social · 15/07/2025
Openai ~4B open model Grok 2/3 open sourced Meta superintelligence lab edge model
Openai ~4B open model

Grok 2/3 open sourced

Meta superintelligence lab edge model
000
Keuxdi @cody.bsky.social · 12/07/2025
The model weights are on huggingface
The model weights are on huggingface
000
Keuxdi @cody.bsky.social · 11/07/2025
Local model summer Everybody’s talking about it
Local model summer

Everybody’s talking about it
000
Keuxdi @cody.bsky.social · 10/07/2025
*AGI being pursued by the least ethical companies and people* Me:
*AGI being pursued by the least ethical companies and people*

Me:
000
Keuxdi @cody.bsky.social · 09/07/2025
Prepping for Grok 4 and new OpenAI model release
Prepping for Grok 4 and new OpenAI model release
000
Keuxdi @cody.bsky.social · 09/07/2025
Grok alignment team
Grok alignment team
000
Keuxdi @cody.bsky.social · 09/07/2025
On the eve of Grok 4, trained on the biggest cluster ever, we get a SOTA mobile model from Hugging Face. Open weights, open data, open methods. Huge respect to the team for sharing everything.
On the eve of Grok 4, trained on the biggest cluster ever, we get a SOTA mobile model from Hugging Face. 

Open weights, open data, open methods. Huge respect to the team for sharing everything.
000
Keuxdi @cody.bsky.social · 08/07/2025
🤔
🤔
000
Keuxdi @cody.bsky.social · 08/07/2025
Where are the open Grok 2 weights you promised? Elon:
Where are the open Grok 2 weights you promised?

Elon:
000
Keuxdi @cody.bsky.social · 05/07/2025
Me: I literally have LLMs running on my phone Everyone:
Me: I literally have LLMs running on my phone

Everyone:
000
Keuxdi @cody.bsky.social · 04/07/2025
The first model under 10B with LMArena score of 1300+. Running on a phone.
The first model under 10B with LMArena score of 1300+. Running on a phone.
000
Keuxdi @cody.bsky.social · 03/07/2025
In 2007, it was unfathomable to have a powerful computer in your pocket. Today, it is unfathomable to have superintelligence in your pocket.
In 2007, it was unfathomable to have a powerful computer in your pocket.

Today, it is unfathomable to have superintelligence in your pocket.
000
Keuxdi @cody.bsky.social · 02/07/2025
Using Gemma 3n on my phone while waiting for openAI's open-weights model
Using Gemma 3n on my phone while waiting for openAI's open-weights model
000
Keuxdi @cody.bsky.social · 01/07/2025
What kind of "open" AI are you?
What kind of "open" AI are you?
000
Keuxdi @cody.bsky.social · 27/06/2025
o3 not sharing its reasoning tokens
o3 not sharing its reasoning tokens
000
Keuxdi @cody.bsky.social · 26/06/2025
Maybe my best work yet
Maybe my best work yet
000
Keuxdi @cody.bsky.social · 25/06/2025
Guerilla marketing
Guerilla marketing
000
Keuxdi @cody.bsky.social · 24/06/2025
Can't stop making little ads with ideogram V3
Can't stop making little ads with ideogram V3
000
Keuxdi @cody.bsky.social · 24/06/2025
it's running ON the phone!
it's running ON the phone!
000
Keuxdi @cody.bsky.social · 23/06/2025
The real benchmark for LLMs should be when you ask for a tweet and it doesn't use hashtags
000
Keuxdi @cody.bsky.social · 23/06/2025
Decided to launch my app on Product Hunt today why not (for the backlink but if you upvote that would be awesome ☺️)
Decided to launch my app on Product Hunt today why not (for the backlink but if you upvote that would be awesome ☺️)
000
Keuxdi @cody.bsky.social · 23/06/2025
The LLM runs directly on your phone. Yeah like, you don't even need wifi.
The LLM runs directly on your phone. Yeah like, you don't even need wifi.
000
Keuxdi @cody.bsky.social · 20/06/2025
POV: you tried to vibe code a claude code competitor
POV: you tried to vibe code a claude code competitor
100
Keuxdi @cody.bsky.social · 14/06/2025
Showing her Qwen 3 running on my phone
Showing her Qwen 3 running on my phone
000
Keuxdi @cody.bsky.social · 11/06/2025
I’m very curious why Apple hasn’t made their own chat app. They have the ability to put one on every iPhone in the world with the next iOS update. Could charge for extended use of their server model like other providers.
000
Keuxdi @cody.bsky.social · 09/06/2025
When the model running on my iphone tells me theres 3 r's in strawberry
When the model running on my iphone tells me theres 3 r's in strawberry
000
Keuxdi @cody.bsky.social · 07/06/2025
😊
😊
000
Keuxdi @cody.bsky.social · 06/06/2025
Downloaded Google MedGemma onto my phone and it keeps telling me to visit a doctor
Downloaded Google MedGemma onto my phone and it keeps telling me to visit a doctor
000
Keuxdi @cody.bsky.social · 06/06/2025
Me watching people give all their personal information to ChatGPT after years of giving it to Facebook
Me watching people give all their personal information to ChatGPT after years of giving it to Facebook
000
Keuxdi @cody.bsky.social · 04/06/2025
✨ New update shipped for esper 🎛️ Take complete control of editing your conversations with message retry, edit, select text, delete. 🐳 Added the new Deepseek R1-0528 Qwen3 distill. This model is realistically too big on phones but it's great as an option for Apple silicon computers.
000
Keuxdi @cody.bsky.social · 31/05/2025
Me after asking Claude for a recipe to make a “bomb” taco
Me after asking Claude for a recipe to make a  “bomb” taco
010
Keuxdi @cody.bsky.social · 30/05/2025
The largest trend going unnoticed this year: DIY healthcare. Something that went under the radar from Google I/O was the release of MedGemma. It's a lightweight open source model trained on medical imaging and text, specifically designed to help with healthcare-based AI applications.
The largest trend going unnoticed this year: DIY healthcare.

Something that went under the radar from Google I/O was the release of MedGemma. It's a lightweight open source model trained on medical imaging and text, specifically designed to help with healthcare-based AI applications.
010
Keuxdi @cody.bsky.social · 29/05/2025
Stop trying to fake the data in your pharmaceutical trials 😤
Stop trying to fake the data in your pharmaceutical trials 😤
000
Keuxdi @cody.bsky.social · 22/05/2025
Finally done working on a big update for esper to allow images and files for all the BYOK providers as well as experimental support for files with local models. Added a lot of little features like reasoning toggle for new Anthropic models, search (over web, news, and X posts) for Grok models, etc.
000
Keuxdi @cody.bsky.social · 15/05/2025
🆕 esper 1.0.13 ⚙️ Advanced parameter controls like temperature and max tokens 📁 Export chat as markdown or json ✨ UI updates
000