Education and digital design.
Canada
Joined May 2018
- Tweets3.6K
- Following899
- Followers171
- Likes14.3K
Many designers and developers still struggle with mobile-first, it is so early for ai-first experiences.
It is easier for me to manage many projects with a few agents each than a single project with many agents. Definitely a skill issue, mostly from inconsiderate workitem management leading to frequent minor merge and write issues.
Grading final assignments for my Code for Designers course. — 'felt like a rollercoaster,' well done students!
ALT Screenshot of testimonial "Overall I'm glad I took the class although it felt like a rollercoaster. Thank you professor Sanderson!
Jonathan Sanderson retweeted
If you can understand how a three dimensional object can cast a flat shadow then you can understand how a flat plane can cast a three dimensional world
Small specialized intelligence
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence.
> 6 model instances, across text, audio, vision, and omni modalities, generative and decision models.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding
> 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages.
> 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language.
> 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input.
> fine-tune on 1 GPU.
> a new tri-hybrid architecture
> architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒
Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device.
1/ 🧵
Jonathan Sanderson retweeted
No, WebMCP isn't replacing MCP.
MCP connects agents to your services.
WebMCP connects agents to your website while it's open and the user's in the loop.
Here's how they fit together ⤵️
Working on new tool for design
ALT screenshot of desktop wireframe tool that shows an unknown user interface in a sketchy style
Jonathan Sanderson retweeted
A fraction of a degree can change an entire pattern. At approximately 137.507764°, the golden angle, given by 360°/φ², helps explain the remarkable spiral arrangements seen in sunflower heads.
Jonathan Sanderson retweeted
One of the most underrated life skills is being able to see something strange and simply move on #̢̛ͥ̀̽̽ͨ̌̓͆̇̀̽ͦͨͫͯ̏̎̇̈̚͜҉̕͢͏͡҉̸̶̛͢͡҉̵̨̭̳̘̹̫͔͖̜̠͖̫̲̖̱̤̻̖͉͢͞͡ͅͅ͏̸̷̢͑̊̀̐ͣͭͭ̉͂ͨ̽͒̊̎͛ͪ̊̎ͩͧ̀͝͠͠͏̸̵̵̷̵
Released a grayscale trial to encourage critical UI analysis, I hope to add notes over a week's use across five projects
github.com/sanderjson/open-b…
Jonathan Sanderson retweeted
everything you need to know about clairic is in this video
Today we introduced a novel system one model, Clairic, based on regression instead of classification (i.e. decision models like Clef). This enables estimation of continuous values without having to exhaustively list choices upfront. It also enables very efficient extraction of structured data from unstructured inputs. Open weights coming soon!
Backlog has a backlog
ALT screenshot of open-backlog's own backlog that shows a number of workitems organized in columns by status
Add a backlog to any web-app with no extra server, just point agent at github.com/sanderjson/open-b…
