Working on generating a synthetic dataset for gaze prediction in UE5, very dodgy, but will be very interesting to see how well it will generalize to my own face
We taught a robot to play cornhole using Gemini 4 Argon ๐ฆพ
Just by watching a video of humans playing, Argon mapped the physics, coded its own simulation to practice, and taught @Apptronik's Apollo humanoid to sink 7 shots in a row.
Made a little game with claude; you combine cards and a language model decides the result.
Bear + Fin = Bearfish
Night + Seed = Moonseed
Brick + Life = Brickling
Sometimes nothing happens (unless you change the mode), sometimes you end up with a Turtle Bear
#AI
A project I've been working on: Jev for context-aware pasting across macOS
Jev reads over your last 50 clipboard items, and 20 permanent saved records (common pastes you use all the time like social links, contact details etc), and selects the best pick when you paste.
An overview of Narrato's dictation and read aloud capabilities. Everything is run on device with leading open source speech models.
Let us know your thoughts and ideas for what to tackle next :)
#Dictation#Transcription#AI#Codex
A level up for coding agent workflows: clipboard-aware dictation - running completely locally.
Ramble while you screenshot and copy anything on your machine, it will all be inserted and linked in your transcript
#AI#dictation#transcription#Agentic
I totally agree, rambling somehow results in me being able to provide much more context and almost plan while I speak. @narrato_tech's dictation automatically inserts clipboard context into the transcript so the model can see any screenshots, text or files I copy which is awesome
One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anything goes, full stream of consciousness. Sometimes I declare it up top, something like "switching to speech recognition sorry for any typos...". Sometimes I turn it into a small interview of a few turns. But I find that the LLMs are somehow very good at reconstructing long incoherent rambles and often their echo of your own tangle of thoughts comes out quite a bit cleaner than what you started with. The result is that you improve the mind meld and have to correct things less from that point on.
Open-sourced an app I've been tinkering on recently to teach myself German.
Learn the words in the app via puzzles and reviews, and see them placed automatically throughout your browser on any website text via the browser extension
@lemonsqueezy has been a week since my application for my store to go live and I haven't heard back from lemonsqueezy on neither the application or the support request regarding it... Very disappointing as I'm hoping to launch asap