Welcome to our maker news where we wrap up some fun and interesting stories in the maker world from the last few weeks. The video will come out on Fridays, but we are posting a topic here on Wednesdays to post the video’s sources and to collect any news from the community to potentially include in next week’s as well as just open up general discussion.
This week we looked at:
A Holographic Display with 3D Touch Interaction
You have likely seen a range of holographic displays online over the past few years but researchers at Upnalab have gone one step further and developed a wickedly cool holographic display that supports genuine three-dimensional touch interaction. Their system, named “Flexivol,” projects holograms onto a series of oscillating elastic strips, creating a dynamic surface that users can physically interact with and manipulate through gestures.
While it looks a tad goofy in slow motion, it appears to allow your hand to physically enter the hologram and interact with it. With how the maker YouTube scene has been going these last few years, I will give it until the end of the year before a creator replicates it.
Arduino Introduces Context-Aware AI Assistant for Coding
Arduino has enhanced Arduino Cloud, their online IDE, with an AI assistant designed specifically for Arduino development. Unlike generic coding assistants, the tool has the context of your current code and can not only generate code for you but also help diagnose and solve compilation errors.
What distinguishes this assistant is its training on Arduino-specific code repositories, documentation, and libraries—giving it specialized knowledge that general-purpose LLMs may lack. In some brief testing, we found that it was quite successful at generating code for a game of life to play on the Arduino Uno R4’s LED matrix. We modified the code to induce errors and it was able to identify and fix them (with a very helpful button to implement the fixed code directly into the script). If you want to give this a go for yourself, a free-tier Arduino Cloud account has access to it with a limited amount of messages.
Build Your Personal AI Assistant with Simple Tools
For those who are instead interested in creating their own AI assistant—not by hosting a large language model, but by building a personalized assistant similar to Siri—Geoffrey Litt has a fantastically simple approach that really demonstrates what is achievable in building your own assistant at home.
On his blog, Litt outlines a straightforward method using just a single SQLite table and a handful of cron jobs. The system creates a digital “butler” that can be fed contextual information from pretty much any source you can provide. Things like Google Calendars, telegram messages, emails, local weather etc.
The assistant then leverages the Claude API to process this information and provide personalized responses, but this could also be run with a local LLM. What makes this approach particularly appealing is its simplicity and high customizability, allowing makers to create an AI assistant tailored to their specific needs and preferences. Its 2025, just go and make your own AI agent already.
And if you have any news from the maker world, feel free to post it below and we may include it in next week’s video, until then we will see you next week!
