ai

Showing Posts From Ai

I Made Myself a JARVIS Over the Weekend

With some spare parts, a USB speakerphone, and my Home Assistant server, I was able to make a competent JARVIS for my home office over the weekend…and the exciting part is that it’s only going to keep getting better! Last year was the “Year of Voice” for Home Assistant, the open source home automation platform, and the team accomplished a ton of work that enables Home Assistant users to make their own voice assistants.

TidyBot - A Cleaning Robot Configured via a LLM

I’m fascinated by the merger of robotics and LLMs so I was very excited to see the TidyBot project from Stanford, Princeton, Google, Columbia, and Nueva School. In this work, we investigate personalization of household cleanup with robots that can tidy up rooms by picking up objects and putting them away.

Could the Apple TV Work as a Local Server for a LLM-Powered Siri?

On the December 26th episode of “Accidental Tech Podcast” there was a discussion about whether or not a future LLM-powered Siri could run locally on a HomePod or if it would need to run in the cloud. I thought this was an interesting question, and while I agree with their ultimate assumption that Apple would most likely will either use the cloud or require an upgrade to a new HomePod version if they launch a LLM-powered version of Siri, I also think that the Apple TV could potentially play a role here.

MindMac - A Native AI Client Experience

AI is everywhere! …but in another sense, it’s everywhere. Meaning, it’s all over the place. By that I mean, I have to go to 10 different websites to experiment with the various models via their developer playgrounds. It’s a pain. So I went in search for a native client on macOS that could at least interact with OpenAI’s API, and I found MindMac.

AI Headshots Sorta Work

I gave AI headshots a try and it’s not bad. Some are quite good, but I’d call it a 7/10 over all. I made and attached a grid of the results, but I mixed in one real picture. Any guesses? The hardest part was that it required 10-20 selfies with different backgrounds and outfits, but since I don’t take a lot of selfies I had to run around my house finding backgrounds and changing clothes for 30 minutes to generate enough training data!

Context Powers Brand Safety at Studio71

Have you heard about the brand safety concerns on YouTube? Maybe you’ve heard it described as the “Adpocalypse”? Even if you haven’t caught wind of the madness over the last 12 months, surely you understand that anything mixed with “apocalypse” isn’t a good thing. As is common in “whatever-pocalypse” situations there was a fair bit of freak out, but Studio71 went to work!