When Is Open Source AI Not Open Source AI?

July 21, 2023 by Jenny List 18 Comments

The world of AI is abuzz, or at least parts of it are, at the news of Meta’s release of Llama 2. This is an AI text model which is thought to surpass ChatGPT in capabilities, and which the social media turned VR turned own all your things company wants you to know is open to all. That’s right, the code is open source and you can download the model, and Meta want you to feel warm and fuzzy about it. Unfortunately all is not as it seems, because of course the model isn’t open-source and is subject to a licensing restriction which makes it definitely not free of charge for larger users. This is of course disappointing to anyone hoping for an AI chatbot without restrictions, but we’re guessing Meta would prefer not to inadvertently enable a competitor.

Happily for the open source user large or small who isn’t afraid of a little work there’s an alternative in the form of OpenLLaMA, but we understand that won’t be for all users. Whichever LLM you use though, please don’t make the mistake of imagining that it possesses actual intelligence.

Thanks to the CoupledAI team for the tip!

RoboPianist Is A Simulation For Advancing Robotic Control

July 21, 2023 by Lewin Day 10 Comments

Researchers at Google have posed themselves an interesting problem to solve: mastering the piano. However, they’re not trying to teach themselves, but a pair of simulated anthropomorphic robotic hands instead. Enter RoboPianist.

The hope is that the RoboPianist platform can help benchmark “high-dimensional control, targeted at testing high spatial and temporal precision, coordination, and planning, all with an underactuated system frequently making-and-breaking contacts.”

If that all sounds like a bit much to follow, the basic gist is that playing the piano takes a ton of coordination and control. Doing it in a musical way requires both high speed and perfect timing, further upping the challenge. The team hopes that by developing control strategies that can master the piano, they will more broadly learn about techniques useful for two-handed, multi-fingered control. To that end, RoboPianist models a pair of robot hands with 22 actuators each, or 44 in total. Much like human hands, the robot hands are underactuated by design, meaning they have less actuators than their total degrees of freedom.

Continue reading “RoboPianist Is A Simulation For Advancing Robotic Control” →

AI Learns To Walk In 3D Training Grounds

July 21, 2023 by Lewin Day 13 Comments

AI agents are learning to do all kinds of interesting jobs, even the creative ones that we quite prefer handling ourselves. Nevertheless, technology marches on. Working in this area is YouTuber [AI Warehouse], who has been teaching an AI to walk in a simulated environment.

Albert needed some specific guidance to learn how to walk upright, something that humans tend to figure out innately.

The AI controls a vaguely humanoid-like creature, albeit with a heavily-simplified body and limbs. It “lives” in a 3D environment created in the Unity engine, which provides the necessary physics engine for the work. Meanwhile, the ML-Agents package is used to provide the brain for Albert, the AI charged with learning to walk.

The video steps through a variety of “deep reinforcement learning” tasks. In these, the AI is rewarded for completing goals which are designed to teach it how to walk. Albert is given control of his limbs, and simply charged with reaching a button some distance away on the floor. After many trials, he learns to do the worm, and achieves his goal.

Getting Albert to walk upright took altogether more training. Lumpy ground and walls in between him and his goal were used to up the challenge, as well as encouragements to alternate his use of each foot and to maintain an upright attitude. Over time, he was able to progress through skipping and to something approximating a proper walk cycle.

One may argue that the teaching method required a lot of specific guidance, but it’s still a neat feat to achieve nonetheless. It’s altogether more complex than learning to play Trackmania, we’d say, and that was impressive enough in itself. Video after the break.

Continue reading “AI Learns To Walk In 3D Training Grounds” →

Sand Drawing Plotter Runs On ESP32

July 21, 2023 by Lewin Day 3 Comments

Humans have always drawn lines in the sand, whether it’s to communicate a plan of attack or to indicate metaphorically a very real boundary. It’s also something we do just for the aesthetic pleasure, and this plotter from [aidenvigue] is great at performing in just that role.

The plotter traces patterns in the circular sand tray by dragging a small marble with a magnet. This is achieved with a pair of NEMA 17 stepper motors, set up in a polar coordinate fashion. One stepper motor controls the angle, while another motor controls the marble’s distance from the center point of the circle. It’s a simple way to build a circular plotter, and works far better than a Cartesian setup would for this geometry. The build uses an ESP32 as the brains of the operation. It hosts a web interface that allows various patterns to be selected and run on the device. It also runs a set of addressable SK6812 LEDs that light the sand rather nicely.

We’ve seen some great sand plotters before, and have always been particular fans of the larger variety. Video after the break.

Continue reading “Sand Drawing Plotter Runs On ESP32” →

Ask Hackaday: What’s Linux Anyway?

July 21, 2023 by Al Williams 55 Comments

Any time we mention Linux, it is a fair bet we will get a few comments from people unhappy that we didn’t refer to it as GNU/Linux or with some other appellation. To be fair, they aren’t wrong. Linux is a kernel. Much of what we think of as a Linux desktop OS is really from other sources, including, but not limited to, GNU. We thought about this after reading a report from [The Register] that Linux has nearly half of the desktop OS Linux market. Wait, what?

If you are like us, you probably think that’s a typo. It isn’t. But the more you think about it, the less sense it makes. You know that half of the world’s desktops don’t run Linux. But maybe they mean Unix? Nope. So how can Linux have almost half of the Linux market? That’s like saying nearly half of Hackaday readers read Hackaday, right?

Continue reading “Ask Hackaday: What’s Linux Anyway?” →

Hackaday Podcast 228: Bats, Eggs, Lasers, Duck Tape, And Assembly Language

July 21, 2023 by Dan Maloney 1 Comment

Summer’s in full swing, and this week both Elliot and Dan had to sweat things out to get the podcast recorded. But the hacks were cool — see what I did there? — and provided much-needed relief. Join us as we listen in on the world of bats, look at a laser fit for a hackerspace, and learn how to make an array of magnets greater than — or less than — the sum of its parts. There’ll be flying eggs, keyboards connected to cell phones, and everything good about 80s and 90s cable TV, as well as some of the bad stuff. And you won’t want to miss Elliot putting Dan to shame with the super-size Quick Hacks, either, nor should you skip the Can’t Miss sweep with a pair of great articles by Al Williams.

Check out the links below if you want to follow along, and as always, tell us what you think about this episode in the comments!

Download a long series of ones and zeroes that, when appropriately interpreted, sound like two people talking about nerdy stuff!

Continue reading “Hackaday Podcast 228: Bats, Eggs, Lasers, Duck Tape, And Assembly Language” →

2600 Breaks Free From DRM With PDF/EPUB Subscription

July 21, 2023 by Tom Nardi 13 Comments

Hackaday has been online in some form or another since 2004, which for the Internet, makes us pretty damn old. But while that makes us one of the oldest surviving web resources for hacker types, we’ve got nothing on 2600 — they’ve been publishing their quarterly zine since 1984.

While the physical magazine can still be found on store shelves, the iconic publication expanded into digital distribution some time ago, thanks largely to the Kindle’s Newsstand service. Unfortunately, that meant Amazon’s recent decision to shutter Newsstand threatened to deprive 2600 of a sizable chunk of their income. So what would any group of hackers do? They took matters into their own hands and spun-up their own digital distribution system.

As of today you’re able to subscribe to the digital version of 2600 in DRM-free PDF or EPUB formats, directly from the magazine’s official website. Which one you pick largely depends on how you want to read it: those looking for the highest fidelity experience should go with PDF, as it features an identical layout to the physical magazine, while those who are more concerned with how the content looks on their reader of choice would perhaps be better served by the flexibility of EPUB. After signing up you can download the current Summer issue immediately, with future issues hitting your inbox automatically. Load it onto your home-built Open Book, and you can really stick it to the establishment.

While the ending of this story seems to be a happy one, we can’t help but see it as a cautionary tale. How many other magazines would have the means and experience to offer up their own digital subscriptions? Or for that matter, how many could boast readers savvy enough to utilize it? The reality is many publications will be injured by Amazon’s decision, some mortally so. That’s a lot of power to be put into the hands of just one company, no matter how quick the shipping is.