Christmas Comes Early With AI Santa Demo

May 18, 2025 by Tyler August 7 Comments

With only two hundred odd days ’til Christmas, you just know we’re already feeling the season’s magic. Well, maybe not, but [Sean Dubois] has decided to give us a head start with this WebRTC demo built into a Santa stuffie.

The details are a little bit sparse (hopefully he finishes the documentation on GitHub by the time this goes out) but the project is really neat. Hardware-wise, it’s an audio-enabled ESP32-S3 dev board living inside Santa, running the OpenAI’s OpenRealtime Embedded SDK (as implemented by ExpressIf), with some customization by [Sean]. Looks like the audio is going through the newest version of LibPeer and the heavy lifting is all happening in the cloud, as you’d expect with this SDK. (A key is required, but hey! It’s all open source; if you have an AI that can do the job locally-hosted, you can probably figure out how to connect to it instead.)

This speech-to-speech AI doesn’t need to emulate Santa Claus, of course; you can prime the AI with any instructions you’d like. If you want to delight children, though, its hard to beat the Jolly Old Elf, and you certainly have time to get it ready for Christmas. Thanks to [Sean] for sending in the tip.

If you like this project but want to avoid paying OpenAI API fees, here’s a speech-to-text model to get you started.We covered this AI speech generator last year to handle the talky bit. If you put them together and make your own Santa Claus (or perhaps something more seasonal to this time of year), don’t forget to drop us a tip!

MCP Blender Addon Lets AI Take The Wheel And Wield The Tools

May 18, 2025 by Donald Papp 8 Comments

Want to give an AI the ability to do stuff in Blender? The BlenderMCP addon does exactly that, connecting open-source 3D modeling software Blender to Anthropic’s Claude AI via MCP (Model Context Protocol), which means Claude can directly use Blender and its tools in a meaningful way.

MCP is a framework for allowing AI systems like LLMs (Large Language Models) to exchange information in a way that makes it easier to interface with other systems. We’ve seen LLMs tied experimentally into other software (such as with enabling more natural conversations with NPCs) but without a framework like MCP, such exchanges are bespoke and effectively stateless. MCP becomes very useful for letting LLMs use software tools and perform work that involves an iterative approach, better preserving the history and context of the task at hand.

Unlike the beach scene above which used 3D assets, this scene was created from scratch with the help of a reference image.

Using MCP also provides some standardization, which means that while the BlenderMCP project integrates with Claude (or alternately the Cursor AI editor) it could — with the right configuration — be pointed at a suitable locally-hosted LLM instead. It wouldn’t be as capable as the commercial offerings, but it would be entirely private.

Embedded below are three videos that really show what this tool can do. In the first, watch it create a beach scene using assets from a public 3D asset library. In the second, it creates a scene from scratch using a reference image (a ‘low-poly cabin in the woods’), followed by turning that same scene into a 3D environment on a web page, navigable in any web browser.

Back in 2022 we saw Blender connected to an image generator to texture objects, but this is considerably more capable. It’s a fascinating combination, and if you’re thinking of trying it out just make sure you’re aware it relies on allowing arbitrary Python code to be run in Blender, which is powerful but should be deployed with caution.

Continue reading “MCP Blender Addon Lets AI Take The Wheel And Wield The Tools” →

This Week In Security: Lingering Spectre, Deep Fakes, And CoreAudio

May 16, 2025 by Jonathan Bennett 5 Comments

Spectre lives. We’ve got two separate pieces of research, each finding new processor primitives that allow Spectre-style memory leaks. Before we dive into the details of the new techniques, let’s quickly remind ourselves what Spectre is. Modern CPUs use a variety of clever tricks to execute code faster, and one of the stumbling blocks is memory latency. When a program reaches a branch in execution, the program will proceed in one of two possible directions, and it’s often a value from memory that determines which branch is taken. Rather than wait for the memory to be fetched, modern CPUs will predict which branch execution will take, and speculatively execute the code down that branch. Once the memory is fetched and the branch is properly evaluated, the speculatively executed code is rewound if the guess was wrong, or made authoritative if the guess was correct. Spectre is the realization that incorrect branch prediction can change the contents of the CPU cache, and those changes can be detected through cache timing measurements. The end result is that arbitrary system memory can be leaked from a low privileged or even sandboxed user process.

In response to Spectre, OS developers and CPU designers have added domain isolation protections, that prevent branch prediction poisoning in an attack process from affecting the branch prediction in the kernel or another process. Training Solo is the clever idea from VUSec that branch prediction poisoning could just be done from within the kernel space, and avoid any domain switching at all. That can be done through cBPF, the classic Berkeley Packet Filter (BPF) kernel VM. By default, all users on a Linux system can run cBPF code, throwing the doors back open for Spectre shenanigans. There’s also an address collision attack where an unrelated branch can be used to train a target branch. Researchers also discovered a pair of CVEs in Intel’s CPUs, where prediction training was broken in specific cases, allowing for a wild 17 kB/sec memory leak.

Continue reading “This Week In Security: Lingering Spectre, Deep Fakes, And CoreAudio” →

Welcome Your New AI (LEGO) Overlord

May 15, 2025 by Al Williams 13 Comments

You’d think a paper from a science team from Carnegie Mellon would be short on fun. But the team behind LegoGPT would prove you wrong. The system allows you to enter prompt text and produce physically stable LEGO models. They’ve done more than just a paper. You can find a GitHub repo and a running demo, too.

The authors note that the automated generation of 3D shapes has been done. However, incorporating real physics constraints and planning the resulting shape in LEGO-sized chunks is the real topic of interest. The actual project is a set of training data that can transform text to shapes. The real work is done using one of the LLaMA models. The training involved converting Lego designs into tokens, just like a chatbot converts words into tokens.

There are a lot of parts involved in the creation of the designs. They convert meshes to LEGO in one step using 1×1, 1×2, 1×4, 1×6, 1×8, 2×2, 2×4, and 2×6 bricks. Then they evaluate the stability of the design. Finally, they render an image and ask GPT-4o to produce captions to go with the image.

The most interesting example is when they feed robot arms the designs and let them make the resulting design. From text to LEGO with no human intervention! Sounds like something from a bad movie.

We wonder if they added the more advanced LEGO sets, if we could ask for our own Turing machine?

Hackaday Links: May 11, 2025

May 11, 2025 by Dan Maloney 6 Comments

Did artificial intelligence just jump the shark? Maybe so, and it came from the legal world of all places, with this report of an AI-generated victim impact statement. In an apparent first, the family of an Arizona man killed in a road rage incident in 2021 used AI to bring the victim back to life to testify during the sentencing phase of his killer’s trial. The video was created by the sister and brother-in-law of the 37-year-old victim using old photos and videos, and was quite well done, despite the normal uncanny valley stuff around lip-syncing that seems to be the fatal flaw for every deep-fake video we’ve seen so far. The victim’s beard is also strangely immobile, which we found off-putting.

Continue reading “Hackaday Links: May 11, 2025” →

This Week In Security: Encrypted Messaging, NSO’s Judgement, And AI CVE DDoS

May 9, 2025 by Jonathan Bennett 3 Comments

Cryptographic messaging has been in the news a lot recently. Like the formal audit of WhatsApp (the actual PDF). And the results are good. There are some minor potential problems that the audit highlights, but they are of questionable real-world impact. The most consequential is how easy it is to add additional members to a group chat. Or to put it another way, there are no cryptographic guarantees associated with adding a new user to a group.

The good news is that WhatsApp groups don’t allow new members to read previous messages. So a user getting added to a group doesn’t reveal historic messages. But a user added without being noticed can snoop on future messages. There’s an obvious question, as to how this is a weakness. Isn’t it redundant, since anyone with the permission to add someone to a group, can already read the messages from that group?

That’s where the lack of cryptography comes in. To put it simply, the WhatsApp servers could add users to groups, even if none of the existing users actually requested the addition. It’s not a vulnerability per se, but definitely a design choice to keep in mind. Keep an eye on the members in your groups, just in case. Continue reading “This Week In Security: Encrypted Messaging, NSO’s Judgement, And AI CVE DDoS” →

AI Brings Play-by-Play Commentary To Pong

May 6, 2025 by Al Williams 6 Comments

While most of us won’t ever play Wimbledon, we can play Pong. But it isn’t the same without the thrill of the sportscaster’s commentary during the game. Thanks to [Parth Parikh] and an LLM, you can now watch Pong matches with commentary during the game. You can see the very cool result in the video below — the game itself starts around the 2:50 mark. Sadly, you don’t get to play. It seems like it wouldn’t be that hard to wire yourself in with a little programming.

The game features multiple AI players and two announcers. There are 15 years of tournaments, including four majors, for a total of 60 events. In the 16th year, the two top players face off in the World Championship Final.

There are several interesting techniques here. For one, each action is logged as an event that generates metrics and is prioritized. If an important game event occurs, commentary pauses to announce that event and then picks back up where it left off.

We really want to see a one- or two-player human version of this. Please tell us if you take on that challenge. Even if you don’t write it, maybe the AI can write it for you.

Continue reading “AI Brings Play-by-Play Commentary To Pong” →