lemuridaelabs.com
Developer, Technologist, Security Fan, AI/ML Enthusiast @ lemuridaelabs.com
52 posts
34 followers
222 following
Regular Contributor
Active Commenter
comment in response to
post
Have a new blog post coming soon about it, but creating subscriptions for notifications based on your movement activity. Not the world‘s most useful function, but sort of demonstrating some LLM things. You can see some of that on the meshscope site in the AI chat, if you want to tinker.
comment in response to
post
I do actually have a message listener in meshscope for responding to commands but mostly for some agent-type interactions. And so far undocumented so…
comment in response to
post
This is a fun sort of decentralized location tagging project. I like it!
comment in response to
post
So the only thing I could see was nodes sending the same position data constantly with dozens of messages a second. That looks like it was resolved a little while back though. Message volume has been much more reasonable without the flooding!
comment in response to
post
Those updates will be great to work with. MCP is rapidly expanding these improvements are definitely going to help!
comment in response to
post
Thanks so much for your work on #SpringAI.. what a great implementation and feature set. We use it on our Mesh Scope project for talking with IoT mesh data, #MCP and everything. Lots of fun and works great!
comment in response to
post
Sure! I'll be curious to hear how it goes, and I'm happy to give any other feedback or thoughts. Been testing different configurations and things, but of course your mileage may vary based on your needs and environment.
comment in response to
post
So giving some easy devices from Heltec or lilygo a spin is probably the easiest way to get started. The t-echo works well or some of the others.. easy to flash new firmware and try it out!
comment in response to
post
I’m not in that area, but I am happy to chat about some experiences on different types of devices. A repeater or two and a few small trackers does a good job!
comment in response to
post
This device looks really appealing to test out. It might be interesting to give it in a spin in some of the mixed terrain we operate in..
comment in response to
post
That's great, and it's good to see such a well developed project be recognized. It's always interesting to see nodes pop up, and it's hard to go anywhere without finding at least another few Meshtastic folks around.
comment in response to
post
The tunnels are great! The a/b camo style tunnel, the art work and Atlanta history.. worth it just to walk the tunnels!
comment in response to
post
Yeah, I’ve always seen that as a standard practice given that I would like some sort of governance into what is placed into my builds. If people are just hitting the wild wild Internet on every build, I don’t know what to tell them. Maybe.. don’t do that.
comment in response to
post
Sure but you can use a caching proxy to help.. running nexus oss should get your key stuff cached, although I’m curious what the total aggregate size of it is.
comment in response to
post
So yeah it’s really interesting and how should things be represented at a protocol level vs an augment. The vector space for behavior but also for similarity matching on malicious information is interesting as well. Content spanning the federated domains can be matched but identity is hard.
comment in response to
post
Today you could allow people to subscribe to block lists, manually or automatically created, and filter at the receive level rather than expecting the network itself to be clean. You could also look at reputation scoring, and using sources you personally trust rather than network level..
comment in response to
post
Looking back at old truly distributed networks like Usenet there was no central moderation, just individual blocks and killfiles. These files could be shared and used to block malicious actors, not across the network, but from your view.
comment in response to
post
It seems like the biggest threat is not as much technical as using the open platform to fill this with abuse and garbage, tempering the excitement about open discourse and making this into a smaller version of the old site. With AI generated abuse and targeted campaigns it would outpace moderation.
comment in response to
post
If you look at the latest release of valkey it’s really interesting to see the architectural shift starting. It looks like more fundamental decisions are being made that will lay the foundation for whole new capabilities.. the new release is awesome though, and excited to see what’s next
comment in response to
post
I feel seen..
comment in response to
post
Obviously you don’t get up. At least until the cat decides it’s time for you to get up.
comment in response to
post
That looks great! I’m definitely going to recommend it to some folks who I think would benefit from that insight. Getting beyond an application or a system and looking at a broader platform is really difficult for many teams, so looking forward to giving this a read.
comment in response to
post
So it’s kinda science and fantasy fusion.. without giving away too much you have multiple dimensions and nanotech but also more classic fantasy settings too. It’s a good balance and worth a read.
comment in response to
post
I ran across this recently and thought it was well done.. The Frugal Wizard's Handbook for Surviving Medieval England
comment in response to
post
On a whim I just listened to a few tracks, really good! She has a smooth but strong voice.. I like it!
comment in response to
post
An interesting idea, and well said. With engineering it’s potentially more tricky in that going out of scope to improve can cause unintended problems or step on toes, but the motivation leveraging clear team communication channels could work really well. Very positive!
comment in response to
post
Amazing work! What a great ecosystem.
comment in response to
post
Thank you so much for putting this together, it’s a great feature!
comment in response to
post
I saw one system where they spun up multiple different virtual machines for the same software, each set to a site local time zone and could never cleanly reconcile the data, rather than just doing UTC and translating at the display.. tragic