They must have changed something recently as when 5.5 first dropped I was unable to make it do anything. It would say it will implement, but would never actually do it, no matter how many times I tried to tell it what it needs to do. It would acknowledge what needs to be done, even create step by step plan and then ask if it should do it. I would confirm and then it will just go around reiterating the plan and that this time it will start. Annoying and funny. Now it doesn't seem to be doing that anymore.
I think that's a failure mode of using the legacy completions API rather than the new responses API. With the responses API, the agent actually goes and does the things it's supposed to do.
Do people communicate to distribute prohibited anti-government propaganda or is it a network of people who otherwise be too shy to talk to each other by other means?
It's primarily just an experimental system. Demonstrating that fixed infrastructure isn't actually necessary to communicate.
Beyond that, it's a mixture of HAM radio for communicating with people outside of your immediate circle, and disaster prep.
The best realistic scenario I can see for using it is after a sever weather event like hurricane, tornado, tsunami, etc. that takes out significant comms equipment. Having an ad-hoc network pop up using battery powered nodes able to setup a secure comms channel to organise aid deliveries would be a powerful tool. But existing infrastructure is resilient enough that it's not actually necessary in modern times.
Beyond that, it's probably more of an IoT type thing. Setup a bunch of nodes across a significant area of land, run machinery, sensors, etc. remotely via a self-healing mesh network.
People forget that this network isnt for everyday use. It is for use in ad-hoc scenarios where cell or even satellite coverage falls apart. The most powerful aspect is that these things are deployable. A communication chain can be established as fast as people can move. Natural disasters are the most obvious use case, but more interesting are things like search and rescue.
Go somewhere properly remote such as the high north. There is no cell network outside of town. And the satellite coverage is spotty at best. Say you need to go look for someone. Meshtastic relays can be up and working in minutes. A chain of rescuers can spread out along a path, and remain talking to each other, as fast as they can move. Sure, radios can do this too, but long range voice radio require serious power and are still largely line-of-sight. Radio relays are an entirely other expensive thing.
Think also of remote camps (logging/planting/fishing/climbing etc). Toss a lora relay on every vehicle and every work party can talk via the app installed on their normal phones. Use GPS-enabled devices and you can passively keep track of every vehicle. Need to operate two valleys over? As the first crew deploys out they can plop down relays at key points. Years ago I setup something like this using wifi relays. It was hell. It never worked right. The range and lower power demands of lora would have been infinitely easier.
The range of these lora nodes is a bit of a myth. It is better than higher frequencies but you shouldn't expect anything more than a km with obstructions, realistically half that.
Not to say they don't fill a niche, but bandwidth and range limit it's viability to small operations; even an optimal network cant handle more than a few hundred tweets per hour.
Keep in mind that this is a very low bandwidth, high latency mesh network.
Great for sending short text messages, absolutely terrible for guiding drones.
Line of sight needed, trivial to jam, power hungry, trivial to fox hunt. These are not boogyman devices. The real boogyman devices are the ones in space. Big militaries don't need things on earth to do any of the things you listed and way more.
Yeah. Actuality I was thinking about less sophisticated adversaries. So called "failed states", nonstate groups, organized crime, amateur surveillance, corporate espionage, sabotage.
People will go on and on about what happens to society when the internet or cell service goes offline, but when they see an emergency solution staring them directly in the face, they wonder what the use case is.
If you see this technology and think "wow! that solves [problem I already have]" - then it's great.
Otherwise, you buy a couple, set them up, spend a week or two sending very slow and unreliably forwarded messages that mostly amount to "hi! i have an ACME 32ABC radio! What do you have?", and then put it in a drawer or sell it on.
It's not all people trying to skirt the law. It's kind of like HAM radio as a hobby. It's fun technology that lets people do cool automation projects and sure with a mesh connect to other people. Imagine you have a few acres of land and want to turn on sprinklers or something.
A lot of people use it just to chat with friends and family in a fun way.
Of course the preppers and privacy evangelists see it as a means to get ready for living in a hostile environment. Being fair to them, things don't look awesome in the US.
I bet a few criminals use it, but it's still very niche.
It runs independently of internet and power. One use case is a group of people in a remote area (hikers, hunters) carrying their own node and being able to communicate via text over several kilometres.
v. a satellite enabled phone that can send text messages over thousands of kilometers?
These people should not be making a short range text solution, they should be building a low bandwidth internet extension with gateways to the real internet. Most of the information content of the internet can readily fit over 56kb lines once you strip out all the fluff. And in an emergency where you need or want a decentralized mesh network, that's more important that being able to text, who exactly?
I used to wonder the same thing and then we bought a vacation home. It's only a 40 min highway drive outside our top 20 U.S. big city metro area. Some friends have said "that's not even far enough to be a 'get away'" and several of our neighbors there commute daily to work in big buildings downtown. Our street is only 12 mins drive from a major interstate highway with the usual suburban superstore sprawl (Target, WalMart, Home Depot, Costco) so it feels like we're in the middle of 'civilization'. But once you turn off the highway, that last 12 mins gets both beautiful (with rugged hills) and also very empty.
Five mins from our house suddenly cell service from all the providers gets very spotty. If you live near the top of a hill facing the right direction, you can maybe jury rig a cellular antenna on a pole. There is legacy POTS phone service via 60 year-old copper but few use it because it's only ISDN barely faster than dial-up and >$100/mo. Otherwise, there was no option for reliable residential phone/data/text service until Starlink became available in our area a couple years ago.
So everyone in our entire area has 2M radios to communicate in emergencies because in four years we've had two fires come close enough to close our roads, been snowed in twice (without power) and a small bridge got damaged in flooding blocking vehicle access in and out for four days. So, sometimes you need to get extremely local information from, and coordinate with, neighbors you can't even see and have to hike to visit. But the old 2M radios have to be monitored in a real-time which feels really antiquated. So, to me, inexpensive LoRa that could enable store-and-forward messaging and conditional unattended alerts suddenly sounds very useful.
I'm using one of those devices with a tiny eink screen... as a pocket ebook reader. Sorry for the bait. Fun and easy project though, lookup Pala one if you're interested and dm me for improved firmware (restrictive license)
I set up my first node after the last major Verizon outage that rendered my cell phone useless as a mobile communications device. Now, when the next outage happens, with the always-on base station that I have at home, I can bring a portable Meshtastic radio out with me, paired to my phone via Bluetooth, and retain the ability communicate wirelessly back home, or with any of the other many nodes in the extensive network here in the NYC / Hudson Valley region. I also enticed a couple of local friends to install them and we often opt to text over the mesh. I see it as a thing that is fun to play around with now, but which may become critical at some point in the future.
There is no real use case, IMO.
I setup a few nodes a couple of months ago. It's mostly no activity, punctuated by some random "can you read me" type messages, and for some unknown reason people who think there is something impressive about them having a node on a commercial flight.
The entire thing would fall over in any kind of scenario where you needed to rely on this janky mesh network as a primary means of communications.
It can be fun/useful for very out of the way things where you have a handful of people out camping, or other off-grid situations. But frankly even in those cases there are far better/established ways to keep in sync if you need to (eg: FRS).
This stuff is mostly a solution looking for a problem.
Just like ham radio, it's a an interesting technical hobby for those that may get excited when their little 0.25W radio hits a repeater 80km away.
More practically, I'm going to try it out while camping this summer. In areas with low or no cell coverage, my phone is useless or dies quickly. Throw a repeater in a tree, and hand your friends nodes.
I would love to be able to get text alerts when an event occurs, from a location that is not connected to the Internet, about a mile away. The need is not critical, so there is no desire to spend money every month. And reliability of the solution does not have to be high either.
It's really for closed user groups. At least meshtastic doesn't allow you to see other people's messages, only those in your own group. They're all encrypted.
I worked R&D on LoRa project a few years ago. Their use case was a long-range emergency communication system for workers in remote areas(no wifi, lte or LEO at that time). Now I see a bunch of applications in this field that aren't what you describe. :)
I built one and found absolutely no use for it. No one ever, and I mean ever, answers you. It's sort of like ham radio, where you get your technician license and get on a net and discover people are just talking about their antennas. Except it's worse, because all the antenna discussions are happening on Reddit and Discord and not on the network itself.
People are very enamored with what you could theoretically do with it, but they never actually do any of it. It's a hardware fetish, it's all about building boxes with solar panels and seeing how many nodes you can light up on the map. Reminds me of another ham radio thing I never got into, "contesting".
This is because big corporations supporting Brexit figured out it will be better for their bottom line if they could source labour from wider pool and have it tied to visa. Something EU workers would never be comfortable with. Hence you had the so called Boriswave - an influx of workers paid below market rates supporting big corporations able to navigate Home Office corrupt system. Conservative party never told the public what it was really about - bringing in very much slave workforce to exploit - at the expense of working class and SMEs.
By the looks of it, Conservative party will never recover from this betrayal and soon followed by Labour who decided to maintain the status quo.
Surely this would be illegal? Personal data without consent?
Or is it a case of too big too fail.
Seems like running governments' infrastructure pays off. No regulator will dare to impose a fine that could collapse the company. But this is very much needed.
£100bn fine and confiscation of assets in the given country could be a start.
> When your workflow consists of asking questions and getting answers immediately, you don't get to see what's nearby.
That's why is very very important to just step out and use saved time to go for a walk, to a park, sit on a bench, listen do birds, close eyes and zoom out.
All should be free. Imagine if government decided to impose 3% revenue tax, yet these companies get a free pass.
If these networks cannot run this for free, then they should be nationalised and tax payer should cover it. It will be cheaper (because it will become non-profit) for everyone and better.
Many banks already require monthly or annual payments for keeping an account with them. They also use the money from deposits to lend it at high interest rates. It is not like the banks are not extracting much more than a fair share of revenue from a captive market.
Not OP, but basically take GiB/s and divide by 30.
You need at least 128GiB to hold the model, too.
It's expensive to get 200 GiB/s, very expensive to get 400 GiB/s and above that you are looking at DC-grade GPUs. Multiple, in fact.
This. At one company we ran out of space in the server room, so the excess machine temporarily landed next to my desk. Dear god. Noise cancelling headphones couldn't cope with the noise.
The Soviet Union gave the equivalent of about 80-100 million USD in support to the ANC in South Africa in the 1970s and 1980s. As soon as they got power, the ANC turned around and aligned with the west, wasting every rouble Moscow spent.
Since the inauguration Trump has supported physical seizures of many different kinds of Russia-aligned merchant shipping and the economic degradation of Russia's allies. Given all of this, we can assume that the Russian asset angle is a much less accurate explanation for Trump's behavior than the alternative theory where he is highly suggestible to the most recent person to heavily compliment him in-person which used to be Putin and has subsequently changed to some mix of Rubio, Vance, Hegseth, Netanyahu and the Trump family.
It's not really about leaders, but people who are supposed to ensure they are not corrupt.
It seems like security services in many countries started outright to scam the tax payers. Get the wage and pretend brown envelopes don't change hands and policies are not shaped by corporations for their benefit, not the public.
reply