Longtime Linux+Unix user here too, I'm in the same boat, and it's been stunning what it can do.
A few days ago we were having networking problems, and while I was flipping over to my cell hotspot to see if it was "us or them" having the problem, a coworker asked claude to diagnose it. It determined the issue was "a bad peering connection in IX-Denver between our ISP and Fastly and the ISP needs to withdraw that advertisement." That sounded plausible to me, I happened to know that both Fastly and our ISP peered at IX-Denver. That night I reached out to the ISP and asked them if that's what happened and they confirmed it. In the time it took me to mess around with my hotspot, claude was doing traceroutes, using looking glasses, looking at ASN peering databases...
It is REALLY good at automating things via scripts. Right now I have it building a script to run our Kafka rolling updates process. And it did a better job than I did at updating the Ansible YML files that control it.
I've been getting ready to switch over to NixOS, and Claude is amazing at managing the nix config. It even packaged the "git butler CLI" tool for me; NixOS only had the GUI available.
I'm getting into the habit of every few days asking it: "Here is the syslog from my production fleet, review it for security problems and come up with the top 5 actionable steps I can take to improve." That's what identified the kafka config changes leading to the rolling update above, for example.
What if I fold you... that using 1960s technology it would be easier to just go to the moon than it would be to fake it? Have you SEEN a 1960s movie with SFX?
>One year ago, the models were only slightly less competent than today.
That has not been my experience. This weekend I pointed Claude Code+Opus 4.6+effort=max at a PRD describing a Docusign-like software. The exact same document I gave to Claude Code+Opus 4.5+Ultrathink around 6 months ago.
The touch-ups I needed after it completed implementation was around a tenth that it took with 4.5. It is a pretty startling difference.
Agree with this. Opus 4.6 thinks of things I didn't even put in the spec, but absolutely need. It thinks around all the edge cases and gotchas. And I love the way modern AI UIs stop in their tracks and have you answer a bunch of questions about all the ambiguities you left in the spec.
They still do dumb shit from time-to-time, but it's getting rarer.
Agreed, this article seems to be dancing around the point: WHY are the Gen Z hating AI? We have a political ruling class that is all too willing to throw everyone under the bus if they aren't living up to some expectation, and the political class is being driven by an economic ruling class that largely seems to have the same opinion.
Gen Z would likely have a very different opinion if their basic living necessities were available to them.
I've been wanting to do this for years. I fully support (and have paid more than most into) John's shareware, but that means that I can't just "apt install" it, which means I rarely have it available on my various machines. Having something I can just "uv run" that keeps most of the same ergonomics would be a nice alternative.
No: It is my understanding that AI Doc premiered at the Sundance Film Festival Jan 27, 2026, and the news about OpenAI and the lethal strikes broke a month later on Feb 27.
However, it does cover at some length the topics of AI use in military engagements and the larger topic of escalation and "cold war".
Any recommendations on inexpensive colo for personal projects/servers? A few years ago ran across a few links for places to host a box and I didn't save them, and have regretted it.
ISTR one was basically just industrial office space that was running a lower-tier colo, and another was some guys in a metro area that got a rack in a data center and were spreading the cost around with other like-minded folks. At my work I have machines in an Iron Mountain facility, but for personal projects I don't need anything like that, but I'd like something that's more capable than AWS that I'm paying $80/mo for a couple VMs.
EC2 is pricey if it's all you're getting from AWS. If you have wierd requirements, colo may be a good option. Otherwise, just get a VPS or 3 and be done with it. You'll get a virtual KVM that lets you boot it off an ISO and set it up the way you want.
Vultr, DigitalOcean, Linode, ... are long established VPS players.
I'm cheap and buy VPSes off deals on lowendtalk.com. e.g. my backups are on a VPS with 3TB disk, 2GB RAM, 1 vCPU, USD7/mo. I suspect your USD80/mo budget would stretch to something amazing, by comparison.
I'm familiar with Vultr, DO, Linode, none of them seem to have pricing that is compelling beyond EC2. OVH and Hetzner have some potentially compelling plans, but then I'm only saving a little bit.
The call-out for colo is largely to save me from having to engineer a setup at my house for getting my Dell R720 with 256GB RAM online (switching to bridging, setting up a firewall/load balancer with backup, segmenting the networks). That does become easier if I decide I'm ok with 1gig rather than bumping it up to 10gig.
I've been working on trying to help fill out the MeshCore network in our area for off-grid communications. Some of us are setting up solar powered, battery backed MeshCore nodes, they have no connection to power or Internet. You can use a small device (like a credit card or a small walkie-talkie) with a phone, or a blackberry-like device, to send/receive encrypted messages, chat on channels, or communicate on BBS-like "room servers".
It's interesting for if there were some sort of disaster impacting the cell network, or for use in the back woods where you have no cell contact. But it's extremely unreliable. My coworker who is into it, he lives 2-3 miles away but we can rarely communicate because he lives in a bit of a bowl that we don't have reachability into. Meanwhile I'm regularly getting messages from 30-70 miles away no problem.
It reminds me a lot of HAM radio, where there are other better ways to communicate, but if those ways broke it would be nice to have an alternative.
Part of why I'm doing it is my office has a nice, unused tower on top of it. The office itself is ultimately not that high, but the 20ft tower on top of the 3 story building helps a lot. I can JUST reach the 8ft TV antenna mast at my house, depending on tree vegetation (we'll see how much worse it gets as the trees leaf out this spring). It operates in the 900 MHz ISM band, so it can "punch through" more than WiFi can.
I had watched the first half of it a couple weeks ago and it does indeed sound pretty interesting. WiFi HaLow is a long distance capable variant of WiFi and Reticulum is a mesh network protocol.
If I had more people buying into MeshCore, I might push it to doing something like this, but at the moment I think this sort of setup is beyond anything more than just some simple testing in my case.
They are very similar in function, as another reply said. MeshCore relies more on dedicated repeater infrastructure, and it's design seems to make it more reliable for longer range communications. Most people with solid mesh networks seem to be moving from Meshtastic to MeshCore and having improved experiences with it.
The ones I built I used the Austin Mesh guidelines, and they use a USB battery setup they've found that is reliable.
I wanted to set up some of the 1 watt radios for the repeaters (more powerful output plus better input stages) have a known problem where if the battery power drops down to 3.3v it goes into a "deep sleep" mode that it won't recover from once the battery pack comes back up to voltage, without physically resetting it. This USB pack is known to provide 5v power as long as it can, and then shut off until it's been charged to the point it can provide 5v again.
Then I plug a solar panel into it. There are also the "SenseCap Solar Nodes" that are a total solution (enclosure, solar panel, radio, batteries) ready to deploy.
I have one of mine on a 20ft tower on the roof at work, I'm not messing with that weekly. I will need to do firmware updates periodically, which can be done via bluetooth.
edit: Forgot to mention that other units can just use little LIon battery packs that plug right into the board and can be charged by a charge controller on the board to a less expensive solution, that's more what the SenseCap uses. Also I think my board has some firmware patches to fix the problem. I just wanted something that was known to not require much messing around with.
Came here to mention this. I've had a lot of fun putting up stealth repeaters in trees to build out the Meshcore network. One or two nodes in critical places can light up dozens of square miles for the mesh.
Find your local Discord and get rolling. In the Bay Area it's baymc.org.
I just started trying to get into this, and so far it's been kinda disappointing. I got a couple of heltec v3s, set them up with meshtastic and found zero other nodes in range. I did a range test by carrying 1 node around with me while I walked my dog, and the other one at home. It maxed out at about a block. I know there are other nodes in the area but I can't reach them unless I turn on MQTT, but at that point you're just using the Internet with extra layers. I'm looking into more powerful antennas, etc, but it looks like a big money sink.
A few days ago we were having networking problems, and while I was flipping over to my cell hotspot to see if it was "us or them" having the problem, a coworker asked claude to diagnose it. It determined the issue was "a bad peering connection in IX-Denver between our ISP and Fastly and the ISP needs to withdraw that advertisement." That sounded plausible to me, I happened to know that both Fastly and our ISP peered at IX-Denver. That night I reached out to the ISP and asked them if that's what happened and they confirmed it. In the time it took me to mess around with my hotspot, claude was doing traceroutes, using looking glasses, looking at ASN peering databases...
It is REALLY good at automating things via scripts. Right now I have it building a script to run our Kafka rolling updates process. And it did a better job than I did at updating the Ansible YML files that control it.
I've been getting ready to switch over to NixOS, and Claude is amazing at managing the nix config. It even packaged the "git butler CLI" tool for me; NixOS only had the GUI available.
I'm getting into the habit of every few days asking it: "Here is the syslog from my production fleet, review it for security problems and come up with the top 5 actionable steps I can take to improve." That's what identified the kafka config changes leading to the rolling update above, for example.
reply