toebert

I wish I could find a way to track my location for this with my phone in my pocket. Nothing to do with Dawarich itself, but after trying a number of apps (the dawarich app, colota, gpslogger) to do it I ended up either with wildly inaccurate location or maybe 1 point for an hour long walk due to the accuracy constraint.

It’s rather frustrating that Google was able to track my location on the same phone in the same conditions with alarming accuracy without much noticeable battery drain, but I can’t seem to manage to achieve it on purpose.

ISP has an outage = you’re cut off from everything, can’t even play offline games.

Data center has an outage = same thing

Once they start paying for enterprise level redundancy for all parts of the system (multiple lines for all homes from isp, your data replicated across DCs, all within reasonable distance to you to retain latency), we get to a point where it’s just a worse experience for probably more money, and a constant dependence on a vendor rather than a 1 time dependence. They have tried to make this work many times with the cloud gaming shit, and it failed every time.

IMO, even ignoring the evil part it is a step down from what we currently have (or rather, had before the asshole crew started buying everything so they rot in storage, driving prices to the moon).

Just remember the pcie port is on the side, it has to go in port first.

Yeah it’s not been uncommon for the AI people to blame QA for “blocking” them simply by identifying the ridiculous amount of regressions.

And of course they’re under delivering because they’re being blocked, not because they can’t even figure out how to get the AI to fix the issues without causing new problems.

I wonder at what point is it better to just let them “fuck about and find out” - of course after some written down confirmation that they accept responsibility for their actions, which is very likely going to be a catastrophe.

Yes, most likely I’m going to earn next to nothing for years. Luckily, I don’t plan on having kids which makes finances a lot easier to sort out and I have done well with my financial planning.

My expectation is that my passive income from my IT career savings will cover my expenses in 2-3 years more or less indefinitely, and my woodworking will cover my spending above that (so something close to minimum wage, after all regular expenses are taken care of).

That being said, if I still needed money and I didn’t already have an established career in IT, I’d probably apprentice as an electrician or plumber right now. IT is just not worth it at the moment IMO.

The tool/tech will stick yes, but the token prices skyrocketing will hopefully reset this forced adoption hype to a more reasonable level.

Or, at least maybe we will be able to buy PC components again - this is the one I’m most excited for.

It’s interesting to hear this side. It sounds like you’re experiencing one of the better outcomes, even if it is still depressing.

My experience has been rather different with people being given AI, particularly in larger teams (I work with a few different companies as a contractor). Some people who were already good remained good but now put in less work (which is IMO the healthiest way), some deliver more but I get the sense they’re burning out because of the amount of reviewing, arguably one of the least fun parts of development - not to mention the cognitive strain of a rapidly changing system.

But the biggest group by far is the average to straight up bad devs being given the magic wishing hat and pouring out shit like there’s no tomorrow with no concept of what is even in their PRs. Especially great when they can just approve each others’ slop code. Think UI being marked ready for release when it doesn’t actually connect to the backend, just returns hard coded mock responses.

The amount of QA’s needed has quadrupled in a team I work with, and they’re still overwhelmed (whilst still having more regressions than ever).

The worst one is the “Claude said…” experience. When someone picks up an issue to troubleshoot and the result is just whatever came out of the AI, with the person putting in zero effort to even try to understand it, but then expecting a response/review from someone else, who is likely going to hate their life.

The interesting part is that we’re both describing the death of the industry, but yours is death by success and mine is death by failure - both caused by LLMs.

Personally, I’m planning a career change into cabinetry because my job has gone from fun to mind numbing, and I don’t think I’d be willing to do the “I’ll unfuck your AI developed codebase” career path for any amount of money.

There’s a long, but interesting article on ollama’s scummy behaviour I’ve been seeing circulating - https://sleepingrobots.com/dreams/stop-using-ollama/

It made me switch from my ollama I never used to localai that I now never use instead.

Single person project with 55k files in it, not pushed to any remote.

I’m gonna go ahead and assume nothing of value got lost here.

You can have *.srv.lan point to your reverse proxy, and then just add something.srv.lan to point to somewhere else as well as another record. The exact record will overwrite the wildcard, but things would default to the wildcard.

Alternatively, your router’s DHCP server should make things accessible under a local domain, sometimes it’s just .lan, or .local, or .home. If it’s nice it’d set the DNS search to that by default so you may be able to just ping the hostname of a device without any domains. Using those, you could keep your wildcard for services behind your reverse proxy vs devices on your network using the hostnames.

You could also consider structuring your domains. E.g. $device.lan = the network interface. *.$device.lan = the reverse proxy on it.