I find some programmers (and this is presumably true of any industry) very narrow in their expertise within technology.

Yeah, most programmers are not curious hackers anymore. They are 9-5 white collar workers with hobbies far outside of programming, systems, hardware, etc. It shows very much as soon as you meet one of them. But, like you said, this is true of any industry.

Oh, and pointy jab: these folks are also, in my opinion/experience, the most eager to vibecode shit. Make of that what you will.

"anymore"? Over a decade ago, a coworker had a path for updating some app's files to a database, and it was taking something like 10 minutes on certain test inputs.

Swore blind it couldn't be improved.

By next morning's stand-up, I'd found it was doing something pointless, confirmed with the CTO that the thing it was doing was genuinely pointless and I'd not missed anything surprising, removed the pointless thing, and gotten the 10 minutes down to 200 milliseconds.

I'm not sure if you're right or wrong about the correlation with vibe-coding here, but I will say that co-workers's code was significantly worse than Claude on the one hand, and that on the other I have managed to convince Codex to recompute an Isochrone map of Berlin at 13 fps in a web browser.

I do feel like the industry has taken a nosedive quality wise over covid in particular. Lots of new people only in tech for the money, no deep idea about computers.

But I know stories like yours from a decade past as well. A tale old as time, but compounding in recent years - IMHO.

Could be, but I think the rot I see now predates the pandemic, possibly with reactive, possibly even before then: https://benwheatley.github.io/blog/2024/04/07-21.31.19.html

I blame it on "software eating the world" (in general) - at some point, about two decades ago, it started to become obvious to everyone that programming is the golden ticket to life - an easy desk job paying stupid amounts of money, with no barriers to entry. So very quickly the pool of students, and then employees, became dominated by people who joined in for the pay, not because of interest in technology itself.

Obligatory link: https://thedailywtf.com/. It's full of stories like this.

That’s quite a simplistic one unfortunately - USB 2 and 3 use different controllers in the PC, which it can indeed detect. The sub-flavours of 3/4 less so.

Not sure about your specific car, but a lot of the “consumer friendly” options like OBDeleven, Carly, etc are fantastic. You often have to pay, but a lot of work goes into them and they often just work.

From messing around with these units from various cars, which often need more enablement than these, sometimes it’s nice to just know your interpretation of the wiring diagrams isn’t at fault when chasing down “no lights at all” issues.

Not necessarily. All they have to do is roll a pub key into the update package. Same as any OTA update.

Do Tesla vehicles get VIN-specific updates?

Not sure - if I was designing it, feels like it would be a good way of getting the right build to the right car so that all the HW versions of each module are in line.

I'd imagine that the update includes all the possible hardware, and the update script actually decides which components to use. Like apt on Debian or yum on RHEL.

Interesting - just found this: https://www.pentestpartners.com/security-blog/reverse-engine...

Not had a chance to read it properly but definitely will be!

That was an amazing read, thank you!

It appears that the Tesla is running a full Ubuntu Linux distro. And here's a small quip to entice passers-by to read more:

  > With names ranging from “INDIFFERENT” to “SUICIDE_BOMBER”, there is a list of escalation strategies in the updater binary, which appear to be strategies for retries of downloads and user prompts on the UI.

There’s an accessibility option in iOS to reduce white point that can be mapped

I live a 25 minute train ride from london in a town with about 16000 residents, on a busy street 5 minutes walk from the main station.

My cell is unusable.

Three in Hitchin?

My personal mobile is on EE. My wife's is on Vodafone. My work mobile was on O2.

(We don't have a fourth device that was on Three unfortunately, otherwise we'd have all of the major carriers covered.)

There are plenty of places I've been around the UK where only one of our devices could get any kind of signal.

Three is now owned by Vodafone, so soon (when they sort the technicals) you will be on all the major carriers!

I live a 5 hour train ride from London and my mobile data is also unusable.

I'm currently on Three. I tried both EE and O2, same issue sadly. Not tried Vodafone.

I’ve found that experience fairly common at our Apple Store. They’ve talked me down.

I’m not sure what model I’d trust locally with anything meaningful in Openclaw. The smaller/simpler the model is, the greater the chance of fluff answers is.

GPT-OSS-120 works well.

I’m not sure how this works. A lot of that tool description is important to the Agent understanding what it can and can’t do with the specific MCP provider. You’d have to make up for that with a much longer overarching description. Especially for internal only tools that the LLM has no intrinsic context for.

I can give example.

LLM only know `linear` tool exists.

I ask "get me the comments in the last issue"

Next call LLM does is

`linear --help 2>&1 | grep -i -E "search|list.issue|get.issue")` then `linear list-issues --raw '{"limit": 3}' -o json 2>&1 | head -80)` then `linear list-comments --issue-id "abc1ceae-aaaa-bbbb-9aaa-6bef0325ebd0" 2>&1)`

So even the --help has filtering by default. Current models are pretty good