anecdotical:

I'm a Ham and I wanted to CAT control multiple radios and a couple serial port devices. I currently do this with multiple USB serial ports which is a pain because if you ever move them you'll get a different enumeration and have to reconfigure everything.

I asked an LLM to create a project for a multiport card with hardcoded serial numbers and configurable settings (speed, parity, polarity etc - even inter-byte delays).

It did everything as I requested. The only issue I had is that when I bootstraped the project i selected a different microcontroller target. But the result is now a serial port that 1) works exactly as intended, 2) has serial numbers which allows me to move it around ports without reconfiguring (Linux by-id path), 3) whatever quirks i can either fix or add as workarounds for badly behaving devices. 4) Is not a chinese clone that will refuse to work on windows because the official drivers detect it as counterfeit.

The experience was honestly just fantastic. I do NOT want to write code for a multiport serial card. It's completely boring drudge. There is nothing fun in it. There is just hundreds of #define and compiler guards for everything. This thing had a working solution in 10 minutes, of something that would probably take me an entire weekend to do myself - a weekend of, you know, not actually using my radios because I was too busy writing code. I was able to do what I really wanted (focus on the high level requirements and then actually use the product) instead of digging deep into a datasheet developing a skill I'll never use in my work (I don't do embedded code for work).

its during those sprints when I'm pumping out thousands of lines per day that I write the code that turns out to be the highest quality, requiring the fewest number of bugfixes later

yeah, all of us write (or copy/paste) great boilerplate code. that's not really something to be proud of.

we all make mistakes when writing business functions which are never 25k LOC in a week.

except no. because when you ask it to generate test cases it comes up with cases you didn't think of. and it then run the tests and fixes the errors.

my experience with both 20 years as a software developer in multiple teams, and LLM user, is that they are very similar in results to what a human produces without enough context.

every time I got a "ticket" to add a feature, the ticket wasn't 100% clear, it only provided happy path solutions, and LOL if it ever provided any test cases. The "extras" I added myself by just thinking "maybe it should have this. The LLM behaves 100% the same. except in the case of humans we call them "peer reviews" which usually just means "the control freak lead engineer who wrote the bad ticket is now extending the scope because what you developed exposed the 1000 ways that he didn't think of"

Good point, and I'm not here to argue with you -- the problem when we talk about Costco is the Wing drone's max capacity of 5lbs. That's not a Costco trip -- that's barely a trip to a Costco food court :).

5lbs feels like not enough to really replace most trips to actually stock your groceries, unless you break up your shopping trip into multiple delivery flights. It's much better for impromptu consumption (though that said, I feel like most of my trips to the local hardware store are "oh crap, I need this one thing ... " which would be under 5lbs)

Go fuck yourself, asshole.

(You're welcome)

Dude, relax. It was a joke.

Not to mention that "Judy's" also conflicts with the contraction of "Judy is" ...

the finance bros "ai can't replace my bloomberg terminal" has some serious techbro "ai can't replace coders" vibes

Reasonable stance, though I'd argue "some quarters are special" are a particularly good reason to stick with quarterly reporting -- because you don't get the smoothing effect of, say, bundling in your best quarter with your worst quarter. Most companies will have the same 'best' or 'worst' quarter YoY, so it's less about comparing, say, Q4 of this year with Q3 of this year, and more comparing Q4 of this year with Q4 of last year.

I'm a Pokemon Go player; been playing it pretty intensely for the last 16 months or so, level 74 (of a max of 80).

Firstly, this is not in any way surprising or upsetting. Niantic's been pretty clear for a long time now that they were making location-based games for the purpose of training systems.

Secondly, I should note that POGO does not actually require you to take pictures of anything. It's an option, one way to do what Pokemon Go calls "Field Research Tasks" (FRTs), but "take a scan of that place" FRTs are a small subset of the FRTs you might choose to do (and when I was attempting to get as many as possible on my way to level 74 I ignored all the scanning ones).

Maybe I'm out-of-date or a control freak, but I don't want my codebases to contain custom code that I need to rely on but that I didn't write myself.

Apparently you're just a person who never worked in a team. Working in a team requires accepting other people's code. Code that you may not have written the same way but it works, and it's efficient. It's just not "your way".

If you're working in a team, I hope you're not making your coworker's lives miserable by dissecting every PR with 80+ comments and dragging it for days on end, because what you really want is that they will rewrite it the way you solved it in your head.

It's a tool. You need to know how to use it. But before all, you need to know what you want it to do.

I don't "vibe code". I explicitly tell an LLM what's the output I want. This works great. It's also helped me take care of long-standing low-priority tickets.

For example, I had it rewrite a backend function that reads from DB/returns JSON. But I had it do it "streaming" from the database instead of buffering-and-stringifying the database response. This has been long in my to-do list. I knew how to implement it (as I had done it in the past). I just didn't want to do it because it was a "nice to have" but not a must for our use case. And it's honestly boring to write.

The LLM did it for me in a few minutes.

I also tried "Vibe coding an app" to see how that would work. It didn't. It shows awesome progress at the beginning and then it starts failing. It deletes entire files, rewrites unnecessary parts, keeps looping and burning through tokens so, I honestly don't know what the "vibe coders" are really doing. It just didn't give me any results when I tried it.

don't worry, you will live at the company town and get paid in company scrip

also: FUCK SLASHDOT ADS.

in a world of "age verification laws" and governments pushing for "cashless", and precedents like Canada freezing protesters bank accounts, yes.

governments are turning ultra fascist everywhere. doesn't matter if they say they're left wing progressive. they're after your internet anonymity and want you to keep your money in banks, mostly to avoid tax evasion.

(cue in europeans saying they're not fascist and i'm a dumb american etc and canadians justifying the bank account freezing because it was aligned with the party's interest)

Also, "helped FBI" is true but also misleading. Proton Mail does business in Switzerland, and is subject to Switzerland's laws. The Swiss authorities made a lawful data request of Proton Mail, which Proton Mail had to oblige, and then the Swiss authorities shared that data with the FBI.