Nice! You packed in so many things which are often drip fed, and take folk a while to find

Kudos

If you don't interpret the intent as "90% of human code will be AI" but instead as "90% of code being created globally" it's probably a really conservative estimate.

Creating 9x more code is depressingly easy - but zero correlation to percentage of meaningful output

You uh.. ask permission first?

We do not for most of them. Most of our stores are built with Shopify or WooCommerce which have publicly available product data.

Permission for what?

If permission is not granted then what damages are caused by not following that rule?

In fairness 2/3rds of that was probably a regex for email addresses :)

Tried to write yacc/lex for that back in the day of bang addresses and decnet :: chains. What actually killed me was dealing with header line folding. I just couldn't wrap my head around the tokens and the value of CRLF and dealing with pushing things back on the input buffer.

Some things are a "you have to be this high" test to get in board and I wasn't.

The most consistent non-durability I see these days is business logic intertwined with external (API / file system / env vars) resource access - all scattered throughout layers of a codebase.

(Basically a violation of Factor IV in "Twelve Factor App" [^1])

Certainly not the _only_ durability concern present, but it seems to crop up more extremely and in more places.

Dependency Injection can help mitigate this a little but seems to often be a crutch which covers for poor architecture and enables eventually atrocious readability (a sad admission as I'm fundamentally a fan)

TLDR I find a good question to ask is whether the core logic and journeys can be executed in total isolation from any real resources. If it can't, you'll likely have a rough time testing, maintaining, or scaling - or at least changes will be more troublesome than they could be

(NB: bunch of sibling comments saying similar - but felt worth highlighting the local-first / isolated / separable aspect)

[1^]: https://12factor.net/backing-services

> but why would you watch this?

Surely if it's the _only_ video which understands MCP I need no more reason /s

Yeah, funny how property markets always seem to have that same response.

I bet most of those same people would lose their minds if their favourite restaurant tried to double prices overnight. "Yeah we sold a lot of burgers yesterday..."

If a restaurant is so popular that they're often running out of food, it's perfectly reasonable for them to raise their prices.

popularity is irrelevant - the context is day-to-day. Prices change slowly over time of course, but that's different.

If there was less egg available for a given day, McDonald's [^1] don't charge more for a McMuffin, they sell fewer McMuffins

I'd argue AirBnb's approach here is more like Uber's surge rates. Which are clearly more extreme than anything taxi cabs did (bar the occasional bad actor)

[^1]: mcdonalds is stretching the "restaurant" analogy here, but they have a higher consistent turnover so seem like a closer comparison

I’d argue that’s a better time to do it? People will eat less, conserving food.

That’s actually what happened around here, during inflation.

Restaurants are now double what they were, just a couple of years ago; even the cheaper ones.

The prices shot up, and have yet to back down.

I wonder how much "value" would actually be lost from the market if prices were simply fixed a month or two ahead of time to exclude those price shocks.

Interesting that the author chose that title despite the existence of NSAIDS [^1] and paracetamol(acetaminophen) - the latter of which they even mention (but only in the context of combination with opioids)

Did I miss something?

EDIT: answered in sibling post, thanks @ggm!

[1]: Non Steroidal Anti-Inflammatory Drugs

A whole lotta nuance lost with this headline..

A good reminder to be cautious , but I don't think the absolute position justified..

I run voluntary exit interviews with departing engineers - I'm not in HR (more of an engineering advocate / practice lead) and any disclosure of content is entirely opt-in, as it's not remotely necessary.

The purpose of these exit interviews _actually is_ to make us better & hold leadership accountable. (And give an escape hatch to The SNAFU Principle [^1])

Ideally we wouldn't only learn this stuff when someone's leaving, but alas sometimes that's how it goes.

FWIW: generally collate insights and themes from these interviews, periodic skip 1:1s, etc.. and use them to report on broad trends we're doing well / sucking at.

For me these things are invaluable.

[1]: https://en.m.wikipedia.org/wiki/SNAFU_Principle

TLDR: please don't stop doing exit interviews! (But do be very clear on who you're talking to and what their intention is.)

You don't know what their intentions are.

They are like the police or the car salesmen: they do this every day or so, you do it once every 5 years. If you're not careful, you won't even know what happened there, not until you see the consequences. Protect your own career and interests.

The best thing one can do both for oneself and the remaining employees is vote with their feet. If enough employees do that, things may change. Don't burn bridges when you leave.

I work for a boss whose anonymous yearly reviews are atrocious every single year. Nothing ever changes, he's still running the department. The organization is rotten to the top, and no honest feedback will ever change that. I bet not even most of us leaving.

Out of interest - What wasn't working for you?

I use Just inside (and outside) mise, almost exclusively with embedded shell/python scripts. Have used mise tasks a little, but didn't meaningfully add enough for me

Or do you mean justfiles with shebangs?

Either way, curious what problem you hit (and may be able to unblock ya)

From what I remember, the task consisted of several lines of shell and just would just choke at some point. I think it was because I used a new line to visually separate code. I'm not sure, I don't remember exactly as it was some time ago and I just gave up.

Ah fair enough, thanks!

possibly patched now but definitely going to try breaking it now before moving any more stuff into it