Turkish isn't pronounced "Turkey-ish". It's just "turk-ish" as in, "of or relating to the ethnic group the 'Turks'". "Turkiyesh" (Turkish is perfectly phonetic, they don't play games with vowels combining to make all sorts of sounds like English) would be a different thing, being of or related to the country Turkiye.
I think the missing piece here is nuance. Of course there are certain tasks that software engineers do that will be replaced. But will AI replace _everything_ a software engineer does?
The most difficult bit about software engineering is to keep a mental model of _everything_ a product does with varying levels of granularity. The way I see LLMs fail at my company the most is that they are very good at the big picture, and very good at the very small picture, but have difficulty moving between those two levels. And especially when changes have occurred or accumulated over time. Most of all production systems have an extremely long tail of gotchas which are only managed by people who have been around for long enough to have some kind of deep storage access in their heads to those little tidbits of information.
And I think current LLMs might be fundamentally incapable of replacing that.
" Bullough gives the example of a Mexican drug dealer who smuggles product across the border to the US. The drug in question would once have been marijuana, then cocaine, and is now likely to be fentanyl, which is cheap to manufacture and easy to conceal. The drugs are sold in the US for cash, which is used to buy, say, agricultural equipment. "
Wouldn't the person buying the tractor in the US for $$$ have to show where that money came from? Can you show up to John Deere with over a million dollars _in cash_?
The short answer is yes. You can buy cars, trucks and tractors for cash. The more expensive the car, the easier it often is. Luxury cars in particular are routinely bought and sold for cash.
Not in the US without the dealer doing all the same work a bank accepting $100k of cash in a duffel bag would be doing. Plus filling out a suspicious activity report on top of it all. They might have a real hard time explaining to the feds that they truly did "know their customer" from a simple form and a photocopy of your ID.
Some dealers might be willing to do this for you, but most will not. They will direct you to your local bank to deposit the money and get a cashiers check instead. They do not want the liability of it all. Perhaps better chances at the Ferrari dealer you've bought 14 cars from over the past 30 years I suppose?
I asked my (luxury) dealer if I could pay cash the last time I bought a car and they basically said “hell no, we haven’t done that in over a decade”. The risk of being caught up in some drug money investigation or whatnot is too great.
Coincidentally showing up to your bank with a duffel bag worth of cash to deposit is a great way to both get your accounts closed, as well as be added to a blacklist so it will be very difficult to open an account anywhere else.
I used to work for GM as a field rep (in the 80s). There had been enough instances where finance managers skimmed/embezzled some of the cash (even before the feds required filing SARs) that dealers stopped taking cash as a policy decision.
Citation needed? Where did you hear that this is a routine occurrence? That seems risky for everybody involved, and it requires a report to the government from the seller.
Because the legal system in most Western countries is set up that the seller bears liability for laundering money if they accept duffel bags of cash for a car without the same documentation a bank would require.
This is absolutely not true in the US. Are you trying to tee up one of those "the US is not a developed country" type quips that are popular around here?
A dealership may take issue with it but a private party accepts no liability by taking cash.
Firstly, I said "Western" not "developed", you need to calibrate your quip detector lest you become what you dislike.
Secondly, while it's possible to construct a private party transaction in the US where this is fine, if the person spending the duffel bag acquired the money illegally, you the seller are liable if you should have known. "Willful blindness" when accepting illegal proceeds makes you liable too. See 18 U.S.C. § 1957
Maybe I misspoke by saying you would be liable for laundering specifically, but certainly accepting that money is a crime if you have any reason to think it was ill-gotten. And that's a huge risk that no one wants to take on.
I don't know about the US. The EU limit on cash transactions differs by country, with a legal maximum of 10k€. Belgium and the Netherlands for example are at 3k€.
This has been my understand as well. I have CKD and my doctors have always been chill about it as long as I stop taking it about a week before having blood work done.
EDIT: I don't do 25g though... sounds like a lot...
You're lucky your doctor is aware of it. I've had several who did not understand, and insisted creatine must be dangerous if it elevated creatinine levels, and/or didn't understand the effects.
If your egfr is above 80 it’s you won’t notice it. But CKD is a late in life problem so you are basically making your end of life worse by making your kidneys work harder now. You might as well take up smoking: claiming it doesn’t hurt now is a lot different to 20 years of constant use.
I find it freaky how you notice the language change between models. Some words which pop up now all the time, that I don't remember reacting to with previous models, such as "honest(ly)" and "load-bearing". Feels like a new AI smell, like em-dashes or "it's not just x, it's y".
I’ve never asked Claude to re explain things so many times with this. The language it chooses is bizarre, not quite technical enough to be precise, and too hand wavy to be useful.
There also seems to be other issues with the sub agents and not inheriting memory.
E.g. I’m working on an analysis, absence of row implies a zero (it’s sparse).
Every time it checks the results it presents this missing data bug, but it’s not a bug, and I’ve explained it 10 times.
I mean, at finding issues it’s been great. I can ask it to look in other repos and its finding cross repo behaviour or subtleties much faster.
But it’s wasting more tokens reporting caveats I’ve already explained. And it’s suggesting stopping analysis when there’s still work to be done.
So overall I’d say it’s mixed. Feels better at code but worse at my explicit preferences and objective analysis.
the difference IMO is that now businesses have to find a product-market-fit faster. you can't spend 5 years building the perfect system. you get to spend 1 year building a system somebody will pay to use.
I keep hearing that claude is supposedly so agreeable. This doesn't agree with my experience. Claude will often tell me that I'm wrong, and insist on its own solution being right even when I tell it it's wrong.
This is a very recent model behavior change: for me, Opus 4.6, Gemini 3.1 Pro, and ChatGPT 5.4(ish) -- prior models and harnesses suffered much more from sycophancy.
(I still prompt some questions and reviews with "our intern suggested..." to allow models to judge the quality of the content apart from the messenger)
- Code lives longer than you expect. You forget sooner than you expect. Make a readme/architecture overview/theory of operation document. Put more in it than you think is needed. Check it in with the code.
This is one usecase where I'm thankful for AI: to help generate a bunch of documents in a well structured manner (giving it audience, having it scan through my comments on why I did xyz, etc)
I mean... so did a lot of other rulers. As far as emperors go, Aurelius wasn't that bad. You have to judge historical people by their peers, not by your own modern standards.
1. the turkish government had reasons for trying to get people to use "Türkiye" instead.
2. It's still not working.
I 100% thought this was about birds until clicking
reply