I had updated a GQL schema and wanted my agent to update the frontend to suit. I told it the server was running, that it could run specific commands to regen types, etc.
It got itself in a loop and killed the running backend process, then searched my filesystem for the changes it thought it needed (not the ones I gave it) in order to run its own copy of the backend.
That is precisely _not_ what I _told_ it to do. The other part of this is that I find, unless explicitly told to ask questions, they don't do a good job of gathering evidence before making such decisions. I'd much rather have my agent ask me a clarifying question than start killing processes at will.
I don't think it's fair to assume that votes are representative of performance. Politics are a popularity contest first and unfortunately that means image is more important than actions.
I lie to my children when I say that the TV needs to recharge after their morning shows. A way to divert their attention elsewhere, but not out of malice.
The business transaction between me and a restaurant is for the food and service they provide. I pay the restaurant, the restaurant pays the labour.
Refusing to tip isn't refusing payment. If it were, I'd be illegal (i.e. theft). Labelling this as "sharing the labour cost" is precisely the problem at hand.
If I hand my pizza delivery driver a $5 bill, I intend it to be a bonus for the driver, not a subsidy for the business. I have zero incentive to pay the business any more than the prices they advertise.
New York City already has a significantly higher population density than Tokyo (~11,300/km2 vs ~6,300/km). It's easier to develop more housing when you have more land.
The intent of the policy doesn't match the real-world implementation of users. Users are lazy. Users will alter a single character or digit in the password and call it changed.
Most people don't use password managers, and some companies block their usage. Now add a requirement of a "secure" password.
To add to this, you don't need to come up with the game mechanics. I'm taking this approach with making a LiveView-based version of the board game Azul.
Time is a consumable. It is not infinite, and it is unique to each individual.
We do not "trade" time in jail between society and an individual. We deny certain freedoms for a period of time. There is no economic equivalent when it comes to time.
Learning is a process that consumes time now for improved capability or efficiency in the future.
Why is B much better here? It enables the officer to claim they felt threatened when no valid threat existed and then exercise lethal force in self defense (whether or not the threat was valid).
I can't think of any reason why someone should be justifiably disappointed they couldn't apply lethal force.
There have been many studies that show bias in the a screening process when gender or ethnicity are known. Ideally, the hiring manager should be given anonymized resumes so they can better make objective decisions about the candidates.
You need _some_ roads, but not to the extent of your typical suburban sprawl. Where I live (Vancouver, Canada), there are sufficient roads for most travel, but the city also has relatively high quality public transit and decent bike routes. I don't think GP meant that a city should have no roads, but rather that transportation infrastructure is almost always going to be an expenditure for a city. America has simply chosen to prioritize cars over other transportation options.
Your analogy works as long as there's a machine doing the washing and drying.
If you replace the machines with people, it makes much more sense for the function definition to define what is and isn't synchronous. If you have to do all the chores yourself, just call the functions synchronously. If you have family members you can delegate some work to, you can call the functions asynchronously and wait for them to complete while you're doing your own work (or twiddling your thumbs).
We have a sample repository with a simple, but broken, web application. We ask our candidates to review the code as they would if a coworker had submitted it as a merge request.
it gives quite a bit of insight about what aspects they latch onto as issues of note. Do they catch the insidious, subtle issues or just the low hanging fruit? Do they understand what stdlib function arguments expect as input? Do they just pass your review through a linter and call it a day?
When deleting a repository in Gitlab, you are required to type the name of the repository in a field. This forces you to change the response based on the particular action.
It got itself in a loop and killed the running backend process, then searched my filesystem for the changes it thought it needed (not the ones I gave it) in order to run its own copy of the backend.
That is precisely _not_ what I _told_ it to do. The other part of this is that I find, unless explicitly told to ask questions, they don't do a good job of gathering evidence before making such decisions. I'd much rather have my agent ask me a clarifying question than start killing processes at will.