jimbobimbo

Thank you, well put! Bots reviewing code written by bots is a self licking ice cream cone.

I use similar approach, but use gpt to review plan written by opus.

I have a picture of my car in street view photo made by Bing. Sold the car long ago, nice keepsake.

Behind the scenes is unbearable.

Alternative headline: my university had bought a (cheaper?) plan which doesn't tick all necessary boxes.

My problem recently is to have people read design docs I'm writing.

The most surprising part, however, is that when one model slops this into a plan, another model somehow is able to interpret it correctly enough to produce code to spec.

I see this at work. People are "writing" specs and design proposals with bots. This is noticeable and is a huge turn off. I don't have issues with using bots to aid research, but I'm not reading the doc you slopped together.

"Хотели как лучше, а получилось как всегда."

My instructions for agents have "find a synonym to 'resolve'". Otherwise I havea bunch of Resolver classes with every other method called ResolveThis or ResolveThat.

It is very efficient until it's not. When it's not, it sucks. You don't have to have full on diabetes to experience the suck.

My story is: I had a really bad crash after eating a chocolate, which never happened before. That prompted the bloodwork, which shown that A1c was not where it's supposed to be.

I bought CGM myself, without any doctor's input. Didn't know they exist before reading some selfhelp book which my wife found.

Echoing another poster, it's eye opening how some of the foods you'd never think of, are spiking glucose levels.

Overall, after both me and my wife adjusted diet, much better energy through the day, no crashes.

That's what I thought too, until I got my last bloodwork results and started wearing one of the OTC continuous glucose monitors.

That device helped a lot to switch diet habits.

To be fair, the guy on the picture doesn't even look like Jim Carrey, so...

This is the way. I use gh copilot and have opus interrogate me and write the plan, then gpt review the plan and provide feedback; repeat this multiple times until gpt is either satisfied or starts to nitpick on unimportant stuff. Then sanity check the plan myself and have gpt implement it.

Each implementation is also reviewed by me before merging to master. I complete PRs only when I'm satisfied with the implementation, my feedback is addressed, and I fully understand what is going on. Agents are the replacement for typing and productivity multipliers.

I have big picture view of the product, each plan implements only a part of it, scoped to avoid merging unreviwed slop. Probably slower, but result is much better.

The new 85th St & I-405 interchange is horrible. I took it twice and both times it's too stressful and takes much longer to get off 405.

As a fastmail user I'm glad there was no announcement. Every time a company starts telling me about some bright future, this usually means my user experience is about to sink.

Why would this even be in the chat? Showing recent transactions is a basic functionality of a bank.

I asked Claude to describe an app I was working on and it managed to describe the purpose of the app by looking only at implementation, no relevant docs in the repo. This was truly oh shit moment and I'm using AI assistance on that app since then.

>I kinda just expect it to be all “AWS” trust.

This would be very unwise from security standpoint. Internal access to customer stuff is granular and made hard for internal staff to gain, to minimize chances of screw up intentional or not.

Managed identity is enabled via that endpoint, for example.

When things must be shipped quickly, shit breaks and corners are cut; large orgs are full of disfunction. Not sure if such insight was worth of setting your own career on fire.

The irony! My router died literally an hour ago, and I was on bestbuy to buy a new one. Over 5g connection. That was probably the worst shopping experience I had in a while...

Seriously! I have too many years of software development experience, but I use Visual Studio UX to handle pretty much all git operations. And always merge.

I have better things to do in my life than "internalizing" anything that doesn't matter in the grand scheme of things.

Rebase is easy and not terrifying. Here's a 1k word article on how to do it correctly.

Or just do a merge and move on with your life.

"Once those additional vaccines are off the "routine" schedule, they'll be pulled by the suppliers, because it eliminates exemption from lawsuits"

Why is this bad? From one of the threads - "There IS scrutiny on vaccines, by the scientific and medical community - your "scrutiny" (as presumably neither a PhD in a relevant field or MD) is not valuable or relevant. There is decades of research that says that currently recommended vaccines are safe and effective."

OK, then there won't be grounds for lawsuits or lawsuits will be easily dismissed.

"you can be sure there'll be a bunch of crackpot right-wingers trying to prove each one is "bad" and they'll disappear sooner or later" - This logic can be applied to literally any product, be it a medicine, a vaccine, or any consumer good. Somehow pharma companies are able to sell any other drug without going into bankruptcy.

It is true. Withdrawals from SSRIs are no joke and can take a long time.

"But accepting the full S3Client here ties UploadReport to an interface that’s too broad. A fake must implement all the methods just to satisfy it."

In NET, one would simply mock one or two methods required by the implementation under the test. If I'm using Moq, then one would set it up in strict mode, to avoid surprises if unit under test starts calling something it didn't before.

That's exactly it. What happens in a private branch is an implementation detail and reflects personal work style. Policing that is counterproductive.

"Forward Deployed Engineer" is a bodyshop with LLM.