Hacker Newsnew | past | comments | ask | show | jobs | submit | throwawayffffas's commentslogin

> A user noticed that their email signature formatting was off in Cora, our AI-powered email assistant. I asked Claude Code to investigate and fix it. By morning, the fix had touched 27 files, and more than 1,000 lines of code had changed. I didn’t write any of them.

Email signature formatting, 27 files, more than 1000 lines of code changes? I would not read that code either, that's automatically rejected. I can't possibly imagine a problem that would result in a subtle formatting bug that would require that level of change.

> migration that moved email_signature from one database table to another

Migration? For a formatting bug? Are you sure?

Stop self-snitching people, this kind of behavior is beyond unprofessional, its negligent. Anyway have fun with your unreadable spaghetti code base.


I hear a lot of "I am not a good enough coder..." "It has all the sum of human knowledge..."

That's a very bad way to look at these tools. They legit know nothing, they hallucinate APIs all the time.

The only value they have at least in my book is they type super fast.


You'll never find a programming language that frees you from the burden of clarifying your ideas.

Relevant xkcd: https://xkcd.com/568/

Even if we reach the point where it's as good as a good senior dev. We will still have to explain what we want it to do.

That's how I find it most helpful too. I give it a task and work out the spec based on the bad assumptions it makes and manually fix it.


He had frontotemporal dementia.


So what are the economics here? How much does it cost? How much power does it draw? What's the cost per gallon?

Side question, why go through the trouble of turning the methanol into gasoline? Methanol burns and modern cars should be able to run on high grade methanol alone. They definitely can run on ethanol.


It's a money pit.

You get x-many joules of energy from combining each carbon atom with two oxygen atoms, then you have to capture that energy and turn it into useful work, but that is a lossy process. So even if your gaoline-from-air machine was 100% efficient you would still need to burn more carbon than it captures just to run the machine, so to make it run you need to supplement it with some other energy source.

If you already have an additional energy source with which you can run your gasoline machine then why make the gasoline at all? Just use the energy source you have directly in an electric car.


It depends if it can produce gasoline at a cost similar to pulling it out of the ground and then distilling it, it might be worth it. There are over 1 billion internal combustion cars around the wolrd. They are not going anywhere anytime soon, powering them this way has incredible advantages for everyone involved even if we ignore climate change.


Oh Yeah, it's the AI. It can't be that GDP growth has stalled to 0.1 per quarter. We are definitely not heading for a recession guys. It's all unicorns rainbows and robot butlers in our future.


> British companies reported that AI had resulted in net job losses over the past 12 months, down 8%

The companies responding know best why they do and don't hire.


Companies lie. Even if companies didn't lie, people within companies lie.


They also answer might these things quite casually. The response may not come from people who know. Its probably been delegated to someone quite junior.

I once saw a survey question (on what the view was of economic outlook and exchange rates) been bounced to someone junior, who then looked to an external source which based its answer partly on the previous version of the same survey.


Saying you are laying people off because AI, makes the company look like it's innovating and embracing new technology. Saying you are laying people off because costs are up and earnings are struggling reflects bad on the image of the company and the performance of the leadership. Everyone has incentives to lie.

What we would be seeing if the AI uptake was productive would be faster GDP growth, and an uptick in the job market as they would be looking for people to leverage the AI into even more productive gains.


I somewhat agree, but:

With your first point that would be in public, but they do not have the same incentives replying to a survey with the promise of confidentiality.

With the second yes, but not in the short term.


Companies tend to lie towards their employees in similar ways as to the public. Very often most if the organization ends up believing the lies. The few people that do not tend to keep shut or go elsewhere.


https://1funny.com/funny-joke-the-forecast/

It is October, and the Indians on a remote reservation ask their new Chief if the coming winter is going to be cold or mild…

:)


If that 8% was representative or even accurate then unemployment would have rocketed.


>It can't be that GDP growth has stalled to 0.1 per quarter.

Growth was 1.1% in 2025, not great but saying it was 0.1% per quarter is deceptive.

>We are definitely not heading for a recession guys.

The consensus for 2026 is 1.4%, definitely not a recession.


It was 0.1 the last two quarters.

2026 just started, these are all forecasts. I guess we will see.


The growth for 2025 was not a forecast.


> Oh Yeah, it's the AI. It can't be that GDP growth has stalled to 0.1 per quarter.

Why not both?


That is always going to be a true statement.

> even those who believe the mission is safe acknowledge there is unknown risk involved.

Also always true.

No guts no glory.


Did they use the jet at all? Or did they have it just for safety?


I made a small tiled map editor in under two hours, including frequent interruptions. Traditionally it would have taken me at least twice that realistically maybe 8 hours with all my nit picking compulsions.


That's funny that's all the things I don't trust it to do. I actually use it the other way around, give it a big non-specific task, see if it works, specify better, retry, throw away 60% - 90% of the generated code, fix bugs in a bunch of places and out comes an implemented feature.


Agreed. Claude is horrible at munging git history and can destroy the thing I depend on to fix Claude's messes. I always do my git rebasing by hand.

The first iteration of Claude code is usually a big over-coded mess, but it's pretty good at iterating to clean it up, given proper instruction.


I give the agent the following standing instructions:

"Make the smallest possible change. Do not refactor existing code unless I explicitly ask."

That directive cut down considerably on the amount of extra changes I had to review. When it gets it right, the changes are close to the right size now.

The agent still tries to do too much, typically suggesting three tangents for every interaction.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: