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

My first piece of advice after such a major break up is that you’re probably going to need to move to a different city. Focus on choosing somewhere that excites you, where you think there will be something new for you. And when you get there, advice from others applies. Need to work damn hard at meeting people. Do three big things. Maybe a social group where you just consume something, like film or books. A social group where you make something, like photography etc. And then maybe a course or similar. If you want a good social life, I’d say three different social groups is necessary. Some will shut down or end. Replace them.

Absolute burn. But accurate.

I cannot stand this kind of absolute thinking from the left or right. It’s usually just cope for personal deficiencies.


The option is to quit your job and go get a different one. It amazes me that people choose to work at Meta etc. I mean, it’s good for them, but they are choosing a bit more money whilst harming the rest of society. That’s a really bold move, to say that you just don’t care about other people.

Agreed. You can quit. That is always an option. "Gotta pay the bills" is definitely valid for some small subset of the us population but that certainly doesn't apply to software engineers in a hub like the bay or seattle. These people delude themselves into thinking they "must" have their ridiculous Meta pay to pay for their $2.5M house and their current lifestyle. Golden handcuffs and turn the blind eye to what they are doing.

Are you using Claude Opus 4.5/6?

If not, then you’re not close to the cutting edge.


Until two weeks from now, at which point you'll be hopelessly obsolete. I've seen this treadmill before and am happy to let it settle down first.

This is not a good thing. If you’re not being exposed and skilling up already, you’re likely to be in the camp that is washed away.

If you can’t be exposed to it in your day job, start using Claude opus in the evening so you know what’s coming.


So far I have not seen much skill gain from using LLM extensively.

Maybe I will be replaced by matrix multiplication in my job, but if I need to use LLM at some point I expect little benefit from starting now.

Yes, I tried to use Claude Code two months ago. It was scary, but not useful.


“Not useful” —- one of those moments where you have to be able to adjust your views in the face of new evidence. Humans are so wedded to their beliefs that it can be agonising to let go. I have nothing but respect for people who admit they were wrong, though. I remained a skeptic for a long time, but 4.5 was enough to convince me to adopt for production code.

So far I updated from "meh, not useful" to "scary, not useful".

If it would be useful I would continue to use it, but at this point I would not use even if it would be free, not proprietary and not funding replacing me.


Eh, people have been warning me "you'll be left behind!" about the flavor of the week for decades now and it hasn't happened yet. If it happens it happens.

LLMs without tooling can not do most tasks. It’s the tools and skills that enable them to do things.

So you may as well ask “is there anything a python script can’t do”. It’s just not a meaningful question.


I use both openai and Claude and in the last few months have moved exclusively to Claude, as it’s better.

What were the empty promises?

I don’t know if this is related, but the standards of CS university courses in the UK are objectively way below what they were 25 years ago. You can compare two syllabuses from the same course and the difference is shocking.

It may be AI has raised the bar, but also that junior devs out of uni are just much worse than they used to be.


Youre not wrong.

Theres been a lot stuff going in the UK that goes back to high-school level stuff. Grade inflation etc.

Its a disaster.


Why is AI such a bad writer? Phrasing like this feels like reading Fox News.

I saw someone point out something like: ai makes every sentence count. There’s no building or allowing a point to breathe. Every sentence is an axiom to get the meaning across, and its so grating

It's an interesting way to view it, because what happens in fact is likely the opposite - AI is asked to expand a few bullet points into a blog post

Maybe that's why the writing feels so terrible. The AI is attempting to maximize every sentence while simultaneously expanding on just a few actually meaningful points. And the net result of that dissonance is this rage-inducing vapidity. It's the written equivalent of the Uncanny Valley.

I think it has got past the uncanny valley really - it does read like a human, just a very attention-seeking one, like your typical LinkedIn salesman.

That's probably just default settings though - I asked it to rewrite, and most of the tell-tale signs are gone as I can see (apart from the em-dash)

https://chatgpt.com/s/t_69a46b290fb08191ad3bd93066b8cad4


Making fluff sound grandiose is probably what makes so grating.

> Why is AI such a bad writer?

A better question is "Why can't the devs producing code with AI spot the same poor patterns in the code they are generating?"

Maybe my point is that, to a poor speaker of English, the AI blogpost looks good and reads well. In much the same way, to a poor programmer, the AI produced code looks good and reads well.

In a nutshell, if it generates poor English, WTF would anyone think it generates anything but poor code?


To be honest it still feels crazy that AI is a writer at all. But yeah, not a good one

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

Search: