theorchid

I once wrote a similar essay on this topic. When I talk to people, they forward my questions to AI and send me the AI’s answer: https://orchidfiles.com/im-tired-of-ai-generated-answers/

— I thought about this and found there are options A, B, and C of doing X, I like A more but C is the fastest; what do you think?

— Ask Claude.

Thanks. I checked the text using the gptzero service. It shows that 100% of the text was generated by AI. I removed the link from my article.

I checked out the other pages on his website. And his other articles. AI-generated content is absolutely everywhere.

Oh, I'm so tired of AI-generated texts.

Why do you think it wasn't written by a human?

Thank you. I've added a link to your article in my article.

Hmm. Using a script. That's explained in the article)

The fact that I don't like phone calls and that I don't like talking to AI can certainly coexist in the same person without being manipulation.

Building a team to operate based on your own personal preferences is selfish leadership... or even dictatorship.

Is there really anything wrong with that? Most managers manage their teams the way they're used to. Founders build their startups the way they're used to—based on their own experience and mistakes. Founders and managers don't adapt to the team's needs. Instead, they look for a team that will adapt to them.

I agree. But we can call each other just to chat about things unrelated to work. Just to catch up.

Sorry, I forgot to add “I’ve never met a single person, except for managers and recruiters”.

It really depends on the people on the team. If they prefer to call each other and hear each other's voices, there's no need to push them to read emails. If you need people to read text, then you need a different kind of team. Every team simply needs its own management style.

your job IS the meetings

No

You’re responsible for a remit and the performance of people inside the group

Yes

One is not the cause of the other.

As a manager, you should learn to change this about yourself rather than imposing your work style on your team.

I think we need a different approach here. I shouldn't have to change anything about myself, because we're slaves to our habits. I shouldn't force this on people who don't want to work this way. I just need to work with teams that are willing to work this way. And so far, I've always managed to find such teams and such people.

I think it depends on the person. For me, for example, text work best for any kind of question, no matter how complex. On the other hand, for simple questions, when I just want to chat, I'll opt for a call or meet in person.

There have been times when a founder was constantly on calls but allowed me to manage teams without them. He also continued to hold occasional calls with the team, but I wasn't present for those calls. I provided him with written progress reports. I didn't try to convince him to stop holding calls, and he didn't try to convince me that calls were mandatory. But such cases are rare.

I completely agree with you. I was just writing about my approach to project management and manage teams. It works perfectly for me. I suggest this approach at the companies where I work, but I don't recommend it to absolutely everyone. Everyone manages projects in their own way.

I don't want to live in a world ruled by AI agents. I sincerely hope that superintelligence won't emerge while I'm still alive. I don't want to live in this dystopia.

if they don't think it belongs on the site

But then it wouldn't have gotten any upvotes and wouldn't have hit the front page, would it?

You have to create keys, copy the text into a file, and when editing the text, you need to update that file instead of the code. When removing the interface, don’t forget to delete the keys from the file. When moving the interface, you need to reorder the keys in the file. When changing the meaning of the text, you need to change the key name. The benefit of these actions is almost zero if the startup has no users.

Are you suggesting that we build a multilingual architecture right from the start of the startup? Without testing the hypothesis, without testing the market and demand?

That's why I'm really tired of interacting with AI. It helps me a lot with development. But in every other aspect, not so much.

Since there were 0 upvotes, there's nothing interesting in the post :)

This method is used to save money on token costs. When you purchase a $20 GPT subscription, you receive tokens worth approximately $1,000. When you purchase a $100 Claude Max subscription, you receive tokens worth $2,000.

To save money on token costs. When you purchase a $20 GPT subscription, you receive tokens worth approximately $1,000. When you purchase a $100 Claude Max subscription, you receive tokens worth $2,000.

Thank you very much for reading! Yes, the purpose of this essay was to shed light on the challenges of launching startups in stigmatized niches and the consequences of working in them.

but I have yet to see studies proving the societal damage caused by porn.

It doesn't necessarily have to be harmful for it to be stigmatized by society.

Thank you for pointing that out! But since HN allows any title, I chose the one that best suits the HN audience.

I think OP made this whole article up.

Thanks for reading! When writing this essay, I drew solely on my own experience. I’ve often noticed that startups post job listings with misleading job descriptions, especially in stigmatized industries. It’s only after the interview that they reveal what the work will actually entail. Perhaps you simply haven’t noticed such job listings.

Ungate is a Cursor extension that lets Cursor use provider subscriptions through a local proxy and a public tunnel, instead of direct API billing. Right now it supports Claude and GPT via OAuth, and MiniMax via API key.

The main reason this exists is that Cursor can talk to a custom OpenAI-compatible endpoint, but its backend cannot reach localhost, so the extension manages the proxy + tunnel + provider auth flow for you.

Known limitation: this only works reliably with custom model IDs configured through the Ungate dashboard. Cursor’s built-in model IDs can bypass the custom base URL.

I have a Max subscription account. I've been using this service for three months and haven't been banned yet. But that doesn’t mean I'll never be banned. The service proxies requests so that they appear to come from Claude Code.