Estimated reading time: 4 minutes
Anthropic makes Claude. According to NewsBytes, the company is now hiring human writers in San Francisco and New York. The official “Member of Editorial Staff” listing says USD 295k-345k a year. NewsBytes puts the top of that range at roughly Rs 3.25 crore (that rupee figure is as reported by NewsBytes).
The job is to be a writer and editor who shapes how Anthropic talks about its research, products and company. Reports also mention a copy lead at $255k-320k and a head of copy and content at $320k-400k.
Think about that for a second. Most of us, when a writing job comes up, say “let Claude do it.” The people who build Claude are paying top-of-market money for people to write. I don’t think that is a contradiction. I think it tells us something.
Competent text is now free
AI has made competent writing cheap. A clean, grammatical, reasonably organised draft costs almost nothing today. When something becomes free, it stops being valuable. What stays scarce is taste, judgement and a point of view. Anyone can produce text. Very few people can tell you which text is worth publishing.
I see this in my own work too. The first draft is rarely the problem now. The problem is knowing whether the draft says anything worth saying.
Someone has to decide what to say
Writing is mostly deciding. What is true? What do we leave out? What do we want people to believe about us after reading this? A model can draft any of those answers, but it cannot own them. Someone has to be accountable for the company’s voice, and that person needs to be a human who will stand behind every sentence.
That is why accountability matters. If a claim in a company post turns out to be wrong, nobody blames the model. People blame the company, and the company needs a person who answers for it.
AI is a tool, not a replacement
An AI draft is only as good as the editor directing it. A great writer with AI gets more done, faster. A weak writer with AI gets more mediocre text, faster. The best use of the tool is a strong writer using it for reach, not the tool instead of the writer.
I would put it this way: the tool gives you speed, and the writer gives you direction. You need both, and speed without direction just gets you to the wrong place sooner.
Readers can smell it
People have read enough generic AI prose by now to recognise it quickly, and once they do, they trust the source a little less. A brand’s voice is its moat. If your website, your emails and your posts all sound like everyone else’s, you have nothing that is yours.
They know the limits better than anyone
Nobody understands what Claude does well, and where it falls short, better than the team that builds it. If they still want a top-paid human holding the pen for their most important communication, that is worth paying attention to.
I am not saying you should stop using AI. I use it, and you should too. I am saying the salary on that job listing is a signal about where the value sits.
What I take from this
If you are a founder, a marketer or a professional in India, my advice is simple. Use Claude for first drafts, for research and for speed. But invest in someone, or in yourself, with real taste and a real point of view. Don’t outsource the thinking. The thinking is the part your customers are actually paying for.
So here is my question: who owns your voice, you or the tool?
Related: this is the same thread as why I started writing. If you build for Indian healthcare, the field guide to selling AI to Indian hospitals is the practical side of that.