ChatGPT and the Cost of Silicon Valley Hype
People keep asking what ChatGPT is for, but the sharper question is what it is doing to the way we write, think, and hand off judgment to machines. That is why the latest wave of criticism matters now. Dave Eggers has argued that OpenAI’s chatbot can flatten language and nudge you toward a thinner, less human style of work. He is not wrong to push back. ChatGPT can be useful, fast, and even clever, but it also rewards speed over care. If you use it badly, it can make your writing sound polished and empty at the same time. And that is the real issue, not the demo.
What stands out about the ChatGPT critique
- It is not a blanket anti-AI argument. The criticism focuses on how people use ChatGPT, not only on the model itself.
- Language quality matters. If the output feels generic, the problem often starts with the prompt and the workflow.
- Speed has a cost. Faster drafts can mean less original thinking and less editorial judgment.
- The cultural risk is real. When many people rely on the same tool, writing styles can start to blur together.
- Productivity claims need scrutiny. Saving time is useful only if the final work still has a point of view.
Why ChatGPT feels useful, and why that is not the full story
ChatGPT is good at producing a first pass. It can summarize, rephrase, outline, and answer routine questions with decent speed. That makes it attractive for anyone under pressure, from students to marketers to product teams.
But the first pass is where a lot of people stop. That is the trap. A draft that arrives quickly can feel finished before it has earned that status. Like a kitchen that serves plated food before tasting the sauce, the process looks efficient while the result may still be flat.
Fast text is not the same as good judgment. ChatGPT can help you move, but it cannot decide what matters unless you do that work first.
ChatGPT and the problem of sameness
One of the strongest complaints about ChatGPT is not that it is wrong all the time. It is that it often sounds plausible in a way that drains personality from the page. The prose lands in the middle. Safe. Smooth. A little bloodless.
That matters for writers, editors, teachers, and anyone who depends on voice. If every team starts using the same model in the same way, the output converges. You get fewer sharp edges, fewer odd turns of phrase, fewer sentences that sound like a person actually chose them.
What that looks like in practice
- You ask for a blog post or memo.
- The model gives you a neat structure and tidy phrasing.
- You make minor edits, then publish.
- Over time, the work begins to feel interchangeable.
That is not a technical failure. It is a workflow failure.
How to use ChatGPT without flattening your work
Look, the answer is not to pretend the tool does not exist. It does. The better move is to use it in places where speed helps and judgment still stays with you.
- Use it for rough structure. Start with outlines, not final copy.
- Keep a human pass at the top. Decide the angle before you ask for text.
- Push for specificity. Ask for examples, names, dates, and constraints.
- Rewrite the opener yourself. That is where generic writing shows first.
- Read the result out loud. If it sounds like a corporate brochure, cut it.
Here is the thing. A good editor treats ChatGPT like a junior assistant, not a ghostwriter. If you would not publish a junior reporter’s first draft without checking the facts and sharpening the angle, why would you do it with machine-generated copy?
What Dave Eggers gets right about the cultural stakes
Eggers’ criticism hits a nerve because it is about more than one chatbot. It is about what happens when a tool built for convenience starts shaping taste. That is a larger fight around AI in writing, education, and media. The output can be serviceable, but serviceable is a low bar.
There is also a values question here. Do you want systems that encourage quicker, thinner communication, or tools that force better thinking? The industry keeps selling efficiency, but efficiency is not the same as quality. A bridge can be fast to build and still a bad bridge.
People do not need more text. They need better decisions about which text deserves to exist.
The bigger question for AI tools and products
ChatGPT is a strong product because it makes a hard thing feel easy. That is exactly why it needs pressure from critics, editors, and users who care about the end result. Tools shape habits. Habits shape culture.
So what should you do next? Use AI where it reduces drudgery, then be ruthless about everything it cannot judge for you. If the model saves you ten minutes but costs you your voice, was it actually saving time?
That is the question worth sitting with before the next polished draft lands in your inbox.
