Ask HN: Is Claude Code bad for ADHD?
4 comments
I don't know if/when/how it replaces devs
Never, but it will replace what devs do with something else. No one has really figured out what that is yet. It almost certainly isn't writing code by hand though.
Your list of small projects is a good example of how things are changing. Before AI businesses would have wanted tools like those things, but spending tens or hundreds of thousands of pounds to get them was too much, so they lived without unless there was either a clear need or a cheaper off-the-shelf option. Now businesses can afford to get a small team to spend a month on something, and only spend a little. Price is usually relative to the value of a project, and AI means the price is far better aligned to value for small things.
I'm seeing the same benefit as an Engineering Manager in a small tech company. I can spend an afternoon building a useful tool that I couldn't really justify taking the time to build before when it would have taken me 3 days, so my toolset is expanding fast, and with it I'm getting more effective (I hope.) AI is not taking dev jobs away because we're getting more software, with narrower focus, delivered faster. That's awesome.
Never, but it will replace what devs do with something else. No one has really figured out what that is yet. It almost certainly isn't writing code by hand though.
Your list of small projects is a good example of how things are changing. Before AI businesses would have wanted tools like those things, but spending tens or hundreds of thousands of pounds to get them was too much, so they lived without unless there was either a clear need or a cheaper off-the-shelf option. Now businesses can afford to get a small team to spend a month on something, and only spend a little. Price is usually relative to the value of a project, and AI means the price is far better aligned to value for small things.
I'm seeing the same benefit as an Engineering Manager in a small tech company. I can spend an afternoon building a useful tool that I couldn't really justify taking the time to build before when it would have taken me 3 days, so my toolset is expanding fast, and with it I'm getting more effective (I hope.) AI is not taking dev jobs away because we're getting more software, with narrower focus, delivered faster. That's awesome.
That justification was always the issue before. The cost of entry to a working solution for a problem was always high enough that low or very-low value propositions couldn't jump the starting gate (cost > value). AI assisted coding has lowered that barrier tremendously. Despite the negativity around vibe-coding - this is the sweet spot. The long-tail of lower value prop demand that couldn't justify the cost of getting built. In the enterprise world, RPA lowered some of that barrier. Vibe coding is next level lowering that barrier.
more software, delivered faster, narrower focus.
this is a great line. very well put
this is a great line. very well put
Hi Chris. I don't suppose you've open-sourced the content-writing tool by any chance?
I have ADHD. I've posted a little here on that. This one was cool, 142 comments: https://news.ycombinator.com/item?id=46610156.
Claude is powerful. This week feels like it has moved up a notch (running for 20mins on longer tasks and then getting stuff right). I don't know if/when/how it replaces devs, I know a lot on here have very strong thoughts on it. But for someone like me it is like rocket fuel.
But then ADHD. I've realised a thought I have for a product, an app, a website, an idea just no longer sits in a list of cool ideas. Now I can do stuff. But do I need this stuff?
Is all this activity healthy?
To give you an idea last 6 months I've built:
- A ChatGPT wrapper with a few bells and whistles and lots of system prompt work and sold it to a big IT company on a £30k annual retainer
- A tool that takes a spreadsheet of data (think people/company/lead data) and allows you to construct a prompt using placeholders for any of the data then run each line/query through an AI prompt. Need websites and mailing address for 30k companies. Easy. They paid us £50k (they got a few other bits of marketing as well for this).
- I built an internal tool for our SEO team which takes a blog post and does all the on-page, and some off page optimisations for them at click of a button. So they can do more interesting work. It sort of works, but they don't use it.
- I built a widget that I can push urls at and at end of the day get a personalised 2 person podcast (GoogleLM style) so I can listen on my cycle home to what I couldn't read. The AI API is some 3rd party tool, its neat very like GoogleLM. It's a cool idea but I don't use it
- I built a AI power content writing tool. Two panes. Paste a content brief in left and it is mirrored right. Select text left and prompt against the selection ("Turn these bullet points into a flesh out para focusing on the stats and shock of the data") and it appears in right hand pane. Slowly your brief left turns into proper content right. Supports RAG for background docs. I use it to write almost anything these days.
- An interactive AI report for my business. "20 AI experts and what their advice means for you". The link to site includes an encoded work email of reader, we scrape Linkedin and personalise text before they arrive. We use it as leadgen, its cool.
There are more, I won't go on.
But I did just build (and polish, and put "out there") a tool - https://tryultrathink.com - that uses a browser extension and electron desktop widget to make it easy to capture any thought, note, text, audio, message, link etc. Then you can sort, search, AI etc.. it's cool. It's an ADHD dream I think. I'm using it every day.
But I made it to help my ADHD "constant stream of thoughts". And I used Claude Code to make it. And I think Claude Code is actually the problem.
What do others think? Is it powerful when your "list" can actually be worked on autonomously by AI agents? Deep research is another one. When I have a thought, why would I not run a deep research or Pro prompt at it and get some background. So when I come to it I can detail. I can. So why not? Maybe water use is the reason? That's a good reason.
On the one hand the future (using basically existing AI tech, not even needing anything more advanced) is cool. I can save every thought and before I come back to one of them it has been researched/built/some how progressed. That's cool. That's productive.
On the other hand its a ADHD nightmare. I can't sleep. I building so much with not always selling, testing, getting people to use. The other I fired up Claude Code and put it in my bike basket so it carried on running on my cycle. I even had my laptop on passenger seat in car and checked it at lights. That's bad right?
In the mean time I will keep using ultrathink (yes I named it in honour of Claude Code). Hit me up for a free trial. Maybe it can help you?