Hacker News Reader

Comment by rmbyrro | Hacker News Reader

> the worst part is that it’s impossible to distinguish it from real human output

Doesn't that make human content look bad in the first place?

If we can't distinguish a Python book written by a human engineer or by ChatGPT, how can we demonstrate objectively that the machine-generated one is so much worse?

View on HN

That argument might work for content which serves a purely informational purpose, such as books teaching the basics of programming languages, for instance, but it doesn't work for art (e.g. works of fiction) because most of the potential for a non-superficial reading of a work relies on being able to trust that there is an author that has made a conscious effort to convey something through that work, and that that something can be a non-obvious perspective on the world that differs from that of the reader. AI-generated content does not have any such intent behind it, and thus you are effectively limited to a superficial reading, or if were to instist on assigning such intent to AI, then at most you would have one "author" per AI model, which additionally has no interesting perspectives to offer, simply those perspectives deemed acceptible in the culture of whatever group of people developed the model, no perspective that could truly surprise or offend the reader with something they had not yet considered and force them to re-evaluate their world view, just a bland average of their dataset with some fine tuning for PR etc. reasons.

The problem is not that no one can distinguish it. It's that the intended audience (beginners in Python in your example) can't distinguish it and are not able to easily find and learn from trusted sources.

Aren't there already bad Python books written by humans?

I bet ChatGPT can come up with above-average content to teach Python.

We should teach beginners how to prompt engineer in the context of tech learning. I bet it's going to yield better results than gate-keeping book publishing.

There are, but it used to take actual time and effort to produce a book (good or bad), meaning that the small pool of experts in the world could help distinguish good from bad.

Now that it’s possible to produce mediocrity at scale, that process breaks down. How is a beginner supposed to know whether the tutorial they’re reading is a legitimate tutorial that uses best practices, or an AI-generated tutorial that mashes together various bits of advice from whatever’s on the internet?

Personally I don't subscribe to the "best practices" expression. It implies an absolute best choice, which, in my experience, is rarely sensible in tech.

There are almost always trade-offs and choosing one option usually involves non-tech aspects as well.

Online tutorials freely available very rarely follow, let's say, "good practices".

They usually omit the most instructive parts, either because they're wrapped in a contrived example or simplify for accessibility purposes.

I don't think AI-generated tutorials will be particularly worse at this to be honest...

Another great contribution would be fine-tuning open source LLMs on less popular tech. I've seen ChatGPT struggling with htmx, for example (I presume the training dataset was small?), whereas it performs really well teaching React (huge training set, I presume)

If beginners in Python programming are not capable of visiting python.org, assuming they are genuinely interested in learning Python, it would be very questionable how good their knowledge on the subject can really be.

100% agreed.

I've seen many developers using technologies without reading the official documentation. It's insane. They make mistakes and always blame the tech. It's ludicrous...

We can distinguish it. That's what publishers and editors do. It's also what book buyers for book chains used to do. Reviewers, writing for reputable publications, with their own editors and publishers, as well.

Humans, examining things, and putting a reputation that matters on the line to vouch for it.

The fact that Amazon doesn't want to have smart, contextually aware humans look at and evaluate everything people propose to offer up for sale on their storefront doesn't mean it can't be done. Same as how Google doesn't want to look at every piece of content uploaded to YouTube to figure out if it's suitable for kids, or includes harmful information. That's expensive, so they choose not to do it.

Reply on news.ycombinator.com