I Thought ChatGPT Was Impressive. But Then Someone Asked ChatGPT About a Topic I've Written Two Books On.

 

ChatGPT continues to generate incredible hype, buzz around Silicon Valley, and grand pronouncements that its soon going to replace writers and teachers, even accountants and lawyers.

 

But before you believe any of that, here's a simple test to try:

 

Ask ChatGPT about a specific subject that you know very well. And take a close look at the results.

 

I thought about that when Redditor "FLAPJACKU" asked ChatGPT a simple question: "What can Second Life do to improve its systems?"

 

The answer (see above) reminds me of that famous story about the amazing harpsichord-playing robot from the 19th century:

 

The robot was the toast of Europe, giving command performances to royalty across the Continent.

 

Until someone finally noticed the midget on a keyboard, hidden beneath the robot's dress.

 

ChatGPT's answer about improving Second Life is not wrong, per se (besides some glaring factual errors). It's just so superficial and basic, it doesn't really address the question in a substantial way. Like the famous saying goes: "That is not only not right; it is not even wrong."

 

Just to review some of ChatGPT's suggestions:

 

• Better User Interface: Second Life's user interface can be a bit complicated for new users, and making it more intuitive and user-friendly could help attract and retain more users.

 

Yes, but this doesn't address the fact that Second Life's user interface was improved, but still didn't grow the user base, because most of the established community had moved to Firestorm and other third party viewers. Just for starters.

 

• Improving Security: As with any online platform, security is a constant concern. Improving security measures to protect users' personal information and prevent hacking would be a priority.

 

This is so generic as to be irrelevant. Also -- to take a hot topic -- what about third party platforms like Bonnie Bots which collect publicly available data but many users still feel it invades their privacy?

 

• Mobile Support: Second Life is currently only available on PC, adding mobile support to the platform would make it more accessible to users who prefer to use their mobile devices.

 

Yes, it's trivially true that Second Life should be on mobile. But the entire immersive 3D program? Or just messaging? Or another function? Also should it be a premium app which few new users will buy, or should it be free-to-play with IAPs? And because Apple and Google take 30% off any IAP revenue, is this still cost effective?

 

And so on (and on, and on). Reading the answer, you realize ChatGPT is mostly just giving you back a collection of Google search answers in readable text. It's like pulling the robot's dress back and going, "Hey wait, there's just a midget on a keyboard in here!"

 

To be clear, ChatGPT is still impressive, but it sure seems like an incremental advance to existing technology. We definitely need a better version of Google search (which has been around since the 90s), and ChatGPT could very well be that.

 

Also to be clear, I'm not at all claiming any expertise on AI. But I have written two books about Second Life, so on that particular topic, I can say ChatGPT's answer is painfully inadequate and practically useless. And it makes me suspect ChatGPT probably can't do all the grandiose things other non-AI experts claim it can do.

 

Speaking of which, does anyone know the origin of the harpsichord-playing robot story? I tried to find it on Google but that wasn't very helpful.

 

Maybe I should ask ChatGPT.

Comments

[Martin K.]

I haven't heard any stories about harpsichord-playing robots in the 19th century; but there was a famous chess-playing robot hoax in the late 18th century: https://en.wikipedia.org/wiki/Mechanical_Turk

> ChatGPT is mostly just giving you back a collection of Google search answers in readable text.

Yeah. And maybe that could have a positive effect in education: in the future, you might not be able to pass an assignment by handing in a collection of Google search answers in readable text, because that answer could be generated automatically and is therefore worthless.

[Adeon]

The magic of ChatGPT comes in that it's a two-way conversation - if you want it to go in detail about one of it's responses, you just need to ask it to elaborate.

[Brad Templeton]

A number of people have posted reactions of this sort. I am afraid this has a serious misunderstanding of what ChatGPT is. You may have heard it called a large language model. It only knows language. It knows no facts at all. It only knows things that are sensible and probable language. Not knowing any facts, it's very impressive that it still says correct facts. It is not in any way a failure that it gets some things wrong and makes stuff up. That is what it is supposed to do.

It's answers are amazing, and only viewed as useless because you thought it was something that it isn't.

[Wagner James Au]

> famous chess-playing robot hoax in the late 18th century

Thanks! Maybe I was thinking about the Mechanical Turk all along, or read a fictional story where it was turned into a harpsichord player.

Funny that Amazon has an human-powered automation system called... Mechanical Turk.

[Download Filmplus App]

I read both books of chatGPT, ChatGPT can be used for tasks including writing emails, scripts, and social media copy, and the sign-up to do so is seamless, requiring just an email, phone number, and first and last name.

[Ravelli Ormstein]

First, I asked ChatGPT to write a LSL script that would turn on a light when the sun went down. It wrote a nice script, but it didn't work because the AI was using an LSL function that didn't exist. So I complained about it. The AI apologised and fixed the script by using another non-existent function. This went on until I specifically asked for a script that only used existing functions. The script was usable and in the style we usually write them, with a 300 second timer. The AI kept apologising for making mistakes until I asked it never to apologise again.

Then I asked it to optimise an SQL statement (code for connecting to a database). The result looked clever and inspiring, but it didn't work.

Later I asked it to write a Christmas poem using eight specific words. The result was a soulless set of short sentences using those eight words. But no real story, no rhyme and no Christmas feeling. The same thing happened when I asked for song lyrics.

When I asked it to write short stories, I got better results. I tried it a few times, asking for changes each time. The resulting stories were all of the same length and complexity. Their level was that of a primary school pupil. Only one of the stories had a remarkable ending, which made me save it. So somehow the same story pattern was used every time.

Overall, the results weren't very good. This AI really seems to be just a word processor: it produces correct texts linguistically, but the statements are wrong.

(I used the Deepl Write AI to optimise this comment)

[Nodoka Hanamura]

I have to agree with Ravelli. As someone who has used AI to write short stories, it often needs constant redirection and does not have factual knowledge, and while the first three statements are more than valid, ChatGPT cannot be trusted as an authority on anything, and anyone who does will at best be sorely dissapointed, or at worst, could get themselves or others hurt.

[Rosie Helendale]

So basically its like a search engine. I can see how this could be used as an aid or startimg point for a project you are working on but google does that already.