Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

konjer.xyz

Post reply on HN
Print view View original

21–30 of 58 posts

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#21

Post by mnkm »

rhtgrg wrote: post #7
gardenhedge wrote: post #3

How did you turn your microeconomics textbook into a chatbot with GPT-3?

Something seems amiss. You can ask this chatbot how old Stephen King is and you'll get the answer.

Each book is best at answering questions on its own source material. For now though, it will sometimes answer stuff not related to the book.

The litmus test I use is always asking a new book "Do you know anything about the Chamber of Secrets" -- often times will say no. Sometimes will return the most semantically similar thing it can find.

This will be fixed eventually!

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#22

Post by mnkm »

pvg wrote: post #19

You just had a big Show HN for this a few days ago https://news.ycombinator.com/item?id=34509335 Reposts are fine but usually on major changes and/or after a year, give or take.

Ohh!! Sorry -- that was my first time to HN. Will remember this for the next time!

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#23

Post by pvg »

mnkm wrote: post #22
pvg wrote: post #19

You just had a big Show HN for this a few days ago https://news.ycombinator.com/item?id=34509335 Reposts are fine but usually on major changes and/or after a year, give or take.

Ohh!! Sorry -- that was my first time to HN. Will remember this for the next time!

Not a big deal and the rules, such as they are, aren't super explicit.

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#24

Post by james-revisoai »

nutanc wrote: post #13

I remember seeing a chatbot for fiction from the same website. Looks like it's the same tech but applied to a different book. As the comments have mentioned, it's a fun experiment, but lots of engineering has to be done still to stop out of context answers.

I've been working on that since last year, and notably on GPT-3 and other models on textbooks specifically with a publisher since September - here's how we convinced people: https://www.youtube.com/watch?v=7V3VkNj2bag

The problem is the tech can be wrong and that destroys retention. You need to a) verify statements b) consider citing statements c) Identify out-of-domain or "base GPT model response" answer.

All I can say is it has been incredibly hard so far, but we've made noticeable progress recently, and hit recall of 98% for psychology. Having been in OPs shoes, I wouldn't underestimate how temporary traction or a UI doesn't make up for bad retention for these products... it's the core engineering and NLP focus that matters (that's our bet anyway!)

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#26

Post by rafaelmelhem »

rhtgrg wrote: post #7
gardenhedge wrote: post #3

How did you turn your microeconomics textbook into a chatbot with GPT-3?

Something seems amiss. You can ask this chatbot how old Stephen King is and you'll get the answer.

That happens with all products built on top of Open AI API. There are a few tricks to avoid this.

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#27

Post by marcosdumay »

cloudripper wrote: post #14
nomy99 wrote: post #6

paranoid engineer here.. will chatgpt replace us :|

If the OP turns their macroeconomics textbook into a chatbot, you can ask it how labor has fared in the past with regard to how profits are maximized through industrial optimization. The response to your question directly may illuminate something to the affect of "it depends" or "yes and no". Certainly, big changes are inevitable. I used to work in industrial rope access and I remember when folks in the industry were…

It circled around every concept, gave a lot of references and refused to answer the question.

So, whatever the answer (yeah, you gave the answer), looks like it can wait a while. Politicians, on their turn must be wary. Here is what it said:

> In Chapter 11, we discussed how firms can maximize profits by producing the quantity such that P=MC, entering industries where P>AC, and exiting industries where P<AC. We also discussed how the movement of labor from weaker firms to stronger firms can increase efficiency. In Handbook of Labor Economics, it was found that in the United States, 60% of the increase in U.S. manufacturing efficiency was caused by people moving from weak firms to strong firms. In this chapter, we also discussed how firms can use limited resources, such as labor and capital, to produce goods in order to satisfy as many of our wants as possible. This means that firms must allocate labor and capital to their highest value uses in order to maximize profits.

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#28

Post by marcosdumay »

nomy99 wrote: post #9
nomy99 wrote: post #6

paranoid engineer here.. will chatgpt replace us :|

When I asked chatgpt if it will replace engineers: No, AI language models like ChatGPT are not designed to replace engineers. AI can automate certain tasks and assist engineers in their work, but it cannot fully replace the creative, critical thinking, and problem-solving skills of human engineers. AI is best used as a tool to enhance human capabilities and improve efficiency, rather than as a replacement for human w…

> AI can automate certain tasks and assist engineers in their work, but it cannot fully replace the creative, critical thinking, and problem-solving skills of human engineers.

It very likely can. That it can't is an extraordinary claim with no evidence at all to support it.

It just doesn't right now, and GPT and LLMs in general are almost certainly not the correct kind of AI to do that.

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#29

Post by moistly »

nomy99 wrote: post #20
mnkm wrote: post #16

Earlier quoted context omitted.

see for yourself :) https://www.konjer.xyz/eiman/the-quran

I asked if God was real, and it gave me verses proving it from the book? But I guess this specific training of chatgpt won't crossreference against other religious texts or relavent opinions from the scientific community? etc. Still a really good way to get contextual religious knowledge on a query use case (ex problems someone might be having with their lives..)

Really good, until the day it hallucinates a novel answer that suggests doing something bad.

Re: Show HN: I turned my microeconomics textbook into a chatbot with GPT-3

#30

Post by marcosdumay »

hot_gril wrote: post #2

It gave me a really good explanation of why costs aren't totally passed onto consumers, and what conditions affect how much cost is passed on. Also, apparently the book says something about communism.

When I asked if monopolies produced more or less goods than competitive markets, it got the effect of economy of scale exactly inverted, and finished with the usual textbook answer that monopolies produce less and thus created less value.
Post reply on HN